Author |
Topic: Fritz's commentary (Read 13042 times) |
|
Fritzlein
Forum Guru
Arimaa player #706
Gender: 
Posts: 5928
|
 |
22s
« Reply #45 on: Jun 1st, 2010, 7:22am » |
 Quote  Modify
|
22s
I am surprised the Mob needed twelve days on this move. Was something wrong with activating their horse immediately, although I didn't see it? Probably they were merely having trouble deciding between good and better, because their actual move of 22s CcvRc>Dc^H^ keeps my elephant out of e3, forcing me to defend the f3-trap from f4. They lose some time building the phalanx, but probably make it up with interest by forcing my elephant to occupy a worse square. I am still puzzled, though, about why the Mob spent a step filling d2, to keep my elephant out of d3, when their horse in f3 keeps my elephant out of e3 in any case. Perhaps I had a tactical resource I overlooked.
My camel is temporarily safe because of the Mob's horse in f3, but my g4-cat is threatened, so I have no reasonable option beyond moving my elephant to f4. There are only four ways for me to do that; either retreat my horse to f5, or retreat my horse to e5 while moving my dog to f5, e6, or d5. Given my narrow range of very-similar choices, it shouldn't take me long to move this time.
With the elephant deadlock about to be restored, it is time to take stock strategically of the effect of the tactical interlude. On the positive side, I regained control of the c6-trap, and snuffed out the goal threat that the Mob's horse was slowly developing. I therefore have many moves to try to free my elephant again before it is forced to leave to stop a western goal. Furthermore, the absence of a horse on each side mitigates the disadvantage of losing my camel. I could now trade my camel for just a horse and a rabbit to land in a materially even position.
On the negative side, I can't prevent the Mob from activating its remaining horse, because I needed so much time to get my elephant home. Furthermore, the limited elephant mobility I spent so many moves building up is all gone now. With a horse absent from each side, it may be even harder for me to regain elephant mobility than it was the first time. Finally, I exposed my rabbit on a4, giving the Mob a much easier capture target than they had the first time around. I will be quite lucky if I can trade it away for the Mob's b4-rabbit rather than losing it outright.
At the end of the day, the Mob's camel hostage is still keeping them in control. It seems improbable that I will be permitted to abandon the camel without conceding material disadvantage. If, on the other hand, the hostage situation is maintained, the Mob has plans in the west that will work faster than I can improve my position in the east. Either way I am in trouble.
...
A bit more analysis doesn't brighten my outlook at all. My only active plan is to swarm the f3 trap, and that is too slow. Of my few move choices, moving my horse to f5 ends up looking most like a swarm, so that's what I'll do. Also, the Mob's camel is probably quickly destined for the west, so my horse will be relatively less powerful in the west than in the east. But there's not much in it for me.
|
|
 IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
23g
« Reply #46 on: Jun 1st, 2010, 7:23am » |
Quote Modify
|
23g
Superficially, it appears reasonable for the Mob to go after my exposed rabbit immediately with 23g Cc<<RC>. However, there is no reason for them to start fighting in the west while we are on equal terms there with cats and dogs. That would be voluntarily giving up the advantage that their camel hostage provides them.
No, in order for the Mob to use their advantage, they must mobilize their camel and horse. I don't see why they wouldn't do so immediately. I predict a straightforward developing move like 23g M^H<M^Cc<, which leaves me very little to go on.
Unfortunately, I am afraid they have even better in the form of 23g M^H<Dc<^, hemming in my elephant and getting their dog to its natural square. I'm not sure how I could make progress in face of this simple plan. Ugh.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
23s
« Reply #47 on: Jun 1st, 2010, 7:23am » |
Quote Modify
|
23s
My first thought on seeing the Mob's 23g Cc<M^H<M< was that I had caught a break. The empty e4 square allows me to play 23s E<EH>Hv, beginning a process of liberating my elephant that could continue with steps EHvE<HvDv> in some order. Belatedly I realize that the whole plan is brought to a screeching halt by 24g Re2> plus three steps such as M<DM>. I now question even my choice to bring my horse to f5.
Since I can't liberate my elephant, my only hope is to abandon the f3-trap with my elephant at an opportune moment, hoping to lose as little as possible. But when will there be an opportune moment? My swarm is now working against me, because all my advanced pieces besides the camel are also in jeopardy. At a minimum, on the two turns after my elephant leaves, the Mob can capture my camel and cat. When will leaving ever be worth that much?
I might now contemplate a waiting move to shore up my trap defense and get my c5-dog to a less exposed square. Sadly, the moment I leave c5 with my dog, the Mob can occupy c5 with their mobile dog. Once the c6-trap is protected by a gold dog, my elephant will stand to gain even less from abandoning the f3-trap.
I have taken more than three weeks to make this move, not due to prolonged analysis in an attempt to find a way out, but due to despair and disinterest. It seems that not only was I wrong to trade horses, I was wrong by more than I first thought. I could willingly resign now, although that would be a disservice to the Mob. No, let me continue to play on for a while a least, if not with the same enthusiasm as before, at least with a hope to make things interesting again at some point before I lose.
It seems the only way I can regain a fluid position with some possibility to make progress, as opposed to passively awaiting disaster, is to immediately attack the Mob's camel, effectively offering to trade it for my own camel and cat. I will then be behind a whole cat with no compensation, but at least I will have freedom to plan. I believe that 23s E<EM^De^ will not end with worse than a cat loss for me, e.g. 24g ME<^* 24s EM^H^Re>. Let me finally take my medicine and pay the price for giving up a camel hostage with my terrible move 10s.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
24g
« Reply #48 on: Jun 1st, 2010, 7:24am » |
Quote Modify
|
24g
What is to discuss? The Mob gains material and takes no chances by playing 24g ME<^*. Not only will their position be widely superior, they will also gain reserve time.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
24s
« Reply #49 on: Jun 1st, 2010, 7:24am » |
Quote Modify
|
24s
Yep, the Mob played 24g ME<^* in only 20 hours. I think instead of 24s EM^H^Re> I will play 24s EM^H^>, since it leaves my eastern flank in better shape to have the horse there.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
25g
« Reply #50 on: Jun 1st, 2010, 7:25am » |
Quote Modify
|
25g
Taking my cat is obvious, either with the elephant or with the dog. The only reason the Mob might deviate would be to prevent me from getting in the first shot in the west with 25s EM<*RC<. So 25g Cb<RC>Rb< deserves consideration, but I think it is too great an invitation for my dog to head towards b3. 25s EM<*Dcv< looks like a refutation.
Of the two cat captures, I expect the Mob will decide against using the dog, so as to leave it in place to pull a second rabbit. My prediction is therefore 25g CE>ECv*.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
25s
« Reply #51 on: Jun 1st, 2010, 7:25am » |
Quote Modify
|
25s
The Mob failed to take my cat, for reasons that are not clear to me. 25g E<DhC<Re2> will still permit a cat capture for next turn, but what has been gained? The Mob can't intend to complete the capture with the elephant, or it would have done so at once. They must intend to complete the capture with the dog. But they also could have finished the capture with the dog this turn, unless they want the dog on f4 instead of on g3. So my best guess is that the Mob is trying to save time by rotating their horse to g3 and dog to the middle of the board. That makes sense because with a material advantage the Mob wants strength to meet strength, elephant to fight elephant, horse to fight horse, all the way down the line so that their extra piece weighs in their favor. They don't want a race where time can matter more than material.
However, the Mob may not end up saving any time despite their contortions, because there is more important stuff going on than their piece re-alignment. If I play the natural 25s EM<*Evv, I don't think the Mob will want to complete the maneuver with 26g H>>DfCv* due to 26s CR>RDc> with a nice rabbit frame to give me some compensation for my lost cat, instead of zero compensation as I have presently. And if they don't complete their piece realignment at once, they might be forced to retreat their exposed dog first, actually losing time in the exchange.
Perhaps the answer to the riddle of the Mob's move is that the immediate cat capture on 25g offered me a different form of compensation that I didn't see. If that is true, then the Mob's actual move need not be better than it looks; only my suggested move needs to be worse than it looked.
I obviously must capture the Mob's camel with two steps. There is nothing to gain from delay. For the other two steps I could re-centralize my elephant as above, but it is tempting to pull their b4-rabbit to a nice hostage position on b5. At first glance this appears to be refuted by, 26g DfCv*E<^, but apparently then I have a nice resource in 26s E<Dv<Ev, where the rabbit hostage is almost as good as a frame, and I have the bonus of putting my dog on a nice attacking square.
I'm in a bit of a dilemma, because my strategic instinct is to immediately recentralize my elephant. It feels like the right thing to do. On the other hand, the rabbit pull seems tactically difficult for the Mob to refute. In the best lines I can work out so far, at the end of my playouts, I am slightly better off with the rabbit pull than the centralized elephant. I guess the only sensible thing to do is come back to it later, although I confess I burned ten days off my clock before even looking at it the first time.
...
Another look at the position didn't turn up anything wrong with 25s EM<*RCb< other than being behind by a cat, which I am anyway. My head is going to overrule my gut on this one. I can't reject a strong move for which I can't find a refutation just on my gut feeling that it is unsound.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
26g
« Reply #52 on: Jun 1st, 2010, 7:26am » |
Quote Modify
|
26g
The Mob could play 26g ECv*RDgv, ignoring my threat to their b-rabbit and countering with a threat of their own. This is a reasonable plan, but I don't expect it, because it is not consistent with their failing to take my cat on 25g. If they wanted to work in the east, it would have been faster and more flexible to go with the immediate cat capture. Besides which 26s HvRg>Re>Ev seems to prove that my rabbit threat is better than the Mob's, and to inch me closer to a type of unbalanced racing game that I must create to have any chances when down material.
More likely the Mob took pains to centralize its elephant one square on 25g precisely to move it west now and save their rabbit. Indeed, they could even delay capturing my cat for a turn by threatening my dog with 26g E^<EDv. But then 26s EvvD< gets my dog to a very nice square, and the Mob has accomplished nothing by again delaying capture.
Therefore I predict 26g DfCv*E<^, which I intend to meet with 26s E<Dv<Ev. The terminal position is terrible for me, but holding the Mob's b-rabbit hostage at least gives me a tiny bit of positional compensation for the lost cat.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
26s
« Reply #53 on: Jun 1st, 2010, 7:26am » |
Quote Modify
|
26s
Move 26 was the seventh time I correctly predicted the Mob's move. 27% accuracy is much better than I expected, but still lower than I would expect if this were a chess game.
A quick look at the position gives me no reason to change my mind about playing 26s E<Dv<Ev. I don't see any crushing responses for the Mob, and even if I did, I'm not sure what else to do. I need to save my threatened dog, and this way of saving it incidentally gets my dog to a nice square. I will move quickly this time.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
27g
« Reply #54 on: Jun 1st, 2010, 7:26am » |
Quote Modify
|
27g
The Mob can't be happy about my dog threatening to invade b3. I'm sure they would like to bring their horse west to take care of that dog, but that line is strategically barely worth considering. Without the Mob's horse in the east, my own horse will spring to life, creating a very sharp position with latent race potential. With a material advantage, an unclear race is the last thing the Mob wants.
On the other hand, it would not be in character with the Mob's fighting spirit to tamely accept being tied down to defense of their hostage rabbit with, say, 27g E^Dc<Cb<D^. Therefore I predict that they will try to complicate by picking on the most obvious target with 27g E^EDvRh1<, final step to preempt any goal-threat tactics by me. With my horse in danger, I won't be able to snatch the Mob's rabbit immediately, but I don't see why the ensuing maneuvering should favor them over me. Apart from being down a whole cat with no prospects of recovering it, my position is in reasonable shape.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
27s
« Reply #55 on: Jun 1st, 2010, 7:27am » |
Quote Modify
|
27s
I was right that the Mob would counter-attack rather than defending, but wrong about the specific move. With 27g Ev<Dc<H>, the Mob threatens my b4-dog with capture, preventing me from immediately capturing their b5-rabbit. I had thought to answer such a move with 27s E>DbR>, winning a rabbit on 28s, but I see now that their horse step permits them to reply with 28g H>Df>RD>, intending to trade rabbits rather than losing one outright. As a bonus to the Mob, they would complete the rotation of their horse to the g3-square, not losing time as I had hoped, but actually gaining time by leaving their dog high and out of the way of the crossing. As usual I didn't look at the tactics deeply enough.
Despite the probability of merely trading rabbits, what can I do apart from 27s E>DbR>, plus some fourth step? I must save my dog somehow. Can I contemplate occupying d3 with my elephant? It seems extremely loose to move my elephant away from the wing on which the Mob has an advanced rabbit, but that rabbit can't easily advance, and indeed I would still be threatening to capture it. Perhaps this complication is the reason the Mob used almost seventeen days to decide on 27g.
...
After a second look, it seems clear that I must hold the tension with 27s E>vEc<. In the long run this may turn out worse for me than a simple rabbit trade, but I have to give the Mob more opportunities to make mistakes. Straightforward play is no longer reasonable given my material disadvantage. If I sink my ship all the more quickly, it will at least be better than a calm, protracted, inevitable loss.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
28g
« Reply #56 on: Jun 1st, 2010, 7:27am » |
Quote Modify
|
28g
Stylistically the Mob is not given to patient defense. It would be very like them to attack something instead of defending their hanging rabbit. For that reason I might anticipate 28g Df>RD>H>. Also this would be consistent with having left the dog high in the first place to ease their horse’s transition to g3. On the other hand, I could reply with HvRf<<Hv, creating a much tenser position than the Mob should be comfortable with given their material lead. Even if they want to play tactically by nature, it would not be wise in the present circumstances. They should make every effort to keep the position as calm and balanced as possible.
Therefore I will instead predict 28g Cb<Db^DE^, defending their threatened b-rabbit. This justifies the westward dog step of 27g. Also, pulling my dog from b4 to c4 frees locks down my elephant to defense while opening a square for counter-attack. The Mob is certainly not committed to passive defense after this move, as they will still be able to harass my a5-rabbit and g5-rabbit, potentially overloading my elephant.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
28s
« Reply #57 on: Jun 1st, 2010, 7:28am » |
Quote Modify
|
28s
After another long think, the Mob played 28g H>CcvCb<Db^. They are conceding their b5-rabbit. I can’t believe it! To be behind a whole cat seems insurmountable, but being down a cat for a rabbit suddenly makes me feel like I’m back in the game, and could even potentially come back to win. Yippee!
Of course the Mob didn’t overlook the capture, so either they thought much more highly of my position than I did, to the point that giving up the rabbit is cutting their losses, or else there is catch, and the offered rabbit is a snare. Of course I can’t take the rabbit with my cat, because then they would just flip my cat into c3 for a cat frame. But why not take with my dog? True, it would allow their dog to invade b6, but I have enough support from little pieces at home that the attack hardly looks crushing. Yes, an elephant-dog attack is valuable, but surely less valuable than a whole rabbit?
…
OK, now I see that 28s DbR>^* can be met by 29g Db^RD^Dfv, after which I could not prevent my rabbit from being framed. With the current piece alignment, a frame would be essentially unbreakable. Furthermore, there would be false protection around c6, which I would have to defend before trying to stir up counterplay, and further furthermore my dog would be frozen on c5 with no friendly piece that can conveniently unfreeze it. The fourth step retreating dog to f3 efficiently completes the Mob’s piece rotation to align strength against strength and deaden the position just when I need to make it tense. All of this makes my rabbit “capture” no better than a rabbit trade. I should have known the Mob would not voluntarily give back material without seeing a concrete refutation.
On the other hand, what else can I do? I need tension, action, and imbalances, but the Mob’s last move put a dog on the western corner where my dog was attacking, and a horse on the eastern corner where my horse would want to attack. They are sucking the life out of this position, exactly as one should do with a material advantage. Their only weakness apart from the b5-rabbit that I dare not take is that their dog is still temporarily exposed on f4. So perhaps 28s H<vDH<?
Moving my horse toward the center could eventually leave my eastern flank exposed to the Mob’s horse, but now is not a time to be shoring up weaknesses. I will have to weaken some part of the board to get play elsewhere, and create any kind of race. Furthermore, I see only three responses for the Mob to save its dog, all of them requiring four steps. Forcing moves are easier to analyze.
29g E>>EH< is optically crushing, getting the gold sheriff next to the silver deputy and hitting hard at my suddenly-depleted eastern flank. But it seems to me that after 30s Re>CR>E^ I may be able to get the rabbit for nothing. It would create a race at a time when I’m very eager to race, with elephants on opposite wings and horses on opposite wings, which is all I can ask for.
The Mob’s second saving move, namely 29g H<^DfR^, seems to at least gain me time after returning my horse to f5 plus three other steps. That leaves only 29g H^HR^Dfv, after which 29s H>Rh5v<Rh6v gives the Mob something to think about.
All in all I have convinced myself that picking on the Mob’s exposed dog is at least not bad, and keeps tension in the position. As this move coincides with the early rounds of the World Championship, it is a great reminder for me to consider more vigorous moves in my live games as well.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
29g
« Reply #58 on: Jun 1st, 2010, 7:28am » |
Quote Modify
|
29g
In addition to the moves I previously mentioned, the Mob can also save its dog with 29g E>^HEv. I should not get my hopes up, though, for either this elephant move or 29g E>>EH<, because either is a dream come true for me, i.e. a live position with possibilities to swindle a win.
No, as much as I would wish for the Mob’s natural taste in active play to guide them to an incorrect choice, I am afraid I must predict the “deadest” move of all their options, namely 29g H^HR^Dfv. With the Mob’s horse out in front of their dog, it will be tough for my horse to keep picking on the dog. It may get to the point that I can’t stir up further trouble without making clearly weak moves, at which time I can only hope the end is mercifully swift.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
29s
« Reply #59 on: Jun 1st, 2010, 7:29am » |
Quote Modify
|
29s
I am shocked that the Mob played 29g E>>EH< in under two days. Is it so obviously superior to the move I predicted? I can’t believe they would move so fast unless they had discussed my move 28s before they played 29g, but in that case they must not have considered my 28s to be any threat, and must now consider their position to be very strong. In short, the Mob is not only disagreeing with my evaluation, but doing so with a confident flair.
In previous analysis I liked 29s Re>CR>E^, but perhaps the Mob has a refutation already in the can for that move as well. I will be in no hurry to move now that the position has shifted from dead to sharp in only two ply.
As I consider moving my rabbit from e7 to f7 on 29s, it is worth remembering move 18s, when I moved it from f7 to e7. I hesitated then to abandon defense of my f6-trap, but I didn’t hesitate enough. Since then, the tactical vulnerability of f6 has caused many of my potential plans to fall through, while the central position of the rabbit has been of no use whatsoever. That rabbit step on 18s was worse than useless: is has had negative ramifications ever since, and finally I am forced simply to undo it, wasting another step in addition. My failure on 18s traces back to my misjudgment that the Mob ought not play a horse invasion on 19g; if I had seen the full power of that move I would have positioned myself better to survive the aftermath. This is a powerful lesson to me to be less sure of my strategic intuitions, and to give even more weight to generally good-looking moves over special-purpose moves.
…
I analyzed for a long time on 29s Re>CR>R8>, changing the fourth step from my previous favorite. The first three steps seem very logical. First, I must protect my e6-dog from capture because I can’t make a counter-threat as big as a dog. Second, I must take action in the west as quickly as possible, because I intentionally permitted my east to be weakened in the hopes of creating a race. Having succeeded in getting an imbalance by distracting the gold elephant to the other wing, I had darned well better take advantage of my side of the imbalance as quickly as possible. The fourth step, though is more problematic.
I like R8> better than E^ for my fourth step because my elephant’s current location is strong and flexible. If the Mob chooses to defend the c6-trap, I will want to frame their rabbit and attack the c3-trap immediately, for which purpose my elephant is better on d3 than d4, and my rabbit on c8 means I need only three steps for the rabbit frame so I would have a general-purpose step left over. Also if the Mob attacks the f6-trap with 30g DfRvED<, I would like to be able to respond with 30s CR^*Rh6<E>, i.e. two steps to take the Mob’s rabbit, one step to defend the f6-trap, and one step to defend the f3-trap. I seem to have a good position whether the Mob plays to defend its own material or make a material counter-threat.
Unfortunately, I seem to have a losing position if the Mob plays for goal threat rather than material with 30g DfRvRh2^^. They would then have eight steps to goal via RD^R^RD^R^^^, which is exceptionally difficult for me to halt or slow without either my elephant or horse in the east. It seems that I have gotten the race I want, and I am losing it! And if I use 29s to position to get my elephant or horse east to stop the threat from developing, then I am trying to call off the race with a net loss of time.
After analyzing the Mob’s eastern goal attack long enough to verge on strategic despair, it finally occurred to me that I can use my fourth step to play Rh5v! This single step slows down the Mob’s eastern goal run by three steps, which appears to be just long enough that I can capture their b5-rabbit and yet defend goal, assuming they still choose a straight goal attack. For me it is counter-intuitive that advancing my own rabbit would slow down theirs so much, but after staring at the board for a while it finally occurred to me.
Thus I will play 29s Re>CR>Rh5v in high spirits. It is true that slowing the Mob’s h-file rabbit advance leaves f6 slightly less defended, which weakens me in the line where the Mob chooses to defend the c6-trap, but if the Mob feels forced to defend rather than counter-attacking, I will feel that I have won a moral victory, and will fight on with renewed enthusiasm.
|
|
IP Logged |
|
|
|
|
Home
Help
Search
Members
Login
Register
Reply
Notify of replies
Send Topic
Print