Author |
Topic: Fritz's commentary (Read 13031 times) |
|
Fritzlein
Forum Guru
Arimaa player #706
Gender: 
Posts: 5928
|
 |
Fritz's commentary
« on: Jun 1st, 2010, 5:59am » |
 Quote  Modify
|
Congratulations, Mob! You played very solidly throughout, whereas I had one major tactical blunder on 10s. I wonder how my play would stack up to the Mob's strategically without that one mistake. Also it might have been fun to have thirty moves of a viable position followed by ten moves of tortuous despair, instead of the other way around. Still, I enjoyed the game, and I hope I at least put up enough resistance to make your side of it enjoyable as well.
There were moves from the Mob that I thought were slightly inaccurate, but in retrospect this might have been sheer wishful thinking on my part. I would at times believe I was slowly making progress, only to have my hopes dashed every time when the chips were down.
Playing against an opponent that never makes a tactical error is a fearsome experience. I feel totally ground down by it. I'm not sure what would give me the confidence to reprise my role as The One. Perhaps chessandgo, who has won two Arimaa World Championships since this game started, would be willing to give it another whirl? If so, I would be happy to be part of the rabble again.
I was able to predict the Mob's move only nine times out of forty, or 22.5%. That's pretty low, but not as low as I expected going into the game. I am curious to see whether the moves I incorrectly predicted ever got serious attention and/or votes. Also I will enjoy examining to what extent my general perceptions of the game coincided with or diverged from the Mob's perceptions.
I thought I would have a time advantage due to not having to communicate with myself like the Mob members had to communicate with each other, but it turned out to be a distraction for me to focus on the Mob's clock instead of managing my time in a vacuum.
I look forward to any commentary on my commentary. I hope what I recorded of my thoughts is of at least some interest to the Mob.
Thanks for the game!
|
| « Last Edit: Jun 1st, 2010, 1:24pm by Fritzlein » |
 IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
pre-game
« Reply #1 on: Jun 1st, 2010, 6:01am » |
Quote Modify
|
Publicly I have been talking smack about how I am going to destroy the Mob, but that is mostly an attempt to stir up passion and get the game started sooner rather than later, because I am eager to play. I want my shot at glory now, because in the future (a) Arimaa will hit the big time and the Mob will be too large for any one player to contend with, (b) aging will further erode my mental capacity, and (c) computer ability will increase, taking the game away from us humans. Now will be the best chance I ever have to take on the world. Privately, though, I have to admit that even if the game starts tomorrow, my chances of winning are poor.
People who haven’t participated in a team strategy game like this have trouble understanding the dynamic that makes the Mob so powerful. It is easy to attribute strong team play to the domination of a few strong players, as Kasparov did during his match against the World Team in 1999. Kasparov couldn’t bring himself to believe in the power of the collective and therefore assumed that a couple of strong grandmasters were dictating to the hoi polloi. Kasparov was wrong; that was absolutely not what was happening in that game.
My participation in Kasparov vs. the World taught me some valuable lessons about team analysis. First, you can’t measure how much someone is contributing by counting up the number of moves he suggests that the team eventual plays. A few World Team members squandered a great deal of effort trying to take credit for various moves, as if being first to post in some thread about the possibility of a move being played gave them ownership. This is ridiculous; all the moves are ‘out there’ a priori, and a brain-dead algorithm could list them all and suggest each of them first. Even I once managed to be the first one to suggest a World Team move, but that doesn’t mean I understood why the move was good relative to our other options. (In fact, some team members claimed that my move lost us the game.) But I didn’t do anything special for weal or woe; I merely tossed in a suggestion that happened to be played by the team after much debate.
Second, you can’t measure how much someone is dominating the team by counting how often the move they recommend ends up being the move that the team votes for. The recommendation of Irina Krush was voted for by the World Team on almost every move for most of the game, but that was in no way a reflection of her imposing her will on the team. Rather it was reflection of her skills at listening to team input and integrating what she learned into her suggestion. Often the team produced analysis that changed her mind about what move was best. It would be more accurate to say that the team imposed its will on Krush than vice versa.
Outsiders fail to understand the power of the team because the team’s power is in the details. How do you explain that an unexpected defensive resource in one sub-variation invalidates an entire attack? But it does. The primary strength of the Mob is in its refutations, i.e. in its ability to say that move X is bad because it can be answered with move Y. That happened several critical times in the 2007 One vs. Mob game, although a superficial observer wouldn’t see it.
My main chance of winning the 2009 One vs. Mob game rests in a disheartened and/or dysfunctional Mob. If there are a dozen eager contributors working well together throughout, they will play solidly at every turn, whereas I am bound to have more holes in my analysis, due to the greater number of good ideas that never occur to me. On the other hand, if the Mob members don’t cooperate well with each other (e.g. argue rather than analyzing) or if there are too few participants, or if the participants don’t spend much time on the game, then I might not be outgunned, and my good intuitive judgment might overcome the Mob’s greater breadth and depth of analysis.
Bobby Fischer once said, “I like the moment when I break a man's ego.” I don’t play games for that moment myself, but I realize that in this game ego-breaking is my best chance to win. I need to get an advantage early, keep the game under control so that it is obvious that I have the advantage, and squeeze. I need the Mobsters to believe that they can’t beat me because I am just too good for them, and I need them to spend time blaming each other for getting into a weak position. If the Mob despairs of beating me they will stop trying so hard. Demoralization will be my best weapon.
If, on the other hand, the game stays close throughout the middle game, I will be in trouble. The World Team was hugely energized against Kasparov when an early novelty busted open the game and made the position totally unclear. A wild game will give comfort to my enemies. They will start to believe I am beatable. If the position remains nearly equal, or even merely double-edged, late into the game, I am doomed. Not only is the endgame my weakest phase, it is the part of the game where computers and extra eyeballs help the most, because it is so tactical and refutation-prone. I can’t afford to reach an endgame even if I enter it with a slight advantage; the endgame must be prevented by Mob resignation, or by enough mobsters giving up that it is an effective resignation although they keep producing moves.
I must, must, must win the opening.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
1g:
In my recent games with Silver I have been responding to all the standard symmetrical Gold setups by decentralizing my camel and starting both of my horses on the opposite wing. Recently Omar endorsed this plan as well, and I expect many Mob members have taken note. That raises the question of whether the Mob thinks it has something to fear from an unbalanced setup. Quite possibly, the Mob does not attribute my Postal Tournament wins as Silver to my new setup, but rather to subsequent play, in which case they will gladly walk into the teeth of my new strategy and try to beat me anyway.
On the other hand, if enough members are believers in the power of an unbalanced setup, the Mob might consider starting with a flank camel even as Gold. This would be bold because as Silver I could place my elephant directly opposite the decentralized gold camel while keeping my own camel away from the gold elephant. I can’t recall ever having participated in a game that started that way.
All things considered, I count on the Mob to be conventional. Preferential voting guides a group toward consensus, not toward experimentation. Even if a majority of Mob members would like to try something different, they probably won’t agree with each other on which different setup is worth trying. For example, although I have opened with four forward rabbits in every game for almost a year and half now, most folks still consider it risky, unproven, and perhaps leading to disadvantage. I will therefore be shocked if the Mob sets up with any rabbit structure other than 99of9’s.
Indeed, there is a fair chance the Mob will play the complete, standard 99of9 setup, with elephant and camel in the middle and with cats behind the traps. Even the slight variation of a cat behind one trap and a dog behind the other, which worked for the Mob in 2007, doesn’t seem as compelling for Gold as it was for Silver, especially if the Mob expects an unbalanced response from me.
Because I will be making a lot of predictions in this game, such as the one above, I should reassure my readers that I haven't gone back to alter my predictions to appear prescient. On the other hand, I can't promise not to change anything, because it just kills me to go back and find typos I can't fix. Therefore, while I won't promise to leave previous sections entirely untouched, I do promise to edit only for style and not for content.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
1s:
I was stunned to see the Mob set up as chessandgo normally does. Since he is almost the only one to use such a setup, how did it get a majority of votes over the 99of9 setup, which everyone and his dog uses? My first thought was that perhaps the Mob is indeed a bunch of sheep, and chessandgo is indeed dictating. If so, that will make the game much easier for me to win. Dissension is essential for a team to be any better than its best individual.
After further consideration, however, I feel I should not be too hasty to infer the dynamics of the Mob from the setup. The only difference between chessandgo’s setup and 99of9’s is that the dog and rabbit on e1 and g1 have switched places. My limited experience against chessandgo shows that if the opening evolves into a rabbit-pulling race, Silver can’t take advantage of the back-central rabbit on e1 as a target, because Gold is ahead in tempo from the start. On the contrary, in a rabbit pulling race, chessandgo’s setup allows a more comfortable rotation of forces in the east, as the g1-dog follows the g2-horse forward, and the e1- and f1-rabbits can slide to safety if need be, without the piece on f2 needing to move.
In other words, I regard chessandgo’s setup as a tiny technical improvement over 99of9’s setup. If I had been part of the Mob, I would have voted for chessandgo’s setup in preference to 99of9’s, even though I have never played chessandgo’s setup myself. Given my positive evaluation of chessandgo’s setup, it would be absurd of me to take the Mob’s choice as an indication of weakness. If anything it should, on the contrary, warn me that the Mob will play with precision when given two nearly-equal choices.
For my response I am tempted to quickly play my standard response with four rabbits forward and my camel on g7. That would follow my Continuous Tournament game against chessandgo from September 20. I got quite a comfortable position in the first few moves, so that game gives me no reason to fear. The Mob would definitely deviate first, and I am not even sure how they could do so to advantage. Chessandgo’s plan to advance his horse on the wing opposite my camel appeared logical, but he got nothing out of it. He had no hope of an effective elephant-horse attack while I had both of my horses defending that wing. He could at most have pulled a rabbit, which I don’t much fear. I’m not out to pull rabbits these days unless the position gets slow. In a fluid position, I expect I can put my time to better use.
Otherwise, there is no game history for me to follow. When I played Silver against chessandgo in the 2008 Postal Mixer, I set up my camel on the same wing as his elephant. I now regard that as a relative loss of time, and won’t repeat it. Why put my camel two columns away from his elephant when I can instead put it three columns away? Therefore, after my setup, we will be following at most one previous game, at least as far as I am aware.
Why shouldn’t I reflexively play my pet opening? I expect it would lead to a dynamic, unbalanced game, which I would much rather have than one in which I copy the Mob with a 180-degree rotation, effectively conceding disadvantage to the initiative of the Mob’s first move. That said, though, this game is a perfect opportunity for me to question I disagree
umptions. Playing automatically under time pressure is necessary, but now I have days to consider, and I should make every effort to refrain from unthinking moves.
So, what about opening with a decentralized elephant? But why? One idea would be to plan an attack on the Mob’s flank rabbit, but if I did that I would only prove my hypothesis that in a fluid position there are more important things to do than pull rabbits. I am more tempted to copy blue22 and put my elephant on c7 for an immediate elephant-horse attack on c3. However, after playing a few opening lines, I find it rather awkward for me if the Mob responds with elephant to c5, when I need to move my elephant back to the center anyway to get it into the game. Furthermore, blue22's setup almost requires one to re-occupy c7 after the elephant leaves it, costing one tempo that I would by trying to save. Therefore my elephant goes on e7 as usual, facing the Mob's camel, the fattest target it can have in any case.
My camel goes on g7 as explained above. A minor additional consideration is that the Mob's setup is a bit more in tune with advancing its eastern horse than its western one, and my eastern camel stops that. With my camel on g7, I quite like rabbits flanking it on h7 and f7. I am not worried about these rabbits being pulled, since I often want them advanced in any case for the safety of my camel, and/or to help a swarm if I give up my camel hostage. I realize, though, that one of my stated purposes in having rabbits behind my traps, namely that they allow me to immediately race in the opening when I am Silver, no longer applies with my camel on the flank. If the Mob takes my f7 rabbit, I take a cat, they take my camel, and I take a horse, I have lost materially. So a direct race is out. I guess I don't mind that, because the Mob can decline the tension represented by a race, but they can't decline the tension represented by my flank camel.
Since I am not immediately racing, I don't need a rabbit on c7. In fact, I find that I am much less likely to advance rabbits along with horses than I am to advance them along with my camel. Also a c7-rabbit sometimes makes it harder to frame a horse in the c6-trap, because I can't allow my rabbit to be pulled into the trap where it is in the way. So I think I will put a piece on c7 this game. Probably it should be a cat, because if I do get a horse frame, I don't want to waste a dog holding it in.
Maybe I shouldn't even have an a7-rabbit, since I have no plans to advance rabbits on that side. I often find that when I set up my horses on b7 and d7, I'm eager to move the d7-horse to the b-file or even the a-file, so maybe I should start with horses on a7 and b7. Playing with that setup a bit, though, uncovers two reasons to be suspicious of it. First, if the Mob opens elephant forward four when I have a dog on d7, it is rather awkward because I don't want my dog flipped out. With my horse on d7 instead, I worry less because I don't mind my horse being flipped out. But the bigger reason for not having two flank horses is that it seems too committed to the single plan of getting a horse frame in c6. In some sharp opening lines my elephant doesn't quite have time to get any attack started in the west before being forced to play in the east, in which case the centralized horse can come in handy.
How about a dog on a7? This would entail a rabbit on d8. I'm not so much worried about that central rabbit being pulled out in a way that endangers the rabbit itself, but again I recall horse-framing tactics where I needed to occupy d7 with a piece, not with a rabbit, because the defending elephant could flip whatever is on d7 into c6. So I guess I'll stick with an a7-rabbit in my setup, not because I have any use for the rabbit there, but because my dog in the center may have a job.
One final consideration is whether to have a dog or a cat behind my camel. I am not at all eager to have my cat enter the game; cats belong at home defending against rabbits. The question, then, is whether my cat would more likely be forced into service from e8 or from g8. I don't have enough experience to know, but my hunch is that since my camel will staying at home like a coward at first, and will have friendly rabbits to help it advance when it does, my cat behind the camel will be safe and will not be called into action. Meanwhile the piece on e8 will have a lane forward immediately, because I will advance my elephant, and in some lines might need to help out at e6, so at first blush it looks more like a dog's job.
I think I have decided on my setup. However, I have over four days left on this move and my reserve is full, so I don't benefit from moving now. I think I will instead try to anticipate the Mob's next move assuming I set up as planned, in case I see something that changes my mind.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
2g:
I note that the position after our two setups has never occurred in the history of Arimaa, so we are already in uncharted waters. That should remind me to stay on my toes and question my intuitions.
The opening move that dominated play in 2004 was elephant forward four. Here is seems lame, as it only threatens to pull out a horse that I want to activate anyway. I could play very conservatively, answering 2g Ed2n Ed3n Ed4n Ed5n with 2s ee7s de8s hb7s ra7e, asking the Mob whether it has any plan. I expect the Mob's elephant on d6 would have to retreat with loss of time sooner or later.
The favorite move of JDB, namely 2g Ed2n Ed3n Hb2n Hg2, would be solid and flexible if I had set up with a central camel, but here it seems like a waste of time for the Mob to advance the eastern horse. The horse on g3 is not likely to attack into the teeth of my waiting camel, and in some lines will actually be a target for my camel on g6. Nor does the horse on g3 make the f3 trap any safer against an elephant-horse attack from me; I have no horse on that side!
More plausible is the favorite move of chessandgo (and many others), namely 2g Ed2n Ed3n Ed4n Hb2n, threatening a western elephant-horse attack. However, the Mob collectively surely knows that my general strategy with my decentralized camel is to try to take a horse hostage with my elephant on the side where I have two horses, and make credible frame threats based on having an extra horse in the region. Failing a frame, I might keep the horse threatened on b5 while my camel can play actively on the other wing because the Mob's elephant is tied to defense of its hostage horse. Given that I want to nab the Mob's western horse, will they seriously plan to advance it voluntarily?
On the other hand, the Mob will eventually need a horse on b3 to defend against my elephant horse attack on the c3-trap, so the move doesn't waste time. Also it doesn't commit the Gold elephant to either flank. It seems a way to be aggressive while maintaining flexibility. I might answer 2g Ed2n Ed3n Ed4n Hb2n with 2s ee7s ee6s hb7s de8s, cutting off the option of the Mob's elephant attacking my camel while my elephant still very much faces the Mob's camel. The gold elephant could dive into e7, but the attack seems too premature to be dangerous. Basically the Mob would have to continue with western action, at which point we would just see whether I am right that I have more to gain than to lose.
An idea that The_Jeh tried against my setup seems more critical. 2g Ed2n Ed3n Ed4n Ed5e indirectly targets my camel while cutting off my elephant from going after the Mob's camel. The success of my setup is predicated on Gold not having the timing to attack my eastern flank with his elephant, not in the opening, and not later in the game. I can hardly defend the east with my own elephant and expect to gain advantage, so if at any point the Mob chooses to attack the east with its elephant, my plan is to give up only a rabbit in the east while getting something more valuable in the west or center. So I can't exactly hang back against this opening move. I might try 2s ee7s de8s hb7s hd7s, planning to answer 3g Ee5e Ef5e Eg5n xxxx with the solid 3s ee6s de7s rf7w mg7w. I don't think the Mob gets anything by jumping into f7 with its elephant in that line; I can choose to force the elephant out either to the wing or to the center, my choice, and it must come out next turn, perhaps at a loss of time to avoid being smothered. Therefore the Mob wouldn't have any better play than pulling my h2-rabbit to h3 (where I would eventually advance it voluntarily), which gives me time to get rolling in the east. Surely the Mob wouldn't move its elephant right back to the center on move 4g, for a net loss of time.
Of course, the Mob wouldn't need to attack my camel on move 3g. It could instead play against my western horses with its elephant and/or shift its camel west. This type of move, however, will at least give me time to activate my elephant in the west, and will tend to make the west the focus of action, which should slightly favor me strategically if I haven't given up anything else. If play is in the west, the centralized gold camel will probably have to pick a wing and move there, whereas my camel will already be on the wing where it wants to be.
Playing through a few lines is persuading me that 2g Ed2n Ed3n Ed4n Ed5e is the only move by the Mob that doesn't already concede a slight advantage to my setup, so I fully expect this move, and won't be analyzing anything else. If they play this strong move I'm not sure who is winning, as the opening can get sharp in a hurry, and many terminal positions in my analysis are unclear, but I don't feel that the Mob having the first move necessarily puts me on my heels. My setup was a latent threat, and they have to act quickly and precisely to avoid drifting onto the defense right out of the gate.
…
After having written the above, I played a live game as Silver against chessandgo in the middle of the Mob's deliberations. Chessandgo responded to my unbalanced setup with the expected move of elephant up three and over one, whereupon I played the highly speculative move of advancing my flank horse two squares, along with advancing my central elephant and dog one each. Chessandgo didn't take the bait of attacking my exposed horse; without hesitation he moved his elephant to attack my flank camel.
I emerged from that opening slightly worse, but I was correct that Gold had nothing better to do than to pull the rabbit on my camel flank on move five. My first mistake of that game was my own fifth move: I failed to get a horse hostage with my elephant as quickly as possible. I have no need to deviate before that; in fact a superficial analysis leads me to believe I will be a tiny bit ahead if the Mob Game follows the opening of my live game up to that point.
I wonder whether that one live game prolonged the deliberations of the Mob, because I noticed a flurry of posts afterwards, and the Mob has gone almost seven days without moving. I hope they dip significantly into reserve, because they can't do better than elephant up three and over one, which I know from chessandgo's play that they were already considering. At worst I have tipped my hand and given them more time to analyze my counter that they would not otherwise have considered, but I am by no means certain I will play so speculatively in a postal game anyway, so I am as likely to have sent them on a wild goose chase as to have helped them. It is too much to expect that my live play will help lead the Mob into error, but I can at least hope that the Mob will start a slow descent into time pressure.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
2s:
Immediately after having written the above, I discovered both my hope and my fear proved false. The Mob moved a couple hours shy of a week, thus preserving their full reserve, but also they didn't play what I consider to be their best move. I wonder why not. Specifically I wonder what move chessandgo was advocating, since he played the critical elephant advance in our live game, but he normally plays 2g Ed2n Ed3n Ed4n Hb2n, exactly as the Mob has just done. Is it possible that the live game, although chessadngo won it, swayed the Mob away from the move 2g he used?
I haven't looked at any response other than 2s ee7s ee6s hb7s de8s, and that move only superficially, but I don't mind losing most of my previous analysis in exchange for having a slightly better situation on the board. For me to do extra analysis is no problem; time is definitely on the side of the One in these games, because for the lonely player there is no communication overhead. I shall answer Ron Weasley's chat with a taunt, and then analyze at my leisure.
…
After studying the position a bit more, I continue to like 2s ee7s ee6s hb7s de8. Every step has an important function. The elephant steps not only centralize my strongest piece, they also block the Mob's elephant from harassing my camel on move 3g. The horse step prevents the Mob from launching an effective elephant-horse attack on move 3g. The dog step is the least critical of the four, but it shores up my camel-side without conceding any weakness, and I can't think of any step that is more important.
The step ra7e would anticipate a horse frame and also shield that rabbit from being pulled by the Mob's horse, but either a horse frame by me or a horse charge by the Mob is more remote than a potential attack by the Mob's elephant against my f6-trap. Indeed, in some lines the Mob is threatening to pull my westmost rabbit, not with its horse, but with its elephant, in which case my rabbit is safer on a7 than tucked into b7, so moving it to b7 now could be a double waste of time. Definitely the rabbit tuck is low priority at the moment.
I could advance my elephant a third step, but really I am not prepared to race. Not only would a further advance re-open a lane for the Mob's elephant to attack my camel, it would leave my elephant further from being able to help against a central elephant charge by the Mob. I like my position strategically, but I'm not yet ready for a tactical melee. By the same token, advancing my western horse two squares instead of one is consistent with my strategic goals, but seems tactically unsound as it would make it just too easy for my horse to be taken hostage.
Given that my preconceived move looks strategically coherent, I'm only going to abandon it if there is some tactical flaw. The most forcing try by the Mob seems to be an elephant invasion with 3g Ed5n Ed6e de7n Ee6n. The Mob's threat is to push my f7-rabbit into f6, but my resource is to threaten to smother their elephant. After 3s mg7s rf7e ee5n ra7e, the Mob would have no time for any rabbit pull, and would have to push out immediately with something like 4g hd7s Ee7w hd6s Ed7s. But then 4s ee6s hd5w hc5n ee5s leaves me simply ahead on time, clearly more than equal despite moving second. The Mob would hardly have a better plan than attacking my exposed camel after all, but 5g Ed6s Ed5e Ee5e Ef5e seems well met by 5s rh7s rh8s rg7w mg6n, when I am even further along in my plans.
A second response for the Mob would be a horse advance on my non-camel wing. Sometimes such an advance carries the threat of taking control of an enemy trap, but here I have two defending horses, so an attacking horse would be more likely to end up framed than secure on a good square. With an elephant-horse attack probably not in the cards, the purpose of the Mob's horse advance would be to pull a rabbit. This is a serious issue, because I only want advanced rabbits on my camel wing. On the other hand, the horse advance also leaves the Mob exposed to counter-attack. After 3g Hb3n Hb4w Ha4n Ha5n it looks to me like 3s ra7e hd7s ee5s ee4s is such a strong threat that the Mob will have not time to pull a rabbit. The only move that immediately pulls a rabbit and keeps everything safe is 4g Ha6n Ha7s ra8s Dd1n, but the exposed rabbit is more than compensated by my taking control of the c3-trap with 4s ee3w hb6s hb5s hb4s. It appears that I get in the first shot so the Mob would not be able to afford dueling elephant-horse attacks, and I would simply have a strategic edge.
The third main option seems to be playing for a horse hostage with 3g Ed5w Ec5w Eb5e hb6s. At first I dismissed this threat entirely, because an elephant and two horses should win a fight against an elephant and one horse, but the Mob can also shift its camel over to the west. In that case the Mob's elephant, camel, and horse can be expected to win against my elephant and two horses. My possible salvation would lie in having a camel as the sheriff of the other wing. There would be a race in which the Mob is trying to pass off a hostage horse from its elephant to its camel, while my lone camel is trying to get as much compensation as fast as it can on the other wing.
First I would get my horses into their ideal position with 3s hd7s hd6w hc6w hb5w. If the Mob is over-eager about getting a horse hostage with 4g Ec5w ha5s Eb5w Me2w, then my elephant-camel(!) attack 4s ee5s mg7s mg6s mg5s is tactically difficult to meet. Better for the Mob would be the patient 4g Ec5w Me2w Cc2w Md2w, after which 4s ee5s ee4w ed4w ec4w leads into a very double-edged fight. I have a legitimate threat to flip the gold horse into a4, so the Mob can't just shuffle its camel around. If instead the Mob pushes my horse only to a4, my horse could pull a rabbit to a3, much delaying a handoff. Therefore it seems the Mob must push my horse all the way to a3 with 5g ha5s Eb5w ha4s Ea5s. Then I must act fast with something like 5s mg7s mg6s mg5s mg4s, launching a type of race that is characteristic of my unbalanced setup.
I expect to have a general advantage in these races, because it costs my camel fewer steps to move up and back than it costs the Mob's elephant to move right and left. When the Mob's elephant switches sides, it needs to follow up the switch with elephant action on my camel wing, to avoid simply losing time shuffling. So I need to be careful that I'm never offering to give up as much to a sustained eastern assault by the Mob's elephant as I expect to gain with my unopposed western elephant. On the other hand, I have to be reckless enough in the east with my camel that I can compensate for an eastern battle that I will be slowly losing. My entire unbalanced strategy hinges on being able to be both cautious enough with my camel and being able to get enough compensation on the non-elephant wing.
Will my racing strategy work? I am certainly not so scared of this race that I will leave my horse cowering on b7, opening b6 for the Mob's horse. Even if I were truly afraid of my horse being pulled, the alternative is worse. But I'm not afraid: I like my racing chances a priori. I recognize that this sort of thing must be determined by practice rather than theory, but the theory is at least as good for me as it is for the Mob.
Since I have decided on my move early, and the variations are not so sharp that I have to triple-check them, it wouldn't hurt me to move in fewer than seven days. On the other hand, I'm traveling a bit between Christmas and New Year, so it will actually work out better if I move just before my reserve runs out on Christmas Eve, let the Mob deliberate while I am on vacation, and return to the game after the holidays.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
3g:
I'm oh-for-two so far at predicting Mob moves, and the current move doesn't promise to improve my record. The possibility that worries me most from my analysis is 3g Ed5w Ec5w Eb5e hb6s, but the lines that are good for the Mob aren't obviously good for the Mob, and I don't expect the Mob to grab a horse on nothing but general principles. Taking a horse hostage with an elephant seems to be widely out of favor these days, perhaps because the most reliable bot-bashing method is to give away a horse hostage and then swarm the enemy's hostage-holding elephant. Of course, just because bots have no idea what to do with a horse hostage doesn't mean the hostage is worthless, but the general perception of worthlessness remains, and will probably rule out this move getting majority support.
I recall from being on the Mob that there was always enthusiasm for the most active, most forcing move, whether or not it was the best. Charging with 3g Ed5n Ed6e de7n Ee6n certainly meets this definition. But unfortunately for me, the Mob is also excellent at refutations, and they will hardly overlook the possibility that I can threaten a smother. I expect their discussion will label the move as interesting but refuted.
The immediate horse charge 3g Hb3n Hb4w Ha4n Ha5n is in keeping with the standard operating procedure of trying to get a rabbit pull, so it might be favored for that reason. A majority of people like to pull rabbits. Again, though, I have a near-refutation tactically. The Mob must consider the possibility that I will hit back rather than defending, and they shouldn't like what they see in those lines. The only benefit the Mob could get from playing this move is adding to my false sense of confidence.
What about moving the elephant to b6 with either 3g Ed5w Ec5w hb6e Eb5n or 3g Ed5w Ec5w hb6w Eb5n? To me it looks a bit dodgy for the Mob to decentralize its elephant so soon. I expect I could meet either move just fine with 3s hd7s ee5s ee4s mg7s, with a threat to either flip out the Mob's camel or to occupy g3 with my camel while the Mob's elephant is on vacation.
Probably the Mob will play some hybrid move instead, i.e. perhaps they will not devote all four steps to the same purpose. Their elephant might well go to b5, but wait there until next turn to decide whether to pull my horse or my rabbit. That would leave two steps for general development, albeit non-obvious steps. Normally putting a horse on g3 would be automatic, but in this position that creates a target for my camel, so it is less clear. Advancing the camel is another generally useful move, but advancing it directly toward my elephant would be unwise. Perhaps instead the Mob camel could shift west to get ready to receive my horses in a hostage handoff; I guess that makes sense even if the east is left weaker. But that shift makes the Mob's rabbit on e1 look rather foolish: rotating the horse forward to g3 becomes almost mandatory so that the g1 dog can move up too and let the central rabbits over.
I will therefore predict 3g Ed5w Ec5w Me2w Hg2n, although I have little chance of being right because the last two steps are malleable. The threat to flip my horse is non-trivial. On the other hand, I have what I wanted from the opening. The Mob has already let me cut off the sharp plan of attacking my camel, and I just don’t believe that pulling my rabbits is fast enough, so one way or another they will play against my western horses. This will tend to tie up both of our elephants in west and free my camel for action in the east, which will directly test my hypothesis that this general situation gives me an advantage.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
3s
The Mob played 3g Hb3n Hb4n Hb5w Me2w, confounding my predictions once again. Superficially the Mob's choice has the same weakness that 3g Hb3n Hb4n Hb5w Ha5n would have had, in that my counter attack with, say, 3s ee5s ee4s ee3w hd7s poses a triple threat of taking a cat, flipping out the gold camel, or launching an elephant-horse attack against c3. Unfortunately, it looks as though 4g Cc2w Cb2n Md2w Hg2n adequately stops the first two threats and prepares to answer the elephant-horse attack 4s hb6s hb5s Cb3s hb4s with 5g Ra2n Cb2w Mc2w xxxx. It never occurred to me that the Mob might be able to fight off my elephant horse attack without bringing home its elephant. Curses! I must analyze to see whether I can salvage an acceptable counter-attack, or must instead immediately play against the Mob's advanced horse with my elephant, probably allowing a rabbit pull in the mean time.
…
I have given some thought to 3s ee5s ee4w ed4w ec4w, attempting to play directly against the Mob's advanced horse. Although I would get some strategic pull, after 4g Ha5n Ha6s ra7s Md2n, it is truly annoying for me to have a rabbit on a6, because I need that square for my horse-framing tactics. This annoyance is so great that it might even counterbalance my other advantages, so I will play another move if I can find one that seems both good and clear.
Further analysis apparently confirms that I can't make the elephant-horse attack work following 3s ee5s ee4s ee3w hd7s 4g Cc2w Cb2n Md2w Hg2n, but I have a glimmer of hope for the offensive play: If instead I change my fourth step to 3s ee5s ee4s ee3w mg7s, then the Mob's otherwise lovely defensive move 4g Cc2w Cb2n Md2w Hg2n can be well met with 4s ed3n ed4w ec4n ec5w, taking the Mob's horse hostage before it can disrupt my western rabbits. I believe that in that position I would have the whip hand, because the Mob would have no clear plans, and I could persecute the hostage horse at my leisure, playing to eventually frame it in c6, or at least tie down the Mob's elephant and camel while my own camel operates in the east. That would be precisely the strategic situation my opening setup was angling for. (Indeed, if my elephant could fly directly to b5 on the present move I would gladly do it without an intermediate threat.)
Of course, the Mob could choose other means to defend itself from 3s ee5s ee4s ee3w mg7s in a way that does not leave its horse hung out to dry on a5. But the sharp 4g Ed5e Ee5e Ef5e Re1n, attempting to threaten my camel, is left flatfooted by 4s rh7s rh8s mg6n hd7s. I don't think I need to fear that the Mob will switch wings with its elephant just yet. And trying to pull my rabbit anyway with 4g Ha5n Ha6s ra7s Re1n just seems like a way for the Mob to enter a rabbit-pulling race at a disadvantage after 4s ed3e ee3w Re2n xxxx. (Incidentally, that last variation demonstrates one disadvantage of the Mob's choice to switch its e1-dog and g1-rabbit relative to the 99of9 setup; maybe in truth the chessandgo setup is slightly technically inferior to the 99of9 setup instead of the other way around.) I am provisionally a fan of the straightforward counter-attack for 3s, and will examine further possible responses by the Mob.
…
My preferred attacking move seems to be holding up under analysis, although there are too many possible replies to list. Generally speaking, the Mob's western horse by itself can't do anything other than pull a rabbit. If the Mob doesn't spend all four steps on the pretty defense, because they want to use at least one step advancing their western horse, my elephant can make a better rabbit pull or the equivalent. If the Mob wants to make a bigger threat than pulling a rabbit, though, it will need to use its elephant to dislodge my b6-horse, and if the Mob moves its elephant out of the center my threatened camel flip becomes very serious.
I am quite fond of using my fourth step to advance my camel, because the Mob may want to respond to my elephant advance by putting its eastern horse on g3, and my camel would be a greater threat to it on g6 than on g7. Nevertheless, I should give due diligence to using my fourth step to tuck my a7-rabbit into b7. If it turns out that the Mob's best answer to my elephant charge is to accept a rabbit-pulling race, then I must at least consider a move that would leave me better off in such a race.
…
The move 3s ee5s ee4s ee3w ra7e is very tempting because it gains me time if the game turns into a rabbit-pulling race. On the other hand, if I end up taking the Mob's horse hostage with my elephant, the rabbit tuck ends up losing me time. If the Mob calculates along the same lines that I do, then they might respond to my camel advance by initiating a rabbit-pulling race, but respond to my rabbit tuck by letting me take their horse hostage.
I analyzed for some time under the assumption that my fourth step would cause a divergence in the Mob's response into two very different types of games. I wanted to see which type of game I would like better, and ultimately concluded that the rabbit-pulling race will favor me even if I don't tuck my rabbit. My centralized elephant is making incidental threats that will constrain the Mob's options, so it looks like I will be able to capture the Mob's central rabbit while returning with my elephant in time save my flank rabbit, if not permanently, then at least long enough to gain time in the long run.
Therefore my logic has doubled back on itself to the point that I don't think the Mob will start a rabbit-pulling race in response to either of my two options, and will let me take its horse hostage with my elephant in either case. According to that conclusion, 3s ee5s ee4s ee3w mg7s is clearly the better move, because in the hostage-related maneuvering I will have the b7 square available to rotate my pieces through if necessary.
My move preference has clarified, but of course the value of my move depends entirely on whether I am correct to believe that taking the Mob's horse hostage with my elephant is a good strategy. I obviously can't calculate concretely that far ahead. I'm excited, though, that the opening may hinge on exactly the strategic dispute I anticipated when I set up my pieces with a decentralized camel.
…
I entered my move this afternoon, and was about to send it, when suddenly 3s mg7s ee5s ee4s ee3s looked attractive. Why not directly go after the Mob's exposed e1 rabbit? I closed out the game window to give it more thought. Unfortunately, after a closer look, I am at best equal in the rabbit-pulling race after 4g Hg2n Cc2w Cb2n Md2w 4s ee2n Re1n ee3w Re2n 5g Ha5n Ha6s ra7s Ed5e, and if I belatedly abandon the rabbit to chase the Mob's horse with 5s ed3n ed4n ed5w ec5w, then 6g Ee5e Ef5e Eg5e mg6s seems to leave me worse off. I guess I will play 3s ee5s ee4s ee3w mg7s while I am only a couple of hours deep into my reserve.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
4g
The Mob dipped a couple of days into reserve on its last move. They have lots of ways to defend my two threatened cat captures or counter-attack, so I do not expect them to be able to regain time. From my experience as a Mob member, only moves that avoided a preferential vote in favor of a consensus vote were able to add to the reserve, and this doesn't look like a situation for consensus. On the contrary, I somewhat expect them to eat further into reserve.
My inclination heading in to the game was to avoid complications because the Mob will play better in sharp positions, but now it occurs to me that seeking complications might also be my best way to put time pressure on the Mob. I should generally have a time advantage over the Mob because I don't have to communicate or vote, which tempts me to try to create time pressure for the Mob and exploit it.
It is not clear to me whether I should just play the board, or play the board plus the opponent, or play the board plus the opponent plus the clock. Obviously the position is the most important, most tangible factor. I should be able to win a won position regardless of the opponent. On the other hand, wouldn't it be foolish to play into the teeth of the Mob's strength? If the objectively best move leads to a type of position the Mob will play well, then I had better have worked it out extra carefully that my move is good. And I know from postal experience that I can look at a position for a long, long time without finding the best move, which undermines my confidence that I can ever justify ignoring the opponent by the argument that my preferred move is "objectively best". I think I must factor in the Mob's strengths at least to some degree.
As for the clock, though, my calmer thought is that I will just get myself into trouble if I pay too much attention to the Mob's time situation. For the present, the one concession I will make to trying to create time pressure is this: if the Mob plays 4g Hg2n Cc2w Cb2n Md2w as I am predicting, then I will immediately respond by taking their horse hostage. I have looked the position over once again, and I really don't have any other plays. Either the resulting elephant-holding-horse-hostage position is good for me or it is bad for me, but I will have to go for it. On any other Mob move, though, I will forget about the Mob's clock and take it slowly.
One possibility for the Mob that I didn't examine before I sent my move is 4g Ed5e Ee5e Ef5e Hg2n, attacking my camel and offering the c2-cat as a sacrifice. For me to take the cat with my elephant doesn't appear to give me sufficient compensation for the camel hostage I would be giving up. The super-sharp 4s hd7s hb6s hb5s hb4s appears more promising, but I can't quite see my way clear through the tactics after the Mob flips my camel on 5g. I am afraid that I would get the worst of it in that line as well. That leaves me only the staid 4s rh7s rh8s mg6n hd7s. Fortunately my boring counter-move looks strong. The Mob wouldn't be able to consolidate and continue to play against my camel with 5g Cc2w Cb2n Md2w Eg5n, due to a powerful elephant-horse attack with 5s hb6s hb5s Cb3s hb4s 6g Ra2n Cb2w Mc2w Ha5e 6s ed3s ed2w hd6s hd5s. So I'll put the elephant switch in the category of "hope for but don't expect".
I was worried for a bit that the Mob might be able to avoid the horse hostage (and avoid any other disadvantage) with a scattering move like 4g Md2e Cc2w Ha5n Ed5w. But all the versions of that move that I have tried for the Mob let me get some compensation by pushing out the camel with 4s ed3s Me2n ed2e hd7s or something similar. It turns out to be important that my elephant is near the Mob's camel and the Mob's elephant is not near my camel. This development is gratifying because that was one of the justifications for my opening setup.
The Mob can obviously defend the c3 trap by putting a rabbit on b3, but I have given such moves hardly any consideration at all. There would potentially be some benefit to the Mob in a western rabbit advance, namely that it would make it safer for their camel to operate in the west, which in turn might make it a bigger threat for them to pull out one of my two western horses with their elephant. This would combine with the threat of a western swarming attack. Realistically, though, the Mob will not be able to stomach trailing in a rabbit-pulling race, even less so if chessandgo is calling the shots. It ain't gonna happen, so I ain't gonna worry about it.
Although the Mob's move 3g was a strong move I had not considered, as I await 4g I am nearly as optimistic as I was then. The critical move, i.e. the move that I think the evaluation of the position ultimately depends on (4g Hg2n Cc2w Cb2n Md2w), is the move I expect. If the Mob plays any of the other moves I have considered, it should give me a small, clear advantage, as opposed to a small, unclear advantage.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
4s
I was right that the Mob wouldn't move quickly. They needed eleven and a half days, putting a noticeable dent into their reserve, now down to two weeks from the original three. Once again, however, they made a move I did not even consider. Their choice of 4g Cf2w Rf1n Ha5n Ed5w has advantages: it doesn't expose a rabbit on e2 for me to pull; it doesn't allow the Mob's advanced horse to be taken hostage; it intensifies the threat to pull my h7-rabbit; it adds the threats of pulling out my b6-horse and taking over the c6-trap; it prevents me from an immediate elephant-horse attack against the c3-trap; and of course it prevents me from making an immediate capture in one of the Mob's home traps.
On the other hand, my immediate reaction is that 4s ed3e ee3w Ce2n hd7s, defending at home while again threatening both of the Mob's cats, should be strong for me. Yes, the cat, unlike a rabbit, can retreat, but how? That seems to require three steps, leaving the Mob no time for a rabbit pull or a horse pull. And if the cat doesn't retreat, getting it offside could easily compensate me for a pulled rabbit. My second horse prevents a quick takeover of the c6-trap, so it seems the Mob's main plan would be to try to take a horse hostage and ferry it back to their waiting camel, with the fourth step Re1n as the only way to prevent a cat capture using only one step. If, however, I can get a cat capture while the Mob is fussing with my horse, it could work out fine for me, depending on the position. I'll have to play out some lines to know.
Also I have to remember to look at moves other than my first instinct. The Mob's elephant going even one square west gives my camel that much more breathing room in the east, and slightly increases my ability to threaten their centralized camel with my elephant. Maybe there is something there. The only thing certain on first analysis is that I won't be making a snap reply.
…
A bit more examination leaves me with a somewhat murky impression of the position. Playing defense with 4s ed3n ed4w ec4w eb4n seems reasonable, but not nearly as favorable as lines in which I could take the Mob's advanced horse hostage without having my rabbit pulled to a6. This will be my backup move if my instinctive choice doesn't pan out. The camel pull 4s ed3n Md2n ed4w Md3n seems to be refuted by 5g Ec5e Md4s Ha6s ra7s, although I could look at continuing with 5s ec4n ec5w eb5e Ha5e if nothing else panned out. It seems rather similar to playing defense immediately, but there may be a subtle distinction that I am missing.
Mostly I have looked at my favorite move, which still seems critical. It seems the critical lines involve my winning a cat in exchange for a camel holding a horse hostage. I am saddened to discover that some of the lines work out badly for me, but others work out well. For the first time in the game I can't keep the most important lines straight in my head, and I need to create an analysis tree.
4s ed3e ee3w Ce2n hd7s
5g Hg2n Hg3w Ce3s Dg1n
5s dd8s hd6s dd7s hd5s
6g Ec5s Hf3e Ha6s ra7s
6s mg6s mg5s mg4n Hg3n
7g hd4e Ec4e he4s Ed4e
7s mg5n Hg4n mg6e Hg5n
6g Cc2w Cb2n Md2w Hf3e
6s hd4e he4s Ce2w he3s
5g Ec5w Eb5e hb6s Re1n
5s Ce3n ed3e de7s de6s
6g hb5s Ec5w Md2n Dd1n
6s de5w Ce4n Md3n ee3w
5g Rf2w Hg2w Ha6s ra7s
5s Ce3n ed3e Ce4n ee3n =+
…
I followed the Mob game in my World Championship Preliminary game against Adanac, because I believe that I am winning, or at least creating more chances than Silver normally gets. I feel that it is up to Gold to deviate or prove that my opening is unsound. Unfortunately, that game revealed my planned 4s and 5s from analysis. This increases the pressure to be right if I actually go down this line, because the Mob will be better prepared to respond. Still, I don't regret the exposure, because it got me a good position and appeared to fluster Adanac. To do well in the Mob game is prestige, but to do well in the World Championship is money.
…
I have dipped a few days into reserve myself this turn. The primary culprit is spending my Arimaa time preparing my book. One result of that interlude will be that I will switch to short notation for this commentary just as I have switched in my book, because now that I have gotten used to the paradigm, it is easier to read and write as well as being shorter.
I looked at the position with fresh eyes, and for a moment was attracted to 4s ME^HdvRa7>, but a bit of examination convinced me that the response 5g Dd^M<<E< leaves me with nothing. Indeed, pulling out the gold camel now would merely help get it to a better square. In this game it may turn out that I never get to launch my western EH attack, but even so the threat of it truly constrains the Mob's options. My present thought is that the Mob camel is misplaced in the center, and that the relatively better location of my camel is the reason that I am doing well in the opening.
Not only could I find no move better than 4s E>CE<Hdv, I couldn't poke any holes in it upon further analysis. I know that the Mob must be expecting this move in light of my game against Adanac, but I can't find a satisfactory response for them. Obviously I can't look at every line, but in general the lines I play out seem to be working out well for me. The Mob's main threat of pulling out my b-horse and passing it off to their camel is a little bit too slow given the position of their camel.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
5g
« Reply #10 on: Jun 1st, 2010, 6:35am » |
Quote Modify
|
5g
I don't think the Mob will have the guts to play 5g E<HE>Re^ as Adanac did against me in the World Championship preliminaries. In part I think that my response from that game, 5s EC^Devv is actually dangerous to them and they will want to avoid it. But also they know that I know that they know what I played against Adanac. I had a long think, and followed that public game anyway, so I must have calculated that I am winning that line, right? I expect this consideration to intimidate the Mob away from Adanac's continuation.
I'm sure the Mob would like to use two steps to pull my rabbit to a6 with their horse. The trouble with that plan is that only two steps are left for home defense, which means they would have to allow me to get their cat offside or at least pull out a rabbit. My central rabbit pull should be at least as valuable as their flank rabbit pull. I don't expect any move from the Mob which concedes that they have lost the opening, even by a little bit; instead I expect them to try to preserve tension and maintain hope that they have won the opening.
The Mob could use three steps on defense and one to put their elephant on b5, for example 5g Hg^<CevE<. They may be attracted to this plan because their major threats to flip out my b6-horse and/or take over the c6-trap are intensified by the elephant step. If they are paying attention, though, I think they will find it is too tactically dangerous to decentralize their elephant another square, bringing to life my central horse and flank camel.
The normal-looking waiting move 5g Hg^<CevH> leaves the horse directly opposite my camel, which I think gains time for me after the obvious pull 5s MvvHM^. Therefore I will predict the awkward-looking developing move 5g Hg^<CevDg^. It's not pretty, but it also is not clear how I could take advantage of it. It fits my opinion of the Mob that they would focus on tactical shots and not mind having pieces on weird-looking squares. The Mob's hope would be that they have conceded no disadvantage and will get at least a rabbit pull after my elephant gives up on the center and plays 5s E^<<^. Before I would respond in that way, however, I would give a long, hard look at 5s HdvDdvvHv, threatening both an immediate capture in the c3-trap, and threatening to swing my central horse around to g3, initiating a hyper-aggressive double-trap attack strategy.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
5s
« Reply #11 on: Jun 1st, 2010, 6:36am » |
Quote Modify
|
5s
The Mob played 5g Hg^<CevDg^. At last I have predicted a Mob move, so I know it is possible! Now that it has happened, though, I wish it hadn't. I predicted that move because it looks strong; I would have rather been wrong and faced a weak move. My lone elephant has no more targets to attack, so I must either play defense now, or I must raise the stakes and add a horse or camel to my attack.
What I should have expected but didn't was that my choice of a predictable move allowed the Mob to regain reserve time. Their discussion this time had only a few posts and lasted less than five days, so I infer they did not have a formal vote, and merely agreed on a move that had been suggested in the previous move's discussion.
Unless I miss my guess, the Mob will not be expecting 5s HdvDdvvHv at all, so if I play this move they will be forced to dip into reserve time again to respond to it. On the other hand, I must not get trapped into playing the clock instead of playing the board. I see that my analysis tree from 4s didn't include the possible responses 6g EvEH>Hf> and 6g EvEH>Hf<, either of which may refute my preferred move. If I can't make the horse charge work, I'll have to consider various camel advances and also the boring 5s E^<<Ra7>.
…
I looked for a bit at 5s MvvvRh7v, and I liked it just fine if the Mob's elephant immediately crosses, because I intended to run away. 6g E>>>v 6s R6vvM^^ leaves me in no danger and with a noticeable time gain. Unfortunately, the Mob wouldn't have to cross immediately, and could instead inch a step closer with 6g Rh1<EvRHhv, blocking my capture threat and threatening to continue with 7g E>>>>. My camel would then have to retreat empty-handed to avoid being taken hostage.
My preferred move 5s HdvDdvvHv seems to be holding up under greater scrutiny. If the Mob pushes my horse to e4, then 6s DevvHe>> seems to work in most lines. In a role-reversal, the Mob would be giving up its elephant-horse attack on the wing where my camel isn't in order to defend against my elephant-horse attack on the wing where their camel isn't.
And what choice would the Mob have? I believe they can't race my EH attack with an EH attack of their own. I have a horse at home on b6, which drastically slows their attack down, whereas they don't have a horse (or camel) at home on b3, which speeds my attack up. Therefore their elephant is strategically obliged to play defense against my advanced horse.
The Mob could play for a "safe" rabbit pull with 6g RHhvEvHf>, but that allows me to start a rollicking fun double-trap attack with 6s DdvDevvv, followed by a camel advance against g3. The Mob's elephant doesn't seem to have time to stop both that and a potential EH attack against c3.
Indeed, there seems to be a general theme in the position that both of the Mob's home traps are weak, and their elephant can only defend one. I don't see how the Mob's elephant can pursue a horse hostage in the east before shoring up the west which they have weakened with a horse advance. There are too many lines for me to keep straight, but I cling to the reassurance that the Mob's camel is not well-placed for a horse hostage handoff.
One creative attempt at tactically refuting my horse advance would be 6g EvHE^, aiming for 6s E<CE>* 7g EHvM<<, sacrificing a cat to get a horse held hostage by a camel. Strategically their move is motivated by wanting to keep my horses on the same wing as the Mob's camel. I need to look at this tactical line more, because it is so far the line that is most making me question the soundness of my advance.
…
After further analysis of my horse charge, I have to admit that I don't really know what to think. I could never get the lines to play out the same way twice, particularly when the Mob cedes a cat capture to gain a horse hostage. In my past move summaries I have expressed confidence that I was winning the opening, but now that I'm going to take the opening into deep waters, it would be disingenuous to pretend that I have everything under control.
At the end of the day I'm going to have to advance my horse on obscure justification. I haven't proved to myself that it works, but I haven't refuted it either. It just feels right. Does it feel that way because I won with swarming in the 2008 Postal Mixer, even from unsound positions? I don't know, and maybe by next year my feelings will be different. That's the beauty of Arimaa.
I will regret not getting the Mob's horse hostage with my elephant. That would have been a theoretical dispute I relish. There are some people who won't believe that giving up a horse hostage to the opposing elephant is a disadvantage unless it is proven to them. Unfortunately, I want to win more than I want to prove that particular point. I will play the strongest-looking move that my limited analysis can uncover.
The Mob has made sure of a rabbit pull (if only from a8 to a7) before I get a horse hostage, and that tips the scale slightly against my taking it. At the same time the Mob's horse has gotten even further from home defense, and the Mob has been unwilling to advance a rabbit to shore up c3. These decisions have helped them in the rabbit-pulling race, but have left c3 a little bit more vulnerable, which tips the balance slightly in favor of my counter-attack. The final weight on the scale is that my camel is poised for action whereas the Mob's camel is hemmed in, unless I play defense and allow the Mob camel to life. In short, although I may have overlooked something tactically, I feel my choice is strategically justified, if not outright invited by the Mob.
Anyway, I won't be too sorry to have missed a chance at the elephant-holding-horse-hostage dispute, because I am quite directly making a statement in the other major theoretical dispute: is the opening of Arimaa essentially a rabbit-pulling contest? I claim that it is not. It may be good to have opposing rabbits pulled forward to weak positions, but who has time to make such pulls? There are more important things going on, or so I have been asserting. We will get the clearest insight into whether I am right if the Mob takes the bait, pulls the rabbit with 6g RHhvEvHf>, and permits my swarming semi-blockade on move 6s.
Moves in which the Mob doesn't yet pull my rabbit leave the nature of the disagreement as yet unclear. If that happens, so be it. Some observers might look at this game as both sides playing soundly and waiting for the other side to make a blunder. I expect that there will be no blunders, especially not from the Mob, and instead the maneuvering will serve to bring the game to a position that each side thinks it is winning. That is to say, the contest is not about two opponents attempting the same strategy, where the winner is the one who has better execution; it is instead about two opponents with different judgment of what is important, where the winner is the one who has a more correct understanding.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
6g
« Reply #12 on: Jun 1st, 2010, 6:53am » |
Quote Modify
|
6g
I don't think the Mob will be able to stomach the cat sacrifice. The move may well be spotted and advocated by someone, but I don't think even a team effort will be able to analyze the cat sacrifice all the way to a forced material gain for the Mob. It will remain a concrete present loss in exchange for a nebulous future gain. When the ballots hit the box, the Mob will vote short-term and concretely.
I will instead predict 6g RHhvEvHf>, not because it is the best move, but for psychological reasons. The Mob will feel good to play a move that pulls a rabbit, because they played for a rabbit pull in the opening. It would be psychologically painful for them to admit that going after the rabbit was a waste of time in the first place; pulling the rabbit has some chance of proving the Mob right, whereas shifting strategy feels like an early admission of error. Everyone likes to have evidence that they are right, even if it is self-manufactured evidence.
I say this because I am projecting my own psychology onto the Mob. Possibly my best plan on move 5s was to chase the Mob's exposed horse with my elephant, but if so, then it was probably also the best plan on move 4s. If my elephant was going to leave the center anyway, it should not have first made a move that gains time for the Mob. But if my elephant stays in the center, then 4s didn't necessarily lose time. Thus I played my actual move 5s in order to justify the 4s that I played, and to avoid admitting that my 4s might have been a mistake.
An interesting twist to the psychological drama is the possibility that a rabbit-pull strategy was initially correct for the Mob, and it has only become wrong now that I have advanced my horse. It may be that to distract them from a small gain, I had to offer them a bigger gain. The Mob could shift gears now in the confidence that my move 5s was a sign of desperation. They could feel that their threatened rabbit pull forced me to make a bad move which they can take advantage of instead of pulling the rabbit. But it is not in human nature to be so psychologically nimble. Instead of taking the best course of action and instructing ourselves to feel good about it, we humans take the course of action that feels good and instruct ourselves to believe that it must be best.
One reason the Mob will be tempted by 6g RHhvEvHf> is that it threatens to push my horse to e3, fencing it and winning it the next move. Superficially it appears that to save my horse I am obliged to advance my camel, after which the Mob could push back my horse and attack my camel. Lucky for me I have the swarming option in my back pocket. I expect the Mob to overlook the swarm or discount its power if they see it. In spite of many swarming successes of late, it is still deeply ingrained that swarming is risky and probably unsound.
...
Today chessandgo commented on my game against clueless that "a horse hostage is worth significantly more than a camel hostage in my opinion, 'under the proper conditions' as you say jdb, basically that gold has an advanced piece (Horse) on the hostage wing that silver's camel cannot mess with within a move or 2". That describes the type of hostage the Mob could get against me by sacrificing a cat, and since chessandgo rarely comments on games these days, I can't help but wonder whether his comment now is part of a raging debate within the Mob. Given that a camel hostage is worth slightly more than a cat, getting my horse hostage by their camel must be worth far more than a cat, in chessandgo's opinion.
I obviously don't rate a camel holding a horse hostage so highly, or I wouldn't have offered it up for a cat. It is interesting to add this positional feature to rabbit pulls, horse frames, and elephant-holding-horse hostage as things chessandgo and I disagree about. How can the two leading players be at odds about so many fundamental questions?
The Mob has just overstepped their week for this move and started dipping into reserve, so even if my move wasn't good, at least it may have been unexpected and complicated.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
6s
« Reply #13 on: Jun 1st, 2010, 6:53am » |
Quote Modify
|
6s
The Mob played 6g Cc<^M<Hf> two days into reserve. This is a very natural looking move that makes me wonder why I thought the Mob would be forced to use its elephant on home defense. I suppose I thought pulling out the Mob's g3-horse with my camel would be devastating, but now I can't remember why I thought that.
So the Mob didn't pull my h-rabbit, but instead maintained both the threat to pull it and the threat to play E<HE>Ha>, securing the nice square on b6 for its horse. Apparently since they didn't execute the rabbit pull threat when they had a chance, it is the horse position that they are really after. I said that this game might prove to be a theoretical contest over whether rabbit pulls are worthwhile, but that won't be true if the Mob is as uninterested in rabbit pulls as I am.
For my move I find 6s MvvHd>> very tempting, because that would certainly force the Mob's elephant to change quadrants after all, before getting it horse onto b6. Unfortunately it appears that in some lines I would give up my camel hostage in exchange for winning just the Mob's f2-rabbit. I regret that the Mob got smart and rotated its cat away from f2. Nevertheless, there may be something in the attack I am overlooking; I will examine it further.
The obvious move 6s MvvHM^ appears much safer. I think it is forcing enough that the Mob can't afford 7g E<HE>Ha> due to 7s HM<Hd>>. So the Mob would have to play defense with either an elephant crossing, or by unfreezing and running away.
I believe that if the Mob's elephant crosses, it will work out in my favor. I can retreat my camel at the "cost" of exposing a rabbit, but if the Mob's elephant then goes back west, the advanced rabbit is actually a time gain for me, because it makes my camel threats more potent. I simply need to verify that if the Mob's elephant stays in the east I can do damage in the west more quickly.
The Mob's other defensive option of running away also looks all right for me. They can't end the move with g3 unoccupied, because my d4-horse will march right in to a strong square. So whatever is left in g3 I can pull right back to g4 for three steps. If we get into a trade of three steps to pull and three to retreat, I think I can make better use of my fourth step than the Mob can, because I want to push a rabbit up the h-file but the Mob doesn't have a similar incremental improvement to its position. Meanwhile the way for the Mob to avoid a three-for-three step trade is to put a rabbit on h4, but then I can pull that rabbit with my camel to good effect, or so it appears to me.
At the end of the analysis, the Mob and I each have an elephant in the west, so we can each prevent the other from making a capture there, but my camel is the sheriff of the east, so I'm the only one that can make a capture there. The Mob has to get a positional feature in the west that is worth more than what my camel is doing in the east.
...
I haven't had much time to analyze this week, and the weekend promises to be busy too, so I'm just going to close my eyes and move here on Friday night and pick up a little reserve. Pulling the Mob's horse with my camel looks safe enough. If they have a strong reply or I have a strong move I am overlooking, too bad. That's just the cost of being busy.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
7g
« Reply #14 on: Jun 1st, 2010, 6:54am » |
Quote Modify
|
7g
I was startled to realize this evening that the Mob has overstepped its week, and is dipping into reserve as I write. My move 6s was very natural and straightforward, so they can hardly have been taken by surprise. If that move wasn't in their analysis tree, they must not have had an analysis tree. Indeed, I would not have been surprised if the Mob had gained reserve after my 6s, just like they did when I played the 4s they expected from my World Championship game against Adanac.
What accounts for the Mob's indecision? Perhaps they have been distracted by the most exciting Computer Championship ever, not to mention the World Championship which is still underway. Certainly I have been distracted myself, but then again, it wasn't my move. Taking time off on the opponent's move is not nearly as costly as slacking on your own turn. Unfortunately, if it is mere slacking, I can expect the Mob to get much stronger when the tournament season is over and the Mob game is (relatively speaking) the most interesting event left.
Another possibility for the Mob's slow decision might be that there are many plausible options in the current position. The Mob has at least three quite different choices: defense with small pieces, an elephant crossing, and ignoring my horse pull in favor of getting the Mob's horse into b6. It can be wise to take a little extra time when the possible courses of action are so divergent. And even if the choices are not drastically different from one another, it can be confusing to have many reasonable variants.
However, a fresh look at the position after a week off gives me some hope that the Mob's slowness has a more promising explanation. Perhaps they are simply having trouble finding any good move. I had initially made an off-hand prediction of 7g Dg^Hg<vDd^, but it seems I can counter that to good effect by blasting my camel into g3. Indeed, just now when I played out a few variations, it seemed to be of considerable importance that my last move brought my camel from g6 up to g5. My more-advanced camel means that if I answer a retreating move by the Mob with any type of attack against the f3-trap, my attack will arrive one step faster than anything I could have launched on move 6s.
This is potentially very good news for me, because I already thought poorly of the Mob's elephant crossing and horse-to-b6 options. If the status-quo retreats are also unattractive to the Mob, then what plans are left at their disposal? I think I must alter my prediction to the much looser move 7g Ra2^Rh2^^Hgv. If the Mob is desperate, they might just give up on the pretense that they can get an advantage in a rabbit-pulling race in any variation.
In the longer term, supposing the Mob is truly in positional difficulties and I can maintain the pressure, they might have a considerable stretch in which it is difficult to find a good move, and therefore difficult to move quickly. I love to imagine the Mob bouncing around alternatives that are all subject to refutation. What fun if they start thinking they can't play X but they can't play Y either, so they can't even decide!
Well, perhaps there is another explanation. Perhaps I am spinning a dream, but the dream comes close to what I will have to bring about in order to win this game. My best hope truly lies in demoralizing the Mob, grinding them down so they feel hopeless and cease to function as a team. Once they are no longer enthusiastic and rallying together, the game should be no more difficult for me to win than any ordinary postal game.
|
|
IP Logged |
|
|
|
|
Home
Help
Search
Members
Login
Register
Reply
Notify of replies
Send Topic
Print