TASVideos

Posted: 12/5/2017 8:23 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

It seemed like it was much easier to make Chessmaster go quickly off-book by making it play white, and then making a move with black that isn't in its opening book. Much harder to do that with Chessmaster playing black, as the white moves need to be somewhat reasonable. I tried it anyway. In this game move 3. Nd2 threw it off its book (incidentally, this move was the primary move suggested by Stockfish, so it was good.) The game ended being one ply shorter than the previous one I posted. And the mate is even more beautiful. At move 17 Chessmaster was greedy and went for the hanging pawn with Qxb2, which was a grave mistake: Stockfish found a mate in 11. Of course Chessmaster didn't see that, and went for the greedy continuation of taking the hanging rook, with Qxa1, which only sped up the mate by a move. Essentially, this mate sacrifices both rooks (seemingly in exchange for a single knight), which is just absolutely beautiful. 1. e4 e6 2. d4 d5 3. Nd2 Ne7 4. Ngf3 Ng6 5. h4 Nd7 6. h5 Ne7 7. h6 gxh6 8. Bd3 Bg7 9. exd5 exd5 10. Nf1 c5 11. Ng3 Qa5+ 12. Bd2 Qb6 13. Nh5 Bxd4 14. Nxd4 cxd4 15. Ng7+ Kf8 16. Bxh6 Kg8 17. Qe2 Qxb2 18. Qxe7 Qxa1+ 19. Ke2 Qxh1 20. Nf5 Qxh6 21. Nxh6+ Kg7 22. Qg5+ Kf8 23. Nf5 Nf6 24. Qg7+ Ke8 25. Bb5+ Kd8 26. Qxf6+ Kc7 27. Qd6# 1-0

Posted: 12/3/2017 11:49 AM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

I figured out how to do the above, so I suppose my question isn't relevant anymore. This is the shortest game I have managed so far (once again at maximum strength, opening book enabled), Chessmaster playing white. 27 moves. Still took 1 hour and 50 minutes. The mate in the end is just beautiful. Stockfish is a monster. 1. f4 e6 2. Nc3 b6 3. e4 Bb7 4. d4 Bb4 5. Qg4 Nf6 6. Qxg7 Rg8 7. Qh6 Bxe4 8. Nf3 Nc6 9. Qh3 Bxc2 10. a3 Bf5 11. Qh4 Be7 12. Qf2 Na5 13. Nd2 Ng4 14. Qg1 c5 15. d5 exd5 16. Nxd5 c4 17. Bxc4 Bc5 18. Qf1 Rg6 19. b4 Bd4 20. Ra2 Re6+ 21. Be2 Qc8 22. Bb2 Bxb2 23. Nc4 Qxc4 24. Nf6+ Bxf6 25. Kd1 Qb3+ 26. Kd2 Bc3+ 27. Kc1 Qb1#

Posted: 12/3/2017 8:22 AM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

Warp wrote:

And also, I wonder if it would be possible to make the emulator pause automatically when the game makes a move.

It appears that when a move has been made, and the game plays a sound, the value at memory address 9E changes from 64 to 32. Then, when the sound has done playing, it changes back to 64. This might be a good trigger for pausing the emulation. Could someone give me some hints on how to write a piece of lua to do this?

Posted: 12/2/2017 7:09 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

goldenband wrote:

Barring that, the next-best option is likely to be (1) either a closed position in which one player builds an attack on one side -- something computers are notoriously bad at evaluating -- until checkmate is inevitable; or (2) some wild sacrificial line where, again, the CPU wolfs down material until it's too late.

I think that the suggestion of finding an opening trap (that's not in Chessmaster's opening book, or if not using the book) could be the most promising approach. It's just that I don't have enough chess opening knowledge to remember such traps by heart, and there are quite tons of them. Also, from some quick testing it seems that it's really hard to induce Chessmaster to select a given opening line. Its randomization does not seem to work by using the timing of user input. (For example, if I let it start with white, it will for some reason start with c4 like 95% of the time. Sometimes it will do e4, or even f4, but I can't figure out how that decision is made. Timing of keypresses seems to have very little to no effect.)

Posted: 12/2/2017 7:03 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

Btw, is there a way to make BizHawk emulate faster than at 400%? It would make testing faster and easier. (It's somewhat ironic that one would want the emulator to go significantly faster when creating a TAS, but this is a rather special situation... :P ) And also, I wonder if it would be possible to make the emulator pause automatically when the game makes a move. (Not that this is absolutely necessary, as it probably doesn't matter how long the player "thinks", when simply running tests, but if there's an easy way, it would be nice.)

Posted: 12/2/2017 9:20 AM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

£e Nécroyeur wrote:

Warp, how does Chessmaster's thinking time after the player makes a non-book move compare to the thinking time after making a book move?

Chessmaster makes a book move almost immediately, within a second. At the maximum level (ie. level 13) it takes about 4-5 minutes to move on average.

Posted: 12/1/2017 9:17 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

This is not easy. It's hard to find suitable opening traps that would work with the program. It's also hard to induce it to choose the desired book moves. But one thing, at least, that can be done to shorten a game, when book moves are enabled, is to immediately make a move that's not in the program's opening book. That will throw it off, and make it lose in fewer moves. On the other hand, this only saves a very small amount of time, because book moves are done in seconds anyways, so it only shortens the game in terms of number of moves, not really in terms of consumed time. Anyway, here's a game where Chessmaster is playing white. I chose as the first black move one that was not in its book, so it immediately went into thinking mode. The game is "only" 36 moves long, but still took over 2.5 hours of thinking time by white.

Posted: 11/28/2017 3:19 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

How exactly is that rule enforced? I doubt it can be automatized (at least on in general).

Posted: 11/28/2017 7:14 AM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

Warp wrote:

On what basis it is determined that two consoles are the "same" console? On what basis it is determined that two games are the "same" game?

A recent question in this forum about language versions added an additional layer which I hadn't thought before: If, for instance, the Japanese version of a game is faster than the English version because there is less text to print in the former, are they still the "same" game? Would the Japanese version always obsolete the English version for the sole reason that the lesser amount of text makes it faster? In other words, the English version has no chance of ever being published (at least if somebody makes a TAS of the Japanese version).

Posted: 11/25/2017 8:35 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

I saw this on a t-shirt:

Is that really so?

Posted: 11/24/2017 2:00 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

The Fibonacci numbers are: 1, 1, 2, 3, 5, 8, 13, 21, ... The Lucas numbers are: 2, 1, 3, 4, 7, 11, 18, 29, ... If you take the ratio between a Lucas number and the correspondent Fibonacci number, this ratio quickly converges to a specific value, the farther in the sequences you take this pair of numbers. This value seems to be approximately 1.38196601125. What value is this? (Obviously I'm not looking for an approximation but, if possible, the exact value.)

Posted: 11/24/2017 12:40 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

I will try to see if I can induce chessmaster to fall into one of the numerous traps that exist.

Posted: 11/23/2017 9:56 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

What is the result of the infinite sum: 3/4 + 3/16 + 3/64 + 3/256 + 3/1024 + ... Does the following infinite sum converge or diverge? 1/10 + 1/20 + 1/30 + 1/40 + 1/50 + 1/60 + ...

Posted: 11/23/2017 6:30 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

I made an animated gif of the second game using an online tool, for easier viewing.

Posted: 11/23/2017 6:19 PM

Post subject: The Chessmaster

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

I know that a TAS of a pure chess game being accepted has slimmer chances than a snowball in hell, but anyways, since I am a bit of a computer chess fan, I got curious about how efficiently a modern top chess engine could beat a super-old chess game made for a console like the SNES. I already commented about this in the general forum, but here it is anyway, in its more proper forum. I chose as the ~~victim~~... eh, game The Chessmaster, which appears to be a seriously-made pure chess game for the SNES. And to assist me I chose the currently strongest open source chess engine in existence, Stockfish (ok, there are arguably clones that are even a bit stronger, but I chose the official version nevertheless). Because I want the console game to play at its strongest, not at its weakest, I chose level 13, which is the highest (other than "infinite"). In this first game I allowed The Chessmaster to use its opening book, which got used for the first about 14 moves. Perhaps because of this the game was relatively long, 46 moves, and lasted for over two hours. 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 d6 8. c3 O-O 9. h3 Na5 10. Bc2 c5 11. d4 Qc7 12. Nbd2 cxd4 13. cxd4 Bb7 14. d5 Rac8 15. Bd3 Nd7 16. Nf1 f5 17. b3 fxe4 18. Bxe4 Nc5 19. Bc2 Nd7 20. Ne3 Bf6 21. Bf5 e4 22. Nd4 Qc3 23. Rb1 Qxd4 24. Bxd7 Qxd1 25. Rxd1 Ra8 26. Be6+ Kh8 27. Bd2 Bd8 28. Bb4 Bb6 29. Bxd6 Rfd8 30. Be5 Bc8 31. Nf5 Ra7 32. Rbc1 Bxe6 33. dxe6 Rxd1+ 34. Rxd1 Kg8 35. Rd6 Rb7 36. e7 Kf7 37. Re6 Rxe7 38. Rxe7+ Kg8 39. Re8+ Kf7 40. Nd6+ Kg6 41. Re6+ Kh5 42. g4+ Kg5 43. Nxe4+ Kh4 44. Kg2 Bxf2 45. Rh6+ gxh6 46. Bf6# 1-0 I tried a second game, this time disabling the opening book, to make for a faster game. I also raised the contempt setting of Stockfish to full. This game was over in just 27 moves, and was, according to the game's own clock, 1 hour and 39 minutes long (only counting black's thinking time, which would be pretty much the only thing that would matter in a TAS). 1. e4 d5 2. exd5 Qxd5 3. Nc3 Qa5 4. d4 Nd7 5. Nf3 e5 6. Bd3 Bb4 7. O-O Bxc3 8. bxc3 Qxc3 9. Bd2 Qa3 10. Re1 Ne7 11. Rb1 Qxa2 12. Nxe5 Nxe5 13. Rxe5 Kd8 14. Qe1 Nc6 15. Ra1 Qb2 16. Rd5+ Bd7 17. Bc3 Qb6 18. Rb5 Qxb5 19. Bxb5 Nxd4 20. Bxd7 Nf3+ 21. gxf3 Kxd7 22. Rd1+ Kc8 23. Qe7 b6 24. Rd7 Kb8 25. Rxc7 a5 26. Rb7+ Kc8 27. Qc7# 1-0 It might be interesting to make a TAS out of this, but I suppose it would be rather futile. It would be rather long (even with luck the time would still probably be well over an hour), and doesn't really fit into the acceptable game genres. But one can always dream.

Posted: 11/21/2017 10:02 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

Sorry for going a bit on a tangent, off topic even, but if you would forgive me. Due to what I wrote above about running Stockfish against an SNES chess game, for the heck of it I tried what would happen if I played against The Chessmaster for the SNES, at its strongest, using Stockfish 8. (Obviously I sped up the emulator, or else the game would have taken something like 3 hours). This is the resulting game, for anybody who cares (Stockfish playing white, The Chessmaster playing black): 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 d6 8. c3 O-O 9. h3 Na5 10. Bc2 c5 11. d4 Qc7 12. Nbd2 cxd4 13. cxd4 Bb7 14. d5 Rac8 15. Bd3 Nd7 16. Nf1 f5 17. b3 fxe4 18. Bxe4 Nc5 19. Bc2 Nd7 20. Ne3 Bf6 21. Bf5 e4 22. Nd4 Qc3 23. Rb1 Qxd4 24. Bxd7 Qxd1 25. Rxd1 Ra8 26. Be6+ Kh8 27. Bd2 Bd8 28. Bb4 Bb6 29. Bxd6 Rfd8 30. Be5 Bc8 31. Nf5 Ra7 32. Rbc1 Bxe6 33. dxe6 Rxd1+ 34. Rxd1 Kg8 35. Rd6 Rb7 36. e7 Kf7 37. Re6 Rxe7 38. Rxe7+ Kg8 39. Re8+ Kf7 40. Nd6+ Kg6 41. Re6+ Kh5 42. g4+ Kg5 43. Nxe4+ Kh4 44. Kg2 Bxf2 45. Rh6+ gxh6 46. Bf6# 1-0 The 14 or so first moves were book moves by Chessmaster, so they are pretty standard. It starts falling apart quite quickly after that. When doing this, I was tempted by the idea of actually making a TAS of it and submitting. But I suppose it would be just a waste of everybody's time because it would probably just be rejected outright (not least because it would be something like 2 or 3 hours long).

Posted: 11/21/2017 7:55 AM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

Radiant wrote:

It strikes me that the rule against board games was written because chess TAS'es mainly focus on breaking a poorly written AI, and this rule hasn't considered the plethora of other board games that aren't chess or checkers or go.

I wouldn't mind a chess game TAS, and I don't see why that genre should be ruled out by default. If someone wants to submit a TAS of a chess game, let them, and we'll see how it fares in the poll. But I will grant that unless the game has something special about it (like some kind of "story mode"), it can be difficult to justify. You could take some, let's say, SNES chess game, play it at its hardest difficulty, and use a modern top engine (like Stockfish) to beat it as optimally as possible, but then it just becomes a display of how a modern chess engine using modern hardware can crush an SNES era engine. Of course even that in itself would be interesting to me, at least. And it would be, in a sense, a tool-assisted "superplay" (the modern chess engine being part of the toolset), so in that sense it would fit quite well. (If the counter-argument would be that it wouldn't be very interesting to most people... well, I'm sure that from the thousands of TASes that have been published, you can find some that most people aren't interested in. I don't think that should automatically be grounds for rejection.)

Posted: 11/21/2017 7:30 AM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

I think we might need yet another rule change...

Posted: 11/20/2017 7:04 AM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

ThunderAxe31 wrote:

Warp wrote:
I suppose Mahjong could perhaps be considered a "sport" game, which has certain requirements for being acceptable for publication: http://tasvideos.org/Vault.html
It looks more like a board game to me.

Do board games have their own section in the rules?

Posted: 11/17/2017 2:01 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

I have been unusually slow at watching new episodes, for some reason. Laziness perhaps. Anyway, just watched the next episode on my to-do list, ie. Triple Threat. It wasn't as utterly nonsensical as the previous episode (as I commented before), but still pretty nonsensical nevertheless, especially the beginning. Spike acting like a complete idiot might be expected (if anything can be expected from Spike, given that his character is anything but established, of all the characters in the franchise), but why did the ponies go along with it? It made no sense. At least the last quarter of the episode somewhat saved it a bit. Not completely, but at least it made it slightly better. I hope these nonsensical plots don't continue.

Posted: 11/17/2017 9:43 AM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

I was curious to see if the new Firefox Quantum would run these tests faster, but for some reason it doesn't. Strange. I suppose it's only faster for client-side dynamic content that actually modifies the webpage, rather than raw calculation.

Posted: 11/16/2017 1:29 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

I suppose Mahjong could perhaps be considered a "sport" game, which has certain requirements for being acceptable for publication: http://tasvideos.org/Vault.html

Posted: 11/12/2017 1:02 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

In his newest video, blackpenredpen tackles the problem: lim(n→∞) (n!/nⁿ)^1/n I found it very interesting. Before watching the video, perhaps you could be interested in trying it yourself. (Btw, he didn't need the gamma function to solve it, although I suppose there's no rule that would forbid using it.)

Posted: 11/10/2017 5:42 PM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

When I read your post, it was a bit unclear to me what exactly it's saying. It's a bit confusing. I know that it's not very helpful for me to just say this without going into a deep analysis (which I might do in a later post), but I'm wondering if other people are likewise not really getting the point and that's why they aren't rushing to reply. (Oh, and this is my 7000th post. Celebrations.)

Posted: 11/9/2017 7:44 AM

Quote

Warp

Banned User, Former player

Joined: 3/10/2004
Posts: 7698
Location: Finland

I wish somebody would integrate the math plugin into the forum... http://tasvideos.org/forum/viewtopic.php?t=19519

Posts for Warp