Video Chess, the First Chess Video Game?

[BillyHW]

I guess to me, as I read the topic, the first chess video game implies a game for a video game system (which an MIT mainframe is not,) so unless the o2 chess game with the giant cpu brick (which I have,) or the fairchild chess cart (have also,) came out before video chess, video chess gets the nod as first chess video game in my book.

 

Well, it would be nice to know which were the first implementations for mainframe, home computer, and home console separately. Also, it has to be playable against a CPU opponent. If it's only two-player it doesn't count.

 

Someone should work on an emulator for IBM's Deep Blue. Or IBM should set up a website where people can play online against Deep Blue in chess, or Watson in Jeopardy. That would be cool.

[NE146]

I've brought this up before years ago (in the threads where Video Chess was pitted against other programs.. like Mac Chess I think) but it should be possible to jack up the framerate/CPU speed in modern 2600 emulators to make the response time of Video Chess much more tolerable even on the higher levels.

 

Actually.. just taking a look at Stella now, it looks like you can jack up the framerate to "300".. that should work fine eh?

 

Does anyone know at what level Video Chess starts to cheat? Is it only "Level 8"? If so, it should be a trivial matter to at least pit Video Chess on Level 6/7 against something like Chessmaster at various levels and check out the results. Would be fun.

[+thegoldenband]

Unfortunately for an Atari 2600 vs. Intellivision match, there are also bugs in the Intellivision game (USCF Chess) on the highest level. On multiple occasions I've had it get into an endless loop of error messages where it simply refuses to move at all.

[Tin_Lunchbox]

Does anyone know at what level Video Chess starts to cheat? Is it only "Level 8"? If so, it should be a trivial matter to at least pit Video Chess on Level 6/7 against something like Chessmaster at various levels and check out the results. Would be fun.

 

Video Chess increases the difficulty gradually from levels 1 through 7, however 8 is a beginner level that makes some amount of random-seeming moves and is much easier than level 1. I personally have never seen Video Chess cheat or move my pieces accidentally.

[+thegoldenband]

I've been playing through the various difficulty levels on Video Chess, helped by Stella's ability to increase the framerate so it doesn't take forever (something I didn't know about until this week -- I assumed Stella was too faithful/accurate to permit it!).

 

I've won on Levels 1-3 with the White pieces. Here's the Level 3 game, in which my play was spotty in the opening -- Black could've gotten an advantage with more accurate play -- but I eventually secured the win with a nice queen sacrifice. An animated GIF is attached (click on it to start the animation).

 

[Date "2012.06.13"]

[White "thegoldenband"]

[black "Atari Video Chess, Level 3"]

[Result "1-0"]

 

1. e4 e5 2. Nc3 Nc6 3. Bc4 Nf6 4. d3 d6 5. f4 Nd4 6. Nf3 Bg4 7. h3 Bxf3 8. gxf3 Be7 9. Be3 Kf8 10. Nb5 c5 11. c3 Ne6 12. f5 Nf4 13. Bxf4 exf4 14. Qd2 g5 15. fxg6 d5 16. Bxd5 Qd7 17. a4 hxg6 18. Qxf4 Nxd5 19. exd5 Rxh3 20. Rxh3 Qxh3 21. O-O-O Rd8 22. Nc7 Bd6 23. Ne6+ Kg8 24. Qxf7+ Kxf7 25. Ng5+ Ke7 26. Nxh3 Kd7 27. Rg1 Rh8 28. Ng5 Rh2 29. Ne4 Rh6 30. Rg2 Kc7 31. Kc2 Kd7 32. Kb3 Kc7 33. Kc4 b6 34. Kb5 Rh2 35. Rxh2 Bxh2 36. a5 Bd6 37. axb6 axb6 38. Nxd6 Kxd6 39. Kxb6 Kxd5 40. f4 c4 41. dxc4 Kxc4 42. b4 Kc3 43. b5 Kd3 44. Kc5 Ke4 45. b6 Kxf4 46. b7 g5 47. b8=Q+ Ke4 48. Qg3 Kf5 49. Kd5 g4 50. Kd6 Kg5 51. Ke5 Kh5 52. Kf5 Kh6 53. Qxg4 Kh7 54. Kf6 Kh6 55. Qg6# 1-0

 

[Pioneer4x4]

COOL, I just watched the animated gif play through the game! How did you create that? manually write the moves down as you went?

I peeked at - http://www.caissa.com/ , looks interesting.

[+thegoldenband]

COOL, I just watched the animated gif play through the game! How did you create that? manually write the moves down as you went?

 

Yep, I wrote them down as "we" played, and then turned the file into a PGN, which is basically a text file with a little bit of markup. Caissa.com has a PGN importer that makes those animated GIFs -- it's pretty slick. Glad you enjoyed it!

[Pioneer4x4]

I was surprised by how many trades it did not initiate, most lower level one to me seem to like to trade, hoping you make a mistake later.

[NE146]

Video Chess increases the difficulty gradually from levels 1 through 7, however 8 is a beginner level that makes some amount of random-seeming moves and is much easier than level 1. I personally have never seen Video Chess cheat or move my pieces accidentally.

 

Ah yeah that's what I meant.. i.e. the highest level of difficulty. I actually had my parents buy me video chess when I was a kid in the late 70's/early 80's (my Fidelity Chess Challenger 7 died) so I was very familiar with the game.. it's just been several years.

 

But yeah apparently at the highest difficulty level (i.e. 7) it starts to move pieces illegally and or pieces disappear. I've never seen it myself, but then again I've never actually seriously tried to play level 7.

[+thegoldenband]

Played and beat Level 4. This time the computer blundered in the opening, and it was just a question of gradually converting the extra pawn(s). I'm actually rather pleased with the way I slowly improved my position from moves 24-28, rather than playing d5-d6 immediately.

 

[Date "2012.06.14"]

[White "thegoldenband"]

[black "Atari Video Chess, Level 4"]

[Result "1-0"]

 

1. e4 c5 2. Nf3 Nc6 3. b3 e5 4. Bb2 Nf6 5. Nxe5 Nxe5 6. Bxe5 Nxe4 7. Qe2 d5 8. f3 Qe7 9. fxe4 Qxe5 10. exd5 Bd6 11. Nc3 O-O 12. Qxe5 Bxe5 13. O-O-O Re8 14. Bb5 Re7 15. h3 Bf5 16. Rhf1 Bd7 17. Bxd7 Bxc3 18. dxc3 Rxd7 19. Rf5 Rad8 20. c4 Kf8 21. Rdf1 b5 22. cxb5 Ke8 23. c4 Re7 24. R1f2 Rdd7 25. Kc2 Kf8 26. a4 Ke8 27. a5 Kd8 28. d6 Rxd6 29. Rxf7 Rde6 30. Rxe7 Rxe7 31. Rd2+ Ke8 32. g4 Re6 33. Rd3 Kf7 34. Kd2 Re5 35. Rd5 Ke6 36. Rxe5+ Kxe5 37. Ke3 g5 38. a6 h5 39. b6 hxg4 40. hxg4 axb6 41. a7 Kd6 42. a8=Q Kc7 43. Ke4 Kd6 44. Qd5+ Kc7 45. Kf5 Kb8 46. Qf7 Kc8 47. Ke6 Kb8 48. Kd6 b5 49. Kc6 bxc4 50. Qb7# 1-0

 

[NE146]

Hah that's great.. nice game too, pushing your pawn up like that. Looking forward to your games with Level 5-7!!

[+thegoldenband]

Hey, thanks! I don't know how soon I'll get to Levels 6-7 since I have a busy time coming up, but here's Level 5.

 

I suspect I missed some tactical opportunities in the middlegame, and made the position a lot messier than it needed to be. But paradoxically I think that hastened the end of the game, since Video Chess doesn't seem to handle tactically complex positions nearly as well as simpler positions. (Maybe I'm playing into its hands by playing normal chess instead of "anti-computer" chess, but that's me.)

 

[Date "2012.06.14"]

[White "thegoldenband"]

[black "Atari Video Chess, Level 5"]

[Result "1-0"]

 

1. e4 c5 2. Nf3 Nc6 3. b3 e5 4. Bb2 d5 5. exd5 Qxd5 6. Bc4 Qe4+ 7. Qe2 Bf5 8. Bxf7+ Kd8 9. d3 Qb4+ 10. Nbd2 Nd4 11. Bxd4 Nh6 12. c3 Qa3 13. Nc4 Qa6 14. Ncxe5 cxd4 15. Nxd4 Bxd3 16. Qf3 Rac8 17. Qd5+ Bd6 18. c4 Be4 19. Qxe4 Qa5+ 20. b4 Qxb4+ 21. Kf1 Bxe5 22. Qxe5 Nxf7 23. Qd5+ Nd6 24. Ne6+ Ke7 25. c5 Nc4 26. Rd1 Qb5 27. Kg1 Rhe8 28. Nxg7 Rxc5 29. Qe6+ Kf8 30. Qf6+ Kg8 31. Nxe8 Rg5 32. h4 Rg6 33. Qd4 a5 34. h5 Rg5 35. Nf6+ Kh8 36. h6 Qb6 37. Qd8+ Qxd8 38. Rxd8+ Rg8 39. Rxg8# 1-0

 

[roberto]

Unfortunately for an Atari 2600 vs. Intellivision match, there are also bugs in the Intellivision game (USCF Chess) on the highest level. On multiple occasions I've had it get into an endless loop of error messages where it simply refuses to move at all.

 

Idea: let's have them play at the highest level anyway. Then, the first one who cheats gets reported to the referee and loses the game

[Pioneer4x4]

I was wondering how the game allocated RAM for display and move calculations. Could stella be used to monitor what moves it is considering while thinking? I ass-u-me it needs to keep a copy of current positions and a table of sorts to test moves and figure out outcomes, all in 128 bytes, I am more and more amazed by that, not as much as the limit on code size. I think you need at least 56 full bytes to store the board,

1 per square, and 7 possibilities, 6 pieces and color. or would you keep a table of the 32 pieces, and their locations, that would take 64 bytes as well, correct?

[Rybags]

Not necessarily. You could hold a piece position in 6 bits - X/Y ranges only 0 to 7. 2 bits spare, 1 bit could represent which player's piece it is. 1 bit flags whether the piece is still in play.

 

It's the sort of game that would be interesting to see the source for. They must have used every trick in the book to get it working in so little Ram.

Edited June 15, 2012 by Rybags

[Pioneer4x4]

Not necessarily. You could hold a piece position in 6 bits - X/Y ranges only 0 to 7

Isn't that 8?0,1,2,3,4,5,6,7

[Tin_Lunchbox]

Hey, thanks! I don't know how soon I'll get to Levels 6-7 since I have a busy time coming up, but here's Level 5.

 

Go! Go! Go! Go!

[+thegoldenband]

Not necessarily. You could hold a piece position in 6 bits - X/Y ranges only 0 to 7

Isn't that 8?0,1,2,3,4,5,6,7

Well, you just need 3 bits to count to eight, so 3 bits x 2 = 6 bits.

 

BTW I wonder -- if you had a counter to keep track of how many pieces are in play, then as the number of pieces goes down, could you use the freed-up RAM for calculations? In other words, at the start of the game you'd need to use 32 bytes to track piece locations, but if you used 5 bits of another byte to track how many pieces are on the board -- plus, say, 2 more bits for castling rights for each side, and 1 more for something else -- then for every piece that left the board, you could potentially use its RAM location for move trees.

 

But maybe that'd be overcomplicated since you'd need to be able to shuffle the RAM locations of the pieces, and perhaps that'd take more overhead than the RAM you'd recoup.

 

Hey, thanks! I don't know how soon I'll get to Levels 6-7 since I have a busy time coming up, but here's Level 5.

 

Go! Go! Go! Go!

 

[Pioneer4x4]

Well, you just need 3 bits to count to eight, so 3 bits x 2 = 6 bits.

 

Dang, I keep blowing it on stuff like that, you are exactly correct.

[+thegoldenband]

OK, here's Level 6! Yet again, I managed to reach victory in fewer moves than the previous game, despite playing against a higher level. However, it certainly wasn't a "quick" win in terms of time: the game itself took almost five hours, and that's with Stella running at 300fps (which I assume is roughly equal to 5x normal speed). The computer's first 22 moves went by at a reasonable pace, but the 23rd and 24th took a while, and the computer took at least 3.5 hours of real time to make its 25th move. If I were playing on an actual VCS, that one move would've taken over 16 hours!!

 

Anyway, this is a more typical anti-computer victory, with the computer underestimating my kingside attack until it's much too late.

 

[Date "2012.06.17"]

[White "thegoldenband"]

[black "Atari Video Chess, Level 6"]

[Result "1-0"]

 

1. e4 e5 2. Nc3 Nc6 3. Bc4 Nf6 4. d3 d6 5. h3 Be6 6. Bb3 Be7 7. f4 O-O 8. Nf3 Nd4 9. Nxd4 exd4 10. Ne2 c5 11. O-O Re8 12. f5 Bd7 13. Nf4 b5 14. Nd5 a5 15. a4 Nxd5 16. Bxd5 Ra7 17. axb5 Bxb5 18. Bd2 Bd7 19. Qh5 Rf8 20. g4 Bf6 21. g5 Be5 22. g6 hxg6 23. fxg6 {long think here, and on the next move} Bh2+ 24. Kxh2 Qh4 25. Qxh4 {computer thought for at least 3.5 hours, real time} Re8 26. Qh7+ Kf8 {the only move, but the computer thought for 8 minutes before playing it...} 27. Rxf7# 1-0 {...and it thought for 12 minutes before acknowledging the checkmate!}

 

[OldSchoolRetroGamer]

Sometimes I regret never learning CHESS !

[+stephena]

However, it certainly wasn't a "quick" win in terms of time: the game itself took almost five hours, and that's with Stella running at 300fps (which I assume is roughly equal to 5x normal speed).

 

This is exactly five times the rate (60fps x 5 = 300fps). About the framerate, there's actually an undocumented 'feature'. If you launch from the commandline and use the 'framerate' option, you can use numbers much higher than 300. The 300 was chosen arbitrarily, and is enforced only within the UI. So it can potentially run much faster than that. Note however that at high speeds the sound goes out of sync quite quickly.

 

EDIT: I just tried on my development system with the following: OpenGL mode, vsync turned off, TV effects turned off, zoom 5x, framerate of 10000. At these speeds, the reported fps won't be correct, but informally I counted 1 second and then immediately quit. In that time, it had drawn almost 2000 frames, and took up an entire core on my CPU! (normally the same settings result in less than 1% CPU usage). So it's definitely possible to run much faster.

[Tin_Lunchbox]

Sometimes I regret never learning CHESS !

 

What are you waiting for, OldSchool? I'll start you off: Bishops move only diagonally, rooks only laterally and vertically, the queen in any direction; all of the preceding for as many spaces as you like until they run into something which they then capture and remove from the board.

[+thegoldenband]

Stella running at 300fps (which I assume is roughly equal to 5x normal speed).

 

This is exactly five times the rate (60fps x 5 = 300fps).

 

Just to be clear, my "roughly" was meant as a hedge against the possibility that my aging PowerBook G4/1.67 GHz might not be able to muster the full 300fps (especially when running in the background and with 1-2 brief sleeps thrown in there), or that I was wrong about the equation you cited (I'm never sure when 29.97 factors in). I know as long as my CPU is up to it, Stella will hit it right on the nose.

 

About the framerate, there's actually an undocumented 'feature'. If you launch from the commandline and use the 'framerate' option, you can use numbers much higher than 300.

 

Hey, that's fantastic news! I've never launched Stella from the command line before, but I'll check that out for sure before I play Game 7. If yours tops out at 2000, I wonder where my PowerBook will?

 

The only downside would be if the UI of the game itself becomes unusable at high speeds (which happens somewhat in Intellivision's USCF Chess), but it behaves OK at 300fps so I guess adjusting it after boot shouldn't be necessary, and if it is I can always use a savestate and reload the app.

 

Sometimes I regret never learning CHESS !

 

What are you waiting for, OldSchool?

 

I was going to say -- never too old, even if your name has "Old" in it!

[NE146]

Easy to learn how to play Chess.. getting good at it takes a little longer.

 

That said.. @thegoldenband: Nice game 5 & game 6 man.. your skills are obviously far and above mine. Definitely the right guy to check it out. Looking forward to your eventual game 7 and to see if the darn thing really does glitch out.

 

It's still amazing to me though how someone programmed the VCS, a pong machine, to actually play a passable game of Chess.. in the 1970's at that. I'd say it's an engineering miracle.