|
||||
|
Title: 2010 Postal Mixer Post by Fritzlein on Mar 25th, 2010, 10:08am Omar, what ever happened to the 2010 Postal Tournament? I see it was included in the events schedule with registration supposedly closing on April 1st (yikes!) but there is no link to it from the events page and it has never appeared in the game room announcements. The URL http://arimaa.com/arimaa/tourn/postal/2010/ is dead. Are we going to have this tournament or not? I personally think it is an essential element of Arimaa calendar. The World Championship is a great event, but it is not for everyone. I can understand people not wanting to sign up for it if they are just going to get mauled. The Postal Mixer, in contrast, is truly for everyone. Nobody should fear signing up for it; the matches are intentionally paired to be close and everyone learns a ton. Usually we break significant new theoretical ground. I would hate for it not to happen because we just forgot about it. :P |
||||
|
Title: Re: 2010 Postal Mixer Post by camelback on Mar 25th, 2010, 1:00pm Can there be more features or improvements in Postal tourney this year? I feel that it should have more distinctive features compared to Auto postal. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Mar 26th, 2010, 4:38am What features and improvements did you have in mind? Maybe the autopostal has made the Postal Mixer obsolete, and I am just behind the times. :o |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Mar 26th, 2010, 6:49pm Awesome, thanks for opening the registration, Omar. I see we have five players in already. Way cool. 8) |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Mar 27th, 2010, 12:05am Are the bounds for simultaneous postal games in autopostal and postal mixer dependent? Should I decrease bound for autopostal ... ? |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Mar 27th, 2010, 6:04am The bounds are independent. You could theoretically get the maximum in both. Plan accordingly... The Postal Mixer is clearer, because you immediately get exactly as many games as you request, unless you request a ton. You could just turn off the autopostal until the Postal Mixer pairings are announced and you know where you stand. In the past, people (like blue22) who have signed up for twenty games ended up getting only about fourteen. I'm not sure whether to expect the upper bound to be more or less this year. On the one hand many folks so far are signing up for a small number of games, but on the other hand we already have eleven registrants in less than 24 hours since registration opened, so we may end up with a huge field this year, resulting in more possibilities for the game-hungry players. |
||||
|
Title: Re: 2010 Postal Mixer Post by Janzert on Mar 27th, 2010, 8:34am Omar, you can sign up OpFor for a maximum of 20 games this year. Thanks, Janzert |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Mar 27th, 2010, 9:37am on 03/27/10 at 08:34:16, Janzert wrote:
Yay, we'll have a bot again this year! Or two? Maybe another developer is interested too... |
||||
|
Title: Re: 2010 Postal Mixer Post by Eltripas on Mar 27th, 2010, 7:12pm on 03/27/10 at 08:34:16, Janzert wrote:
Will opfor ponder in postal games? |
||||
|
Title: Re: 2010 Postal Mixer Post by Janzert on Mar 27th, 2010, 8:04pm on 03/27/10 at 19:12:05, Eltripas wrote:
No, it won't. I've always set OpFor to think a minimum of 2 hours and maximum of 4 for each move. So far this has kept it thinking literally continuously for the first few months of the tournament. Janzert |
||||
|
Title: Re: 2010 Postal Mixer Post by camelback on Mar 27th, 2010, 8:13pm on 03/26/10 at 04:38:22, Fritzlein wrote:
Ok, Here are my thoughts regarding postal tournament I just noticed that there is no more a $20 deposit for registration this year. Thats good. Now the only difference compared to auto postal are a. A seperate web page and scoring system to track the games b. A chance to paired up with higher ranked players if you choose more games. This is not big advantage since you can always invite strong players personally if need be. Some ideas to make it better and more exciting: 1. Everybody chips in some small amount and the most improved player in the tournament gets rewarded. It can be cash or some Arimaa goodies. I guess this will be an attracting feature for the newbies. Among advanced players, player with most wins can also be rewarded. or 2. Give some title for the most improved and most won player in the tournament, example "Super owl" for most improved "owl howler" for most won or high score. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Mar 29th, 2010, 7:07am Hmmm, only two days left to register, but no RonWeasley. Don't get me wrong: with 21 players already signed up, this is guaranteed to be a great year for the Postal Mixer no matter what. It's just that Ron has been there every year since the beginning. If he doesn't sign up, it would be the end of an |
||||
|
Title: Re: 2010 Postal Mixer Post by Eltripas on Mar 29th, 2010, 7:31am on 03/29/10 at 07:07:45, Fritzlein wrote:
Aren't there still like 10 days left to register?, I think you are confusing the deadline of the screening with the deadline of the registration for the postal mixer. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Mar 29th, 2010, 8:50am on 03/29/10 at 07:31:10, Eltripas wrote:
According to this page, registration ends April 1. http://arimaa.com/arimaa/wc/2010/sch.html ...But maybe that's an April Fool's joke ;) |
||||
|
Title: Re: 2010 Postal Mixer Post by novacat on Mar 29th, 2010, 9:22am Registration ends at 6 pm EST on Friday April 9th according to the registration page. http://arimaa.com/arimaa/tourn/postal/2010/ |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Mar 29th, 2010, 10:12am OK, that makes more sense anyway, since the play starts April 11 on both schedules. There's no need for a large gap between the end of registration and the beginning of play, since all that needs to be done is fix the ratings and produce the pairings. |
||||
|
Title: Re: 2010 Postal Mixer Post by RonWeasley on Mar 29th, 2010, 12:04pm on 03/29/10 at 07:07:45, Fritzlein wrote:
I just noticed that the registration is open. I was thinking of playing fewer games this year than usual, because I'm having trouble keeping up with the two to four games I usually have. But wait! This is supposed to be fun. I'm getting too caught up in playing well and forgetting that I would rather play for fun. So I'm going to play lots of games and apologize in advance if I play carelessly and lose, as if that's going to make anybody feel bad. Look for Fritz to ask me to explain why I didn't attack on move 12. And for me to completely ignore him, because I won't know the answer. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Mar 29th, 2010, 1:52pm Yes, it is supposed to be fun. I think it can be fun to play in a few minutes on a lot of postal games, or think for a long time on a few postal games, or think for a long time on a lot of postal games! The thing that is not fun is feeling that I have to win every game, and agonizing over every move because I am so afraid of losing. I need to get over that at every time control :-/ |
||||
|
Title: Re: 2010 Postal Mixer Post by omar on Mar 31st, 2010, 8:31am I signed up OpFor for the postal. It would be great if a few more bots played. Although it's supposed to be fun keep in mind that these games will be rated and considered 'event games' and will effect your WHR and WHRE ratings. I am taking a chance and trying to see how it works out without the registration deposit. I hope everyone takes the games seriously and doesn't abandon them. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Mar 31st, 2010, 9:47am Yes, we're already going to have the biggest Postal Mixer ever, even if nobody else signs up. Hopefully we will also have the most completed games ever. Some historical trends suggests that when the number of players goes up, the number of mismatched games goes down (yay!) while the number of games resigned or lost on time goes up (boo!). We shall see whether the trend continues in 2010. Year . . 2005 2006 2007 2008 2009 2010 Players . . 16 19 20 15 14 31? Games . . 80 95 85 84 70 119? Mismatches . 9 5 3 21 19 Not Finished . 20 31 18 8 4 % Mismatches . 11 5 4 25 27 % Not Finished 25 33 21 10 6 |
||||
|
Title: Re: 2010 Postal Mixer Post by jdb on Mar 31st, 2010, 11:04am Omar, you can sign clueless up for a maximum of 20 games Thanks Jeff |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Mar 31st, 2010, 11:33am on 03/31/10 at 11:04:58, jdb wrote:
Woohoo! Thanks, jdb. This will be humanity's first glimpse of clueless's postal strength, correct? Will you, like Janzert, basically run a computer full-time on this task? That's very generous. By the way, for both of you, a fun little feature would be to have to the bot chat something like time spent, depth searched, nodes evaluated, or such-like on each move. Chatting the eval would be particularly fun for the human opponents, but probably gives away to much. I don't know how hard chatting would be, though, and how much effort you would want to spend on a feature that doesn't improve playing strength. Simply entering your bots at all is a big contribution to the fun of the tournament. |
||||
|
Title: Re: 2010 Postal Mixer Post by BlackKnight on Mar 31st, 2010, 1:36pm on 03/31/10 at 11:33:22, Fritzlein wrote:
Because Rat is implemented in a different way, she always talks a lot before she starts searching for a move. Maybe too much. ;) So I also was wondering whether or not the bot interface and the server actually allow the bot to send messages to the chat window during the game. |
||||
|
Title: Re: 2010 Postal Mixer Post by Janzert on Mar 31st, 2010, 9:30pm If I'm remembering correctly, when I was making the AEI gameroom interface I thought the server protocol supported the bot chatting, but I never tried it out. Anyway I'm pretty positive neither interface currently supports the bot chatting. It is something that I have thought about adding to AEI before though and really should try out. A related thing I do wish the server supported was a way to chat (one-way) to the spectators, instead of the opponent. Edit: Duh, immediately after posting I realised/remembered that yes the server protocol does for sure allow the bot to chat. That is how both interfaces (AEI and Getmove) chat the initial bot greeting in a game. I do still need to add a way for the program to access it though. Janzert |
||||
|
Title: Re: 2010 Postal Mixer Post by Eltripas on Mar 31st, 2010, 10:13pm on 03/31/10 at 11:33:22, Fritzlein wrote:
The bot should also make fun of you when he "thinks" it is winning and beg for mercy when he "thinks" it is losing. |
||||
|
Title: Re: 2010 Postal Mixer Post by chessandgo on Mar 31st, 2010, 11:47pm I used to have a chess program which, when losing, liked to say things like: "Your computer has been infected by a virus, reboot IMMEDIATELY!" or "My code must have been tampered with" And when winning: "I learnt a game with X's and O's yesterday and ... no, forget it, it would be too complicated for you" |
||||
|
Title: Re: 2010 Postal Mixer Post by omar on Apr 2nd, 2010, 10:24pm on 03/31/10 at 11:04:58, jdb wrote:
Awesome. Clueless is signed up. Wow, the postal mixer is going to be a mega event this year. |
||||
|
Title: Re: 2010 Postal Mixer Post by Tuks on Apr 9th, 2010, 7:10am 40 people signed up, that's triple last year! |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Apr 11th, 2010, 7:01am Apparently Omar decided to pair the games based on game room rating this year instead of based on whole-history ratings. I think I would have preferred WHR so that I would have been paired with the best anti-human players rather than the best bot-bashers, but it looks like a raft of fun games nonetheless. In another change from last year, the games page appears not to be updating with the move number (and side to move) of each game. The continually updating status is a quick and easy way to check on the progress of the tournament. This year more than ever the feature would be useful because it will be impossible to look in on every game. What a turnout! 42 players and 144 games! That's double the games and triple the participants of last year. Game-hungry players get to gorge on fifteen to seventeen games this year. Unless people forfeit their games in droves, this will be by far the most awesome Postal Mixer ever. Year . . 2005 2006 2007 2008 2009 2010 Players . . 16 19 20 15 14 42 Games . . 80 95 85 84 70 147 Mismatches . 9 5 3 21 19 23 Not Finished . 20 31 18 8 4 % Mismatches . 11 5 4 25 27 16 % Not Finished 25 33 21 10 6 *edit* recounted the games |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Apr 11th, 2010, 9:34pm First 24 hours summary: 30 of the 42 players started moving. Hippo vs. Eltripas is the most advanced game, making it all the way to move 9. :o |
||||
|
Title: Re: 2010 Postal Mixer Post by RonWeasley on Apr 12th, 2010, 5:48am Expect this week to go slowly since the owls are busy sending tax returns. |
||||
|
Title: Re: 2010 Postal Mixer Post by 722caasi on Apr 12th, 2010, 3:59pm I am playing all twenty games, in order to gain experience. Will that cause too many mismatches, or is it okay? |
||||
|
Title: Re: 2010 Postal Mixer Post by omar on Apr 12th, 2010, 4:53pm I was contemplating using WHR ratings, but a number of players did not have any WHR rating. Karl, I'm showing 147 total games. |
||||
|
Title: Re: 2010 Postal Mixer Post by omar on Apr 12th, 2010, 5:02pm on 04/12/10 at 15:59:07, 722caasi wrote:
I guess it depends on how much of a rating difference you consider a mismatch. If you consider 400 points as the cutoff then only your games with Tuks and Adanac would be considered a mismatch. However, it will be good experience trying to do the best you can against such strong players. Also the more players that are signed up, the less chances of a mismatch. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Apr 12th, 2010, 8:43pm on 04/12/10 at 16:53:24, omar wrote:
I also count 147 games on a second try, with 23 mismatches. I wonder how I counted wrong the first time. Quote:
Sure, we had that problem already last year because OpFor had no WHR. We just substituted the game room rating for OpFor. This year I was assuming you would do the same for OpFor, clueless, and any new players with no WHR. Not using WHR seems weird after you put so much emphasis on it and also said you were no longer particularly concerned about distortions in the game room ratings because we were going to rely on WHR for seeding all events. If the game room ratings are less accurate, as we suppose, then using them will create more true mismatches even though the number of nominal mismatches will be the same. But anyway, what's done is done. It will be interesting to see whether people with game room ratings higher than their WHR (e.g. Hippo, Eltripas) under-perform their expected scores due to having gotten harder pairings, while people with WHR higher than their game room ratings (e.g. ChrisB, camelback) out-perform their expected scores due to getting easier pairings. Indeed, if we don't observe this phenomenon, then it will indicate that it didn't really matter which ratings you used for pairing. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Apr 12th, 2010, 8:49pm on 04/12/10 at 15:59:07, 722caasi wrote:
Oh, no, don't worry about causing mismatches. Game-hungry players like you will all end up playing each other, because you keep getting paired after everyone else has filled up their game quota against near opponents. Anyone who asks for more than ten games has to know and accept that they will play mismatches, so it is no problem. I am only worried about people who request ten or fewer games and have to play mismatches anyway. Ideally there would be only a few "involuntary" mismatched games. |
||||
|
Title: Re: 2010 Postal Mixer Post by 722caasi on Apr 13th, 2010, 8:15pm on 04/12/10 at 17:02:42, omar wrote:
You must have missed my game against Fritzlein, but I will definitely try my best. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Apr 18th, 2010, 5:39am Almost exactly a week into the tournament, the last player to start (Strategos) made a move in all his games, and the fastest-moving game (Sconibulus vs. Heyckie) finished in 37 moves. Congratulations, Heyckie, on the early tournament lead! The average game is on move 7, although that statistic will no longer make sense now that games have started to finish. Since both players are making one move per day on average, that means we are collectively moving twice as fast as the increment of one move per day per player. |
||||
|
Title: Re: 2010 Postal Mixer Post by omar on Apr 19th, 2010, 9:10am on 04/12/10 at 20:43:34, Fritzlein wrote:
Good idea. I forgot we did this. I've made a note to do this next year. |
||||
|
Title: Re: 2010 Postal Mixer Post by Janzert on May 3rd, 2010, 10:38am For future reference, this morning (May, 3rd) at 4am EST was the first time OpFor did not have a move to think about in this year's postal mixer. This seems like the earliest this has happened in a postal tournament but I can't really remember the dates from previous tournaments for sure. Janzert |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on May 3rd, 2010, 11:16am on 05/03/10 at 10:38:10, Janzert wrote:
So will you increase the thinking time, let the server idle, or use the server for other things? It seems OpFor has more than its original reserve time in every game. And by the way, congratulations on OpFor's 2-1 start this year. |
||||
|
Title: Re: 2010 Postal Mixer Post by Janzert on May 3rd, 2010, 12:06pm The thinking time will stay the same for the whole tournament. What other uses I'll put the freed up cpu time to I haven't completely decided yet, there are some tests I've been wanting to run for a bit (both related and completely unrelated to arimaa). For now I'll probably just let the poor server cool off a little. ;) Janzert |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on May 3rd, 2010, 3:27pm I compiled the statistics to confirm what I thought was the case: openings have gone berserk this year. For the first year ever, unbalanced setups are the plurality, indeed nearly the majority. Of the setups that are balanced, most of them are not symmetrical. Flank elephants, which were extinct for two years, have crept back on the scene, mostly for Silver. Elephants behind a trap have also increased their share from last year, and are being used by both Gold and Silver. The only year on record with fewer centralized elephants was 2006. Having two rabbits forward on the wings narrowly retains a majority, but dips to its lowest popularity ever. No single alternative is taking up all the slack; rather folks are being very experimental about the placement of their forward rabbits, with the most variation in any tournament to date. Opening moves, as well, are more dispersed than ever. The most common pattern, namely moving the elephant forward three steps and one other piece forward one step, claims only one-third mindshare, the lowest winning share ever. At one point I thought my categories were becoming obsolete, i.e. not narrow enough to capture the experimentation that was actually going on. This year the experimentation is so great that even my broad categories show that we are all over the map. If we used a more granular measure, perhaps we would find that this is the first year that no two games were alike after both players set up their pieces. Many people expect that the reason Arimaa has no opening theory is that we haven't studied it much yet. What a surprise, then, that openings are now less standardized than they were in 2005 when I first started keeping track. Elephant File 2005 2006 2007 2008 2009 2010 ------------- ---- ---- ---- ---- ---- ---- d . . . 98.1 79.5 85.3 89.9 90.0 81.6 c . . . . 13.7 9.4 9.5 10.0 13.2 b . . . 1.9 6.3 5.3 . . 5.2 a . . . . 0.5 . 0.6 Rabbits Forward 2005 2006 2007 2008 2009 2010 --------------- ---- ---- ---- ---- ---- ---- ah . . 55.6 56.3 76.5 59.5 57.1 52.4 a . . . 2.5 15.3 4.1 0.6 5.0 12.8 none . . 14.4 22.6 10.0 6.0 2.1 10.8 cf . . 23.1 . . 8.3 1.4 6.6 ac . . . . . 0.6 . 5.2 c . . . 3.1 . . . . 3.8 acfh . . 0.6 . . 14.9 22.1 3.5 af . . . . . . 0.7 2.1 ach . . . . . . 1.4 1.0 acf . . . . . . . 1.0 ad . . . . . . . 0.3 ce . . . . . . . 0.3 adh . . . 2.6 4.1 10.1 10.0 adeh . . . 2.6 4.1 ag . . . 0.5 1.2 abgh . . 0.6 Setup Balance 2005 2006 2007 2008 2009 2010 ------------- ---- ---- ---- ---- ---- ---- Symmetrical . 56.9 50.5 54.1 50.0 35.7 20.1 Balanced . 22.5 30.0 34.1 42.3 42.9 31.2 Unbalanced . 20.6 19.5 11.8 7.7 20.7 48.6 Gold Move 2w 2005 2006 2007 2008 2009 2010 ------------ ---- ---- ---- ---- ---- ---- E up 4 . . 68.8 26.3 25.9 13.1 20.0 22.9 E up 3 over 1 11.3 11.6 1.2 3.6 14.3 9.7 E up 3; X up 1 6.3 41.1 40.0 45.2 45.7 33.3 E up 2; X,Y up 1 3.8 10.5 23.5 27.4 17.1 21.5 E, X, Y, Z up 1 3.8 3.2 . 1.2 . 4.2 Other . . 6.3 7.4 9.4 9.5 2.8 8.3 |
||||
|
Title: Re: 2010 Postal Mixer Post by megajester on May 4th, 2010, 7:39am Wow. That's some great analysis. Just a thought: Is it possible this is happening because of the influx of new and inexperienced players? What if you only analyse the games played by players rated above, say, 1800? Do you get the same result? |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on May 4th, 2010, 10:21am on 05/04/10 at 07:39:38, megajester wrote:
My impression from scanning the games was that in some ways the dispersion of openings is due to inexperience, but in other ways the newcomers were actually more conservative than the old hands. For example, the move 2g of four pieces forward one step each wasn't played by any experienced players, which could explain why that opening was gone last year and back this year. On the other hand, newcomers were more likely to choose symmetrical setups, whereas experienced players were more likely to use unbalanced setups. The unusual placement of rabbits seemed to come more from experienced players, especially the record 1/6 of all setups with exactly one forward rabbit. Decentralized elephants were more common regardless of experience. You might look at the "dumb" forward rabbit in the center in the choices "ad" and "ce", which each occurred for the first time this year, and conclude that those setups, at least, must be the work of ignorant noobs, but you would be wrong. One of them was the choice of an ignorant expert, namely me! ;D The stats would look different if we filtered for experience, but I think they would still reflect the recent fragmentation of opening theory. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on May 4th, 2010, 11:20am on 05/04/10 at 10:21:43, Fritzlein wrote:
But you used it against clueless which ignores possibility of pulling rabbits in the openning. Human could change mind, but the bot's strategy is fixed so you just played anti-clueless, knowing it will not punish the inacurracy. |
||||
|
Title: Re: 2010 Postal Mixer Post by jdb on May 4th, 2010, 1:09pm IN my opinion, ngerhart used the best opening of any against clueless. He has the best position of anyone, so far. |
||||
|
Title: update Post by Fritzlein on Jun 22nd, 2010, 10:25pm 87 games have completed, almost three-fifths of the slate even throwing out the 8 games lost on time before they got started. Only two players, megajester and camelback, have yet to complete a game. In addition to Simon and Wyntre who timed out all their games, six participants have played all their games to completion: azgreg, Korhil, Sconibulus, student_t_w, toby1kenobi, and willwould. This is a large percentage of completed games given that only 77 of the allotted 300 days have elapsed. Every year I fear that some slow game which drags on for a hundred moves will be arbitrarily decided by the 300-day time cutoff, and every year my fears prove unfounded. The current lead of +5 is shared by chessandgo and Nombril, each with five wins and no losses. Nombril, however, has a better chance of winning the overall tournament, because he signed up for ten games total while chessandgo signed up for only six total and thus can't finish above +6. Last year a score of +7 was required for victory, and that was the lowest winning total ever. Another inducement to sign up for many games! OpFor and clueless are both acquitting themselves very well so far. OpFor's victory over Tuks would have been triumph enough, but then OpFor took the biggest scalp ever by a bot in a postal tournament, defeating Adanac in just 30 moves. Clueless is 5-1 so far, at +4 only a point behind the tournament leaders. Also clueless took advantage of a couple of weak tactical moves on my part to win a dog from me for nothing. I may have recovered enough that my position fully compensates for my lost material, but clueless still has potential to take me down in a tense position. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Jun 23rd, 2010, 3:55am Yes, clueless plays rather well :) and each completed game of it reduces it's potential problems with lowered time reserves (I am already horse down, but still with hopes for improvements). Actually ... clueless will be leading tonight, but just for a while if I correctly predict its moves. Nombril's chances to remain around best three are really high. Interesting, how many directed cycles in results were already created. BTW: While I was writing this, clueless swarmed against me even with cats ... interesting :) I would be really interested on bot's evaluation scores of all positions obtained in their games. I hope both of them did log such informations and I hope it is not secret :) |
||||
|
Title: Re: 2010 Postal Mixer Post by chessandgo on Jun 23rd, 2010, 4:44am impressive performance by both bots ... |
||||
|
Title: Re: 2010 Postal Mixer Post by rbarreira on Jun 23rd, 2010, 5:22am jdb and Janzert: How did you set up the bots to play many games simultaneously? Some sort of automatic script, or do you manually go and start up the bot for a specific game every once in a while? About how long are the bots thinking for on each move? |
||||
|
Title: Re: 2010 Postal Mixer Post by jdb on Jun 23rd, 2010, 8:22am on 06/23/10 at 03:55:22, Hippo wrote:
I have the log file for clueless for most of the postal tournament. The first couple weeks or so are missing. If someone reminds me, I'll post it after the games are all finished. Clueless' search is getting between 17-20 steps in the 4 hours it thinks for each move. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Jun 23rd, 2010, 9:28am @jdb: Do you have separate log files or just one common? In the former case it would be nice to publish each one after the game is finished. In the later case it would be nice to have parser allowing to convert it to the former case :). |
||||
|
Title: Re: 2010 Postal Mixer Post by jdb on Jun 23rd, 2010, 10:32am on 06/23/10 at 09:28:47, Hippo wrote:
Its all one big log file. It is easy enough to parse, but not by me. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Jun 23rd, 2010, 11:24am @jdb: OK, than I would be looking forward to the last clueless postal mixer game finished. |
||||
|
Title: Re: 2010 Postal Mixer Post by 722caasi on Jun 23rd, 2010, 1:56pm To whoever runs this tournament: My mother would like to know if my ban from the Arimaa World League also bans me from this tournament. Please reply ASAP. PS: Who runs this tournament, anyways? The official page, here (http://arimaa.com/arimaa/tourn/postal/2010/), doesn't say. |
||||
|
Title: Re: 2010 Postal Mixer Post by jdb on Jun 23rd, 2010, 3:23pm on 06/23/10 at 13:56:15, 722caasi wrote:
I would guess that the world league has no effect on the postal mixer. |
||||
|
Title: Re: 2010 Postal Mixer Post by omar on Jun 23rd, 2010, 7:53pm on 06/23/10 at 13:56:15, 722caasi wrote:
Most everyone qualified to be a director tends to play in the postal mixer so for many years this event has slipped by without having an official director :-) From what the world league coordinator and director stated I think it is can be concluded that the penalty does not impact your participation in any event outside of the league. "Because this is a case of willful cheating, the League Director and I have agreed to declare 722caasi's game a forfeit, and bar him from any participation in this League and any other League starting in 2010." So no ruling by the postal mixer director is needed :-) |
||||
|
Title: Re: 2010 Postal Mixer Post by Janzert on Jun 23rd, 2010, 7:59pm I use the script included with the AEI software package called "postal_controller.py". Basically it connects to the gameroom, checks for any postal games that are on the bots turn, and then has the engine play in each of them. It does a pretty good job but does occasionally crash or go haywire if it encounters connection problems. Janzert |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jun 23rd, 2010, 11:33pm And just while we are praising clueless, it pulls into a tie for the lead by beating Eltripas in this game (http://arimaa.com/arimaa/gameroom/comments.cgi?gid=147301). Six wins and one loss leave clueless at +5, with nine games to play, and thus plenty of chance to top the standings by tournament end. If you had showed me the first sixteen moves of this game four years ago and asked me which side was played by the bot, I would have said Silver was the bot, because Gold was playing with a plan, namely the camel hostage strategy. It is sobering to see computer behavior that appears plan-driven. It is sobering also to see a computer get an edge in the opening, not due to a tactical blunder by the opponent, but rather due to the accumulation of positional advantage. For the present I can take comfort in the fact that clueless still has many strategic blind spots. There are still multiple ways to exploit its weaknesses. But I am reminded that there is no reason in principle that a bot shouldn't be able to do what I do strategically. It's a hard problem, but not an impossible one. The more success bots have, the more I cast an eager, nervous eye to 2011. The progress of the bots feel inexorable, while the progress of the human community in understanding more feels speculative. How long will we be able to stay ahead? The march of the deathly machines makes me all the more grateful for the Arimaa World League, and for Jean's forthcoming book, and for the planned Arimaa Festival. We need to share our knowledge, cross-pollinate our strategies, and in all ways work together to demonstrate the power of our biological brains. For now it is up to the biological brains of JDB and other developers to grasp communal Arimaa strategy knowledge and translate it into machine-comprehensible formulas. This is our margin of safety. If Omar gets his way and computers start learning Arimaa strategy on their own, the end will be at hand. |
||||
|
Title: Re: 2010 Postal Mixer Post by rbarreira on Jun 24th, 2010, 12:38am Fritzlein would you say that if a bot could play like this in a game with 2 minutes per move, it would have a chance to win the Arimaa challenge? Given that humans are probably thinking for longer than 2 minutes about their moves. Anyway I don't think you have to worry much about 2011. If clueless and Opfor are thinking 2-4 hours, that's 6 to 7 doublings of CPU power, which is a lot to overcome by improving the software. But my guess is that with improvements to the bots, and the ever-accelerating hardware, by 2020 bots will probably be playing lot better in the challenge than clueless and Opfor are playing right now in the postal tournament. Of course as you said, there's always the even more uncertain question of human improvement... |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jun 24th, 2010, 12:49am on 06/24/10 at 00:38:37, rbarreira wrote:
A chance, yes, but I think not quite. I suspect the Challenge is still safe from this level of play. 7 CPU doublings should give the bots in the neighborhood of 500 rating points, plus or minus a lot. Chessandgo is still more than 500 rating points ahead, although for defenders below him it gets cloudier. My point is that when I look back I remember thinking, "It will be difficult for a bot to do X," and now a bot can do X. X is the discovery of the past, and our current strategic understanding has moved well beyond that. Today I feel quite comfortable saying, "It is still difficult for a bot to do Y." What I am nervous about is only the future, given the past hurdles bots have overcome. Clueless' performance in the challenge screening was 1960. Its performance in the Postal will depend on the number of games won out of 16: Wins Performance ---- ----------- 7 1945 8 2008 9 2072 10 2139 11 2210 12 2287 From the table, JDB's doubling CPU one more time to four hours per move instead of two should only give clueless one extra win. Also, humans should gain far more per doubling than clueless does, so even if humans think only, say, 16 minutes per move they are getting four doublings, which should wipe out all the gains of clueless' long thinking time. However, my intuitions and calculations about what should happen appear to be off in this instance, because they don't match the excellent tournament clueless is having. Clueless could easily end 12-4, with a score of +8 that could win the tournament! For me thinking 16 minutes on a move is about average, but I have heard some humans (Eltripas? Tuks?) saying they think far less than that. There may be other explanations of clueless's excellent performance as well, including a statistical fluke, and including that my formulas about the value of doubling thinking time for humans and for computers are completely bogus. ;D |
||||
|
Title: Re: 2010 Postal Mixer Post by jdb on Jun 24th, 2010, 10:54am Having clueless play in the postal mixer has been very informative for me. I have a much better idea what needs to be done to improve things. There are things that show up at postal speeds that do not show up with shorter thinking times. About 95% of the time it is sufficient to be on the lookout for enemy goals in two. However there are times when it is required to look for enemy goals in 3. In order to have a chance at the challenge, a bot will need to be able to detect enemy goals in 3. Once a bot has that level of goal awareness, the rest of the game boils down to trap control. Clueless got out manouvered in trap fights in several games in the postal mixer. This will be harder to fix. Adanac is very good at this aspect of the game. I think I need to improve my own understand of trap fights before clueless will make progress in this area. |
||||
|
Title: Re: 2010 Postal Mixer Post by RonWeasley on Jun 25th, 2010, 5:03am on 06/24/10 at 00:49:15, Fritzlein wrote:
I think for far less than that, even after pondering a position for days. There must be a think/ponder ratio that is higher for some My goodness Hermione does wonders for the school uniform. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jun 26th, 2010, 5:38pm on 06/25/10 at 05:03:04, RonWeasley wrote:
I guess that I don't have such a high average thinking time if you don't count thinking away from the board. I am often considering what move to make at odd times, such as in the shower or walking to the bank. I would say that about half the time I don't make a move at one sitting, but let my thoughts percolate and come back to the position later. It is surprising how often I have a new perspective given a second look. For example, I have just seen Hippo's latest move, and have considered it for about five minutes. I have a strong move candidate, and could just make the move now if I were pressed for time. I am not pressed, though, so I will go shopping instead, and think about the position as I drive. When I come back, I may see new candidate moves that I didn't before or see crushing responses for him that I didn't see before. If nothing jumps up to change my mind, I'll make the move; otherwise I'll think longer. So how you add up my total time per move definitely depends on whether you count the thinking while shopping. I tend to think longer on games that I am losing (such as my game with Hippo), and less on games where I have the position under control. I notice from woh's rating lists that my postal ranking is 132 points ahead of chessandgo's whereas my overall ranking is 169 points behind chessandgo. I attribute this entirely to the fact that I think longer than average on my postal games. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jun 27th, 2010, 12:38am As a footnote to my opening analysis, I went back and found duplicate positions after both Gold and Silver had set up. In 144 positions, there were 138 unique positions and 6 duplicates. Half of the duplicates were variants of the 99of9 setup against each other. After 2g only three duplicates were left, and after 2s, no duplicates were left. The duplicates were bot_clueless 722caasi woh 722caasi diverge 2s starjots 722caasi starjots willwould diverge 2g 722caasi Weirdo87 azgreg 722caasi diverge 2s Heyckie ocmiente RonWeasley Eltripas diverge 2g Nombril Tuks 722caasi Tuks diverge 2g Heyckie amalgam Korhil amalgam diverge 2s This total scattering of opening setups should be enough in itself to discourage anyone from trying to codify openings, but another fact makes the situation worse. I looked up my previous research from past postal tournaments and found that opening are less standardized now than then: In 2006, there were 3 games out of 95 that had not yet diverged after 2s. After 3g there were no duplicates. In 2007, there were 4 games out of 85 that had not yet diverged after 2s. After 3g there were no duplicates. So in 2010 we diverged sooner than ever. Quote:
Isn't that amazing? Two moves on each side (or rather the setups plus one move each) were enough to distinguish every game in this tournament, i.e. 144 out of 144. But in a chess tournament two moves on each side only led to 20 unique positions out of 91 games, i.e. 78% duplicates. I think opening theory is one aspect in which we can just stop expecting Arimaa to be like chess. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Jun 27th, 2010, 6:51am on 06/26/10 at 17:38:11, Fritzlein wrote:
OK, let it be noted you said that after 40g :) I hope you just are not paying mind games ... :). I am happy to hear this after your 40 turns, but my potential lead is such small that I hardly notice it. And each turn I am losing 5 days of reserve and probably position as well :). And congrats Nombril for returning to the lead. I suppose today (1.7.2010) clueless would lose one of the games ... |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jul 1st, 2010, 12:54pm on 06/24/10 at 10:54:14, jdb wrote:
Congrats on beating clueless, Tuks. This is a game that suggests seeing goal in three isn't enough, at least not for postal games. http://arimaa.com/arimaa/gameroom/comments.cgi?gid=147834 I suppose, though, that I would never have the guts to sacrifice a rabbit for that game-ending goal attack, even if I had two minutes per move to think about it. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Jul 1st, 2010, 1:20pm Yes congrats Tuks ... this is the game I was watching hoping for win in 3 and I was put 2 turns back in history:) I was scared if clueless have seen some defense I didn't .... And grrr ... I have already made 2nd blunder against clueless (36g) ... with 73 days in reserve. Just when I start feeling I am getting out of troubles (what is the most dangerous feeling in arimaa). |
||||
|
Title: Re: 2010 Postal Mixer Post by jdb on Jul 3rd, 2010, 5:33am on 07/01/10 at 12:54:28, Fritzlein wrote:
Well done Tuks. Nicely played. There was a goal in 4 earlier in the game. I'll leave it as an exercise for the reader to figure out where it was. Today is the first time clueless caught up in all its postal games. |
||||
|
Title: Re: 2010 Postal Mixer Post by Tuks on Jul 3rd, 2010, 10:00am thanks guys, one of my few postal games where i had a strong position throughout. Yes Fritz, i play postals at normal game times, the first 10 moves of each game is usually 30 seconds per move and its never much more than 2-3 minutes throughout a game unless its either really complicated or i'm losing badly :) |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jul 4th, 2010, 6:30am on 07/03/10 at 05:33:59, jdb wrote:
Let me ask you what I asked Janzert about OpFor: catching up gives you the option of increasing clueless' think time. Will you try to challenge humanity even more by keeping the server maxed out with longer thinks? Or will you let the server cool down and/or use it for other purposes? I personally am glad you set it for four hours in the first place, because clueless came up with some excellent tactics that put quite a scare into me. Even now that I think I have regained control, I am treading very cautiously because clueless has proven it deserves more respect than I was giving it at first. |
||||
|
Title: Re: 2010 Postal Mixer Post by jdb on Jul 5th, 2010, 5:38am on 07/04/10 at 06:30:31, Fritzlein wrote:
I need my computer for other things, so it will stay at 4 hours per move. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jul 5th, 2010, 12:01pm on 07/05/10 at 05:38:35, jdb wrote:
Thanks for having devoted that server for so long already so that clueless could play in the Postal Mixer. It has been a treat for me, and I think for others as well. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jul 9th, 2010, 1:40am on 06/23/10 at 08:22:44, jdb wrote:
I'd be interested to see clueless' evaluations throughout my game, in particular around move 19 when clueless threw away its considerable advantage. How long was it before the bot knew it had messed up? I congratulate clueless for tactically outplaying me on a couple of crucial moves. After that, however, my opinion of clueless' ability was inflated so much that I was shocked at its helplessness the rest of the game. Clueless was completely incapable of cutting its losses or stopping the bleeding in any way. In contrast, I realized after the fact that my 15g was a blunder, and on 16g I gave up the camel hostage I had been holding to try to limit the damage. Clueless saw its forced loss on move 32, I believe, but between move 19 and move 31 its evaluation must have deteriorated by about a rabbit per move. That's what I call going down in flames. I'd like to see that plot of evaluation versus time to see whether it was a steady slide to oblivion, or case of clueless remaining in denial for many moves before falling off a cliff. Either way I'm not sure how it would be fixed. Funny side note: Clueless pulled my rabbit into hostage position at a6 on move 17s and didn't have time to take it until move 35s as a "spite capture" just before I goaled. http://arimaa.com/arimaa/gameroom/comments.cgi?gid=148577 |
||||
|
Title: Re: 2010 Postal Mixer Post by jdb on Jul 9th, 2010, 12:09pm Good job Fritz. Clueless saw the forced loss on move 31b. It figured it was pretty much doomed for several moves before that. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jul 9th, 2010, 12:39pm on 07/09/10 at 12:09:05, jdb wrote:
Ah, thanks. In that case I probably had a goal in four that I missed? Not that I feel too bad about missing a goal in four in exchange for a goal in five. Or was the forced loss clueless saw on 31s a goal in five? That would be an amazing search depth even for a four-hour think. Quote:
I'll hit you up for the log file after the tournament so that I can see the exact decline in clueless's expectations. If there was a sudden drop in evaluation, one could look to see whether it came the move after a blunder, but I have a hunch it didn't. My guess would be that clueless' eval dropped only slightly between the blunder of 19s and its first chance to realize it on 20s. Meanwhile I'll guess that the biggest drops in eval came after moves in which clueless did something reasonable but it was too late. Thanks again for entering clueless in the tournament. I expected a cakewalk but got a learning experience and a caution for the future. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Jul 19th, 2010, 2:13am So the most important games for the postal mixer are Nombril/ChrisB and as expected chessandgo/Fritzlein ... the games between undefeated players with lot of wins so far. |
||||
|
Title: Re: 2010 Postal Mixer Post by chessandgo on Jul 19th, 2010, 4:25am on 07/19/10 at 02:13:27, Hippo wrote:
I fear there is less and less suspense for the latter with each new move ... |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jul 19th, 2010, 6:36am Jean, I will optimistically interpret your comment to mean you are despairing of saving your wayward central rabbit. :) As an interesting footnote to our game, I note that all three material evaluators (DAPE, FAME, and HarLog) say that I would be hurt to trade H for DR and I would be hurt to trade DR for H. In other words, they all reflect in some way my own rule of thumb that the player with the deputy doesn't want piece trades. This is confirmation that material evaluation formulas have surpassed anything that can be done with static piece values. Hippo, you are too modest. You, too, are undefeated, and the position in our game is more tense and balanced than the position in my game against chessandgo. Unless your comment coincided with your decision to take our game off the "games to watch" list? If so, I accept your psychological resignation. As Bobby Fischer once said, "I like the moment when I break a man's ego." ;) |
||||
|
Title: Re: 2010 Postal Mixer Post by chessandgo on Jul 19th, 2010, 11:41am Make that a *Horse* and you'll be quite right :) Btw, I think indeed that trading your horse for DR would be great for me (how about when I've lost an extra Rabbit though?), I'd guess I might be somewhat indifferent to trading dr for H ... how bad do the Evaluators evaluate the latter trade? |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Jul 19th, 2010, 11:49am Thanks Fritz :), in current time I am fighting not to finish 2:6. I have more time now, but it seems to me more time I think about the games worse result it gives :(. In our game the 39s was much weaker then I expected, as well as 40s. Especially the 37g demonstrated your powers. I would probably delay playing that game for a while :) not to make too weak 42s :). At least we would be rather equal for longer time if not for longer number of moves ;) :D. And of course I cannot comment your game with c&g :) or make any predictions :) |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jul 19th, 2010, 3:26pm on 07/19/10 at 11:41:30, chessandgo wrote:
If you get my H for DRR, I will be slightly happy. You would be so depopulated (eight pieces left to my ten) that I could hardly make a bad rabbit advance from then on, even though my camel is overloaded trying to contain your HD so I would have to centralize it on defense. The material evaluators agree with me, I think mostly because you would be down to three rabbits and they value those late rabbits even more highly than I do. Here is what they think the net value of the two trades would be, from your perspective: H for DR DAPE: +0.27 FAME: +0.70 HarLog: +0.55 H for DRR DAPE: -3.60 FAME: -0.81 HarLog: -1.44 That second DAPE score is not a typo; its evaluation swings almost four points when you take off your fourth-last rabbit in that situation. Quote:
Again, from your perspective: DR for H DAPE: +0.50 FAME: +0.14 HarLog: +0.86 I tend to disagree with your indifference and agree with the formulas, because the value of the deputy lies greatly in the ability to make two simultaneous material threats. As the board depopulates, the opposing elephant has an increasingly easy job to take the threat the deputy is making and turn it into a threat against the deputy itself, if not into an immediate goal threat. A supporting cast of small pieces is necessary to make the deputy effective. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jul 19th, 2010, 4:48pm on 07/19/10 at 11:49:00, Hippo wrote:
Thank you! At that point I felt I was teetering on the edge, barely holding my position together, but now that my rabbit has wandered from c1 to h4 my position feels much more tenable. Quote:
It is my great advantage to be unemployed and therefore able to think as long as I like about every move. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Jul 21st, 2010, 5:41pm on 07/19/10 at 06:36:19, Fritzlein wrote:
Hmmm, interesting. Fritz psychic powers made the rabbit jump to the trap itself ... |
||||
|
Title: Re: 2010 Postal Mixer Post by chessandgo on Jul 22nd, 2010, 8:11am on 07/21/10 at 17:41:19, Hippo wrote:
haha! Kinda handy, isn't it? :) Thanks for the numbers Karl! I guess I'm probably wrong in still viewing H as strong as dr. And I meant the h for DR exchange with an other gold Rabbit down (as is now the case). Indeed, I'd take a drr for H exchange basically no matter what (say, unless a pair of elephants and a pair of camels have been traded^^). |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Jul 22nd, 2010, 5:17pm I am not sure where to write it ... here or to the site discussion ... The order of columns in "view games" is alphabetical, but the order inside of columns seems to be random. I would be much more happy with ordering inside of columns by descending time of finish. So unfinished games would be shown at the top and first finished one at the bottom ... May be even better but more dynamic ordering, which is generalisation of the mentioned one ... by the time of the last move played. I don't know how big complication it would cause ... |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Jul 23rd, 2010, 12:07pm One thing should be mentioned and was not yet .... here alheady is one player with perfect record: novacat won 4 of 4. Congratulations |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jul 23rd, 2010, 1:27pm Yes, well done novacat! Too bad you only got to play three of your four games. :( Abandoned games were the reason we initially instituted an entry fee, but it seems that for the tournament overall, most games are being played to a natural conclusion, so perhaps abolishing the entry fee has worked by and large. Hippo, it appears the games are displayed in the order that they were created by the pairing algorithm. I, too, think this could be improved on, perhaps from fewest to most moves played, with completed games at the bottom. Actually, what would be most convenient is to list at the top the games in which one has the least time remaining, but Omar explained to me that this is technically difficult for some reason. And there is some virtue in having an order that doesn't change, although if the order is fixed I think I would prefer it to be by opponent rating. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Jul 23rd, 2010, 3:43pm OK, if the order should remain static, I would prefere the same order inside columns as the order of columns ... so either both alphabetically or rather both by rating. Omar please ignore my proposals whenever you want. You have created beautiful site and you have a lot of other stuff to do. I fully respect it. I just hope they could be helpful to have them in mind when working on stuff they are addressing. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Jul 25th, 2010, 6:42am OK so finally clueless seems to not win the postal mixer :) this year. I would be interested on the move where clueless started consider its position to be losing (in my game). I am almost sure the turning point was when it grabbed my rabbit instead of profitting from my "blunder" allowing MR/mc exchange with freeing both hostages meanwhile. But I suppose it was around 3 turns later for clueless. http://arimaa.com/arimaa/gameroom/comments.cgi?gid=149848 |
||||
|
Title: Re: 2010 Postal Mixer Post by jdb on Jul 25th, 2010, 8:46am It looks like Adanac will be done with clueless shortly. The two remaining games with woh could take a while. If anyone wants clueless' log file from the tournament, I'll email it to you after Adanac's games are done. |
||||
|
Title: Re: 2010 Postal Mixer Post by novacat on Jul 25th, 2010, 8:49am Thanks Hippo and Fritz. Congrats to knarl too for finishing undefeated as of yesterday! Next year I will try playing more games. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jul 25th, 2010, 10:57am Well done, knarl! |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Jul 25th, 2010, 11:05am on 07/25/10 at 08:46:08, jdb wrote:
As I have already written I would like to look at it :), thanks. And yes, kongrats kanrl to the perfect record. |
||||
|
Title: Re: 2010 Postal Mixer Post by knarl on Jul 25th, 2010, 6:49pm Thanks guys, but I only played 3 games, and only scored one goal :P . But it will encourage me to play more games next year. Cheers, knarl. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jul 26th, 2010, 12:34pm on 07/25/10 at 08:46:08, jdb wrote:
Thanks for the offer, JDB. Adanac has just triumphed in both games. Sign me up (along with Hippo) for the log-file distribution list. |
||||
|
Title: Re: 2010 Postal Mixer Post by Janzert on Jul 26th, 2010, 2:30pm on 07/25/10 at 08:46:08, jdb wrote:
I would be happy to put these up on my site for you if there is enough interest that you don't want to have to send them out individually. I'm also willing to make the same offer to release OpFor's logs. But I think Opfor's will be much harder to look through since it is simply a file per move named by date and time. Janzert |
||||
|
Title: Re: 2010 Postal Mixer Post by Janzert on Jul 26th, 2010, 6:49pm Clueless' postal log can now be downloaded compressed as a zip (http://arimaa.janzert.com/clueless-2010postal.zip) or gz (http://arimaa.janzert.com/clueless-2010postal.log.gz) file. Janzert |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Aug 3rd, 2010, 10:35am Looks like Nombril is on a long trip. 3.5 days and he starts losing on time ... :( |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Aug 3rd, 2010, 12:52pm on 08/03/10 at 10:35:59, Hippo wrote:
He was in the chat room yesterday. |
||||
|
Title: Re: 2010 Postal Mixer Post by Nombril on Aug 3rd, 2010, 1:01pm Back from a long trip, plus we have moved, so I don't have many extra minutes right now. I am back, so will lose because I am behind, not because of time! |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Aug 3rd, 2010, 1:50pm Yes, it seems to me, time is not the main problem. ... Making single move in blitzspeed against Fritz, Ron, The_Jeh ... in the postal games means losing ... :). I hoped for 60 or more turn loses, but it would be surely shorter :(. Nombril, please don't lose on time ... |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Aug 19th, 2010, 12:03am Woh has just beaten Opfor to finish OpFor's 2010 Postal Mixer with a 7-8 record. For calculating performance rating, however, I will discard OpFor vs. Simon since it timed out on move two. I know it is a judgment call to exclude this game while including Bomb vs. MrBrain from 2005, which MrBrain abandoned mid-game, but the latter game was significantly progressed. OpFor clocks in at a performance rating of 1848, which is exactly its pre-tournament gameroom rating! This is further evidence that OpFor's stunningly strong performance in 2008 was a statistical outlier. Thanks, Janzert, for entering Opfor in the tournament three years in a row for our enjoyment and education. Clueless has two remaining games, both against woh, and both could still go either way, so we can't quite close the book on clueless' tournament yet, but the range is given below. Given the swing in performance rating caused by a single game, it is no wonder that performances have been all over the map in past events. Bot . Year Record Postal Performance --- ---- ------ ------------------ Bomb 2005 4-6 1716 Zombie 2007 1-14 1409 Sharp 2008 1-8 1568 OpFor 2008 8-9 2060 OpFor 2009 3-7 1798 OpFor 2010 6-8 1848 Clueless 2010 7,8,9 1945, 2008, 2072 |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Aug 19th, 2010, 2:33am Jdb, thanks again for clueless' log file; I finally got around to analyzing it. For those of you who want to follow along with my commentary, here's a link to the game (http://arimaa.com/arimaa/gameroom/opengamewin.cgi?gameid=148577&role=v&side=w). Between the start of the log on 4s and the point at which clueless saw a forced loss for itself on 31s, it correctly predicted my move only three times, namely 8g, 23g, and 31g. The first of these was me taking a hostage frame, and the second two were me making captures. A success rate of 3/28 = 11% at predicting my move would seem to argue against the chess strategy of pondering while the opponent thinks by assuming the predicted move will be made. The first surprise in the logs is move 11s, when clueless didn't think I would take its camel hostage on 12g. It seemed to repent immediately, however, as its evaluation dropped by nearly a rabbit between 11s and 12s. My second surprise was discovering that clueless didn't consider my move 15g to be a blunder, only my move 16g. My analysis was that 15g turned a fine position for me into a terrible one and that my 16g was my best desperation try to avoid losing two horses for a camel. Clueless, however, expected 16g E<HM<E>, and I confess that it appears to let me keep the camel hostage without losing material. Clueless can reverse the move, of course, but I win the undo war. So it seems that I bungled tactically in more ways than I knew. I beleive that 19s was a critical error by clueless, allowing me to take its camel hostage because it was avoiding a dog frame in f6. This is a serious strategic misjudgment; the dog frame would have been worth far less to me than the camel hostage was. I expected that clueless considered itself to be winning comfortably as it played 19s, but the situation was more complex. Yes, clueless quite liked its move 19s when completing a 4-ply search, but at depth 17 steps it lost confidence in its move, thinking it was winning by less than a rabbit rather than by a whole dog. However, it didn't find any better move at depth 17 steps before running out of time. This may be a case where clueless' four hour thinking time was just a smidgin too short, luckily for me. My position was at least equal after 19g, possibly even winning slightly, but clueless didn't think it was losing until playing 22s, by which time it in a very poor situation. Clueless' evaluation then rapidly deteriorated until 25s, when it offered to give me MH for M, and fully expected me to take the trade. By then, however, I liked my position too much to accept being up by only a horse. On move 26s, 27s, and 28s, clueless deluded itself into thinking that its horse was somehow going to rescue its camel. I calculated once, before move 25, that the rescue wasn't going to happen, and after that it became part of my "quiescence search". This seems like a big advantage in understanding that I had relative to clueless. I knew all along I wasn't going to lose control of f3, and that I would eventually score at least a free camel there. Clueless, meanwhile, even when making 28s didn't expect me to immediately re-establish the hostage on move 29g, stronger than ever. Serious blind spot. From there the game is completely over, and interesting only insofar as I might have forced goal a move or two sooner. Clueless, which had been feeling down by only a dog or so, suddenly had its eval crater on move 29s and moreso on 30s. I don't feel too bad if I failed to execute the last few moves perfectly, because I feel very good to have anticipated the eventual outcome at least five moves before clueless realized its overloaded elephant was never going to save the day. This kind of analysis makes me more respectful of clueless in some ways, more afraid of its awesome powers, but in others ways smug about its remaining weaknesses. The man vs. machine drama continues... |
||||
|
Title: Re: 2010 Postal Mixer Post by omar on Aug 21st, 2010, 7:49pm Awesome analysis Karl. Thanks for sharing this with us. It was really neat following this game with your analysis. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Aug 23rd, 2010, 5:24am Seems according the score system for postal mixer I won the uphill battle and my game against The_Jeh lost on 12g will last 60 turns. I hope John was not bored too much by this and it was in the postal mixer spirit :). I am not sure I would make such an effort to other lost games. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Aug 23rd, 2010, 8:44am on 08/23/10 at 05:24:05, Hippo wrote:
Apparently ever since I won a rabbit against The_Jeh, he has been in delaying mode. I am scarcely better off now than I was a month ago, even though I haven't been in any danger either. Perhaps you will sleep better knowing that what comes around goes around. :) Fighting to the death is very much in the spirit of the Postal Mixer. Not only is there satisfaction just from playing hard, there is also always the possibility that your opponent will slip. If I blunder on turn 87, allowing The_Jeh to come back and win on turn 113, his tenacity will have been well-rewarded! |
||||
|
Title: Re: 2010 Postal Mixer Post by rbarreira on Aug 23rd, 2010, 11:41am on 08/23/10 at 08:44:29, Fritzlein wrote:
I thought that actually went for Arimaa in general, since people are discouraged from resigning? |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Aug 23rd, 2010, 11:54am on 08/23/10 at 11:41:11, rbarreira wrote:
True, but in the Postal Mixer you get points for fighting to the death. The prize money used to be distributed according to the number of points (not the number of victories), which initiated a tradition of hanging on for as many moves as possible. Perhaps nowadays the points are a less of a motivator because there is no prize money. |
||||
|
Title: Re: 2010 Postal Mixer Post by speek on Aug 23rd, 2010, 1:21pm I don't understand the denigration of resigning. As a chess player, it seems far more honourable to resign a lost position than play on tediously, hoping for an opponent to slip up. |
||||
|
Title: Re: 2010 Postal Mixer Post by rbarreira on Aug 23rd, 2010, 1:50pm on 08/23/10 at 13:21:01, speek wrote:
I think it should be more of a personal thing. You should resign if/because you don't want to play anymore, not because you don't want to force the other guy to play. Waiting for an opponent to slip up seems fine by me, after all you probably got into the lost position due to your own slip up. Just because you slipped up first doesn't mean you should throw away the game if you think the other guy might blunder. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Aug 23rd, 2010, 5:09pm on 08/23/10 at 11:54:02, Fritzlein wrote:
Not exactly ... you get points for fighting to the 60th turn :). Which I am interpretting as the turning point good enough for resignation. (John let me chose to selfmate on 60 or to be killed on 61. It seems to me he missed goal in 2 at leasst on 59g.) |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Sep 7th, 2010, 4:01pm Seems "winner" of the postal mixer didn't played Fritzlein :) |
||||
|
Title: Re: 2010 Postal Mixer Post by Nombril on Sep 7th, 2010, 10:55pm on 09/07/10 at 16:01:47, Hippo wrote:
Who are you declaring Winner? Fritz elected to play 11 games instead of the 10 games that Chris and I picked... that could very will be the edge needed to get the largest win-lose. And the bots have the most points. (Interesting to note that Clueless had a better win record, but Opfor had lasted longer so has more points.) And if you are looking at me, I think my remaining games are not decided, very close in 2, and behind in 1. But I will certainly be stubborn and try to win a few more... |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Sep 8th, 2010, 12:33am I am just kidding :) to remember the mixer is still on :). |
||||
|
Title: Re: 2010 Postal Mixer Post by Nombril on Sep 8th, 2010, 6:30am Ah, I see. There is a certain kind of stamina to play so many postal games! After starting with 60 days, it almost feels like there is some time pressure now as the reserve is disappearing. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Sep 8th, 2010, 6:37am I have reached point where thinking about postal mixer games is meaningless about two weeks ago ... so I do not have problems with reserve time now :). But it was tense in the end of july (when I had some hopes in Adanac and Fritzlein games). |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Sep 11th, 2010, 6:50pm So welcome next undefeated player ... really impresive. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Oct 8th, 2010, 1:13am I see that clueless triumphed in both of its games against woh, for nine wins on the tournament. That makes for a new bot postal performance record. Congratulations, jdb! Bot . Year Record Postal Performance --- ---- ------ ------------------ Bomb 2005 4-6 1716 Zombie 2007 1-14 1409 Sharp 2008 1-8 1568 OpFor 2008 8-9 2060 OpFor 2009 3-7 1798 OpFor 2010 6-8 1848 Clueless 2010 9-7 2072 |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Oct 8th, 2010, 1:22am I count six unfinished games: ocmiente vs. ChrisB amalgam vs. ocmiente Nombril vs. Harren Weirdo87 vs. starjots Eltripas vs. woh ChrisB vs. RonWeasley As usual, I was worried that the time control would lead to an unnatural conclusion of some game, and as usual I appear to have worried needlessly. |
||||
|
Title: Re: 2010 Postal Mixer Post by Hippo on Nov 1st, 2010, 7:07pm Next undefeated player finished ... CONGRATULATIONS. Nombril's arimaa abilities are amazing for being 13-14 months in the comunity. Next year I suppose he would be paired with Fritzlein (not to have 2 undefeated players with the number of wins consisting of 2 digits). |
||||
|
Title: Re: 2010 Postal Mixer Post by omar on Nov 1st, 2010, 10:37pm Wow, congrats Eric. You are definitely a rising star. I wish my rating graph looked like this: http://arimaa.com/arimaa/gameroom/graph/bh/rating_plot/?Nombril |
||||
|
Title: Re: 2010 Postal Mixer Post by Nombril on Nov 2nd, 2010, 7:30am Thanks! I definitely benefited in the postal mixer by being in the middle of my learning curve at the start, so part way through the game period many of my pairings had become favorable to me. I owe the continued ratings climb to all of the great analysis and discussions that have been available from the community. (Fritz's book and commentary, Adanac's write ups about the WC games, and everyone that is willing to do post-mortum analysis with me after a game.) |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Nov 29th, 2010, 1:59pm Well done, Nombril. When ocmiente vs. ChrisB finishes, we should do a post-mortem on the tournament, including everyone's performance vs. expectation, the number of abandoned games, and so on. |
||||
|
Title: Re: 2010 Postal Mixer Post by ocmiente on Dec 24th, 2010, 5:09pm The longest duration game (http://arimaa.com/arimaa/games/jsShowGame.cgi?gid=165608&s=w) of the 2010 postal mixer is finally over. |
||||
|
Title: Re: 2010 Postal Mixer Post by omar on Dec 25th, 2010, 10:56pm Wow, ChrisB really put up a long fight after having lost the horse on 47g. I'm guessing that horse lose was a slight miscalculation. Good game guys. |
||||
|
Title: Re: 2010 Postal Mixer Post by ChrisB on Dec 26th, 2010, 3:47am on 12/25/10 at 22:56:30, omar wrote:
Yes, a slight, but costly miscalculation! ..... largely brought about by ocmiente's strong, solid play and my disastrous attempt to try to generate some kind of counterplay. Now, that the postal mixer is over, I'm planning to post in the next week or so a couple of performance vs. expectation rankings, similar to what Fritzlein did after the 2009 postal mixer. If others have been working on this also, I would be curious to see how our results compare. |
||||
|
Title: Re: 2010 Postal Mixer Post by Fritzlein on Jan 7th, 2011, 2:29pm Final tally on games mismatched and unfinished: Year . . 2005 2006 2007 2008 2009 2010 Players . . 16 19 20 15 14 42 Games . . 80 95 85 84 70 147 Mismatches . 9 5 3 21 19 23 Not Finished . 20 31 18 8 4 31 % Mismatches . 11 5 4 25 27 16 % Not Finished 25 33 21 10 6 21 I thought that because we charged no entry fee for the tournament, we would have a record year for abandoned games. As it turned out, however, we only equaled the median from earlier years. Given that we got a record turnout without horribly diluting the quality of the tournament, I'd say the strategy was a big success. Maybe once again in 2012 we can try the signup method of waiting until a week before the tournament starts and then declaring a free-for-all. :) |
||||
|
Arimaa Forum » Powered by YaBB 1 Gold - SP 1.3.1! YaBB © 2000-2003. All Rights Reserved. |