OTHER RECORDED COMPUTER CAMELOT GAMES
Anyone who wishes to submit a Camelot game score of any game played against a computer program (Chaxx or Camelot-Zillions or any other program) for listing consideration should forward it, in algebraic format, preferably dated, to Michael Nolan at firstname.lastname@example.org. The Website Manager reserves the right to make all inclusion decisions on the very subjective basis of whether or not he finds the game to be of some interest, however, given the severe shortage of published computer Camelot games, most games that get submitted will probably be listed. Remember, by the way, game scores are not copyrightable—only annotations may not be used without permission. I hope you enjoy these games!
Michael Nolan vs. Chaxx 1.05 2008
Dan Troyka vs. Chaxx 1.13 2008
(225 secs. per move)
(Annotations by Michael Nolan)
The Black Men on F11 and C2 are needed for defense. Their failure to retreat, regroup, and join up with their comrades loses the game, since their attack on White's Castle will be a race that cannot be won—White will force his way into Black's Castle first, as demonstrated as follows by World Champion Troyka.
50.G12-F13 Resigns 1-0
White: CHAXX 1.1x (time limit as stated)
Black: Michael Nolan (no time limit)
May-June 2009 on BrainKing.com
(Commentary by Michael Nolan [MrWCF] and Rick Fadden [Chaxx])
Chaxx: When I first wrote Chaxx I was hoping for the day when I could see how the program plays against a great Camelot player. As I may have mentioned we also needed this "online" capability (BrainKing), and I am really thankful that we now have this ability to play over the Internet. With my recent minor mod to the Transposition Table I could call the latest exe version 1.15 (with SSE). I ran this on my 2.0 GHz Core 2 Xeon for 3600 seconds last night, and this move "F6-F8" was the result. I'm looking and wondering if this is a blunder. I'll assume that this is an OK move and we'll see how this goes.
MrWCF: My opening book, based upon 24-hour analysis, says that 1.F6-F8 is the strongest opening move with a plus score of 20. The next strongest first move for White, according to that analysis, is 1.E6-G8, with a plus score of 11. 1.E6-G8 is my personal favorite--I use it in every game that I have White because I think it is more flexible than 1.F6-F8. Out of the 100 first moves for White, only 12 yield a positive score. My opening book says that 1....G10-E12-C10 is Black's best response to 1.F6-F8, but I have my doubts. It shows that nasty inclination to move to the outside. That's why I've suggested that strength should be added to the "keep-to-the-middle" evaluation factor. In any case, we are now out of my (very, very small) book, so I'm on my own against the monster!
MrWCF: Attacking the weak G8 square.
Chaxx: 8 ply with updated Quies Search
MrWCF: Attacking the D8 square with the D10 Knight, and opening squares for the J11 Knight. Here not 3....I10-G10-I8xG6xE8 because of 4.G8xI10xI12xK10 E8xG8xG6, 5.D7-F5xH7 leaving Black down a double exchange (losing two Knights for two Men).
Chaxx: Using the original algorithm, 8 ply...
MrWCF: As I mentioned in one of our many emails, I have recently started thinking that initially, a Knight is worth slightly more than two Men (and thus my adjustment in the Knight/Man formula values), and therefore, not such a great move choice for CHAXX in this game . Of course, if we trade down all the way, two Men will be worth a great deal more than a Knight. The question is whether I can stay alive long enough to get value for the extra Knight.
Chaxx: Original QSearch, 8 ply
MrWCF: An easy choice because I've got to guard against the new D7-F7-H9xH11 etc. threat, but I can't expose myself to J6-H8xH10 by moving, say, H10-H9. 5....I10-H11 has the added benefit of threatening the H9 square. I would like to consolidate by bringing the F13 Knight back into the battle, but that will have to wait.
MrWCF: This is the type of position where CHAXX, as Black, would play 7....D9-F11-F9 or 7....D11-F9 with complex calculations. It's much easier for the human (me) to play 7....I11-H10, though--the other lines would take lots of time for me to figure out--much more than three or four minutes.
Chaxx: I questioned this move at first because I thought Chaxx might be messing up, so I used another copy of Chaxx to explore a but further and this verified it was not encountering a horizon effect. So this is an 8 ply search, new QSearch logic, plus verified to apparently not be a stupid machine move...
MrWCF: 8.H9-I9 is pretty clever by CHAXX--it has several ways to recapture, e.g., I7-I5-K7xI9 or D7-F7-H7-J7xJ9 or I7xK9. It must like one of them.
Chaxx: Yes, upon the other possible capture for Black: H11-H9xJ9 then it would play F8-G9, then upon F10xH8 it plays G7xI9xK9.
MrWCF: Getting the Knight back into play.
MrWCF: Well, CHAXX wouldn't make that move unless it can regain at least as much as that Knight, so I expect something from one of the pieces that can make contact: H6, I7, I6, G7, F6, E7, E6, or F8. That's why Camelot is so difficult--you have to look at all moves that result in a capture being forced (forcing moves) as well as all Jumps and Knight's Charges, and that only gets you to a depth of one move. I immediately rejected 10....D10-F12-H12-H10xJ10 and 10....C11-E9-G11-I11xI9 because they both would have left a further weakness on my dark squares (D8, E9, C9, B10, B12, D12, and D10 or C11). (That seems like a funny statement for Camelot because a weakness on light or dark squares is very much a Chess concept.) Anyway, CHAXX has a huge weakness on the light squares (F9, I8, E8, G8, F7, D7, H7, J7, G6, D5, F5, H5, and J5), but I haven't seen a way to exploit it, at least yet. Light or dark square weaknesses is the reason why it is important in Camelot to keep your pieces bunched together in one mass.
Chaxx: It seems to be finding ways to trade down in order to increase it's score based upon Knight value when there are less pieces on the board... default setting, your values, sort of funny that it's your idea in there... But of course when Chaxx crushes me tactically, it's my logic hitting me over the head...
MrWCF: Pushing forward on the right flank. It will be interesting to see if CHAXX splits its forces, moving forward on its right, and defending on its left, or if it groups all of its forces on one side or the other.
Chaxx: I'm definitely learning about Chaxx by playing through this serious game. Looking at the 8 ply PV it "thinks" you want to go toward the Castle, and it intends to block your path no matter what you do... Roughly it will use the forces available, it won't reinforce or shift pieces.
MrWCF: Continuing the advance! It will be very interesting to see if CHAXX groups and retreats all of his forces to defend, or if it retreats only the F6,E6,D6,C5 Men, and attacks with the rest of its pieces--I think that that's what I'd do.
Chaxx: It definitely intends to block you...
MrWCF: Yes, but with all or part of its forces. Right now, it's just getting out of the way of my continual attacks.
Chaxx: If you were White what kind of thing roughly would you do? I would move at least one Knight over to the left using the ladder that is already there. I'm not a great player though and I don't have that much experience.
MrWCF: Well.... CHAXX doesn't give up two Men for nothing, so a quick inspection shows that after I take with 15....D10-B8-B6xD4xD6, CHAXX will play either 16.I7-G5-E7xC5 or 16.I7-I5-G7-E5xC7xA7. The first one wins a Knight for two Men and puts us back to even. The second one wins the exchange (wins a Knight and a Man for two Men, i.e., a net win of a Knight for a Man ) but I can get the exchange back with a simple Checkers-like shot (combination): 16....D11-F11-D9-B7, 17.A7xC7 C8xC6, and we will be back to a Knight vs. two Men. I am pretty sure that CHAXX will choose the second line because it is theoretically the best one. (The Knight continues to lose value relative to a Man as the pieces get traded off.) However, I want to see how CHAXX does in the endgame, so we will see how that decision works in a real game.
As for your question, if I were White, I would attack on Blacks left flank with forces equal to whatever Black forces were defending there plus two Men (so after a trade-down, two Men would be attacking Black's Castle), and defend on White's left flank with whatever forces I had left. But to be honest, it might be a better strategy to regroup all of my forces and go from there. Who knows..... Against a human, I would probably (55-45) elect to regroup. Against CHAXX, however, I would be up against a superior tactician, so I would probably (90-10) opt to split my forces and hope that CHAXX would not play correctly strategically.
Chaxx: My intuitive feeling is that you will win in the endgame. I think the machine going for this trade situation actually makes sense because it likely would have explored all of the variations where you powerful set of men came in and possibly broke through. It would see the position after these captures as better as far as distance to Castle.
Chaxx: After a 10 ply search. With less going on Chaxx searches deeper.
MrWCF: I predict that at some point soon, CHAXX will move its K9 Man, by itself, toward my Castle. That will not be strategically wise, in my opinion.
Chaxx: 10 ply search with 512 Mb T-Table, new Q-Search logic, version 1.17SSE.
MrWCF: I would have liked to have played 19....C9-C8, but that would have allowed CHAXX to exchange off two more Men--something I can't afford to let happen, at least too easily.
Chaxx: 10 ply.
MrWCF: I don't see how I can make progress toward CHAXX's Castle as well as avoid trading pieces with CHAXX, so I'll move 20....B6-A5, and at least force CHAXX to separate its pieces if it decides to force an exchange.
Chaxx: Interesting game.
MrWCF: Yes, it is. It will be very interesting to see how CHAXX chooses to play with its pieces spread all over the place. My forces should be able to take advantage of some kind of mass movement as mentioned in Paul Yearout's "Camelogistics" article. First this move, 22....C9-D8, then ....F11-D9-D7 and ....E11-E9-C7-E7 or some such to follow.
Chaxx: 12 ply search in 2118 seconds. That's a record, I've never had Chaxx search that deep previously.
MrWCF: I could also have moved 23....E11-E9-C7, but I would prefer to wait and see what CHAXX does. If CHAXX shifts the I6 and H5 Men back for defense, I can bring the Men on E10 and E11 forward. If CHAXX moves the I6 and H5 Men forward toward my Castle, I can bring the Men on E10 and E11 back to defend.
Chaxx: 12 ply search in 1456 seconds.
Chaxx: 12 ply, 970 seconds, silly machine...
MrWCF: I don't know, Rick. I'm not so sure CHAXX is being silly. In fact, I might have tried this same strategy. Soon to follow is I6-K8-K10, and it may win the race. Meanwhile, I've got to waste time in order to move my four pieces toward the center to avoid a potential defensive setup by CHAXX with Knight on C2 and Man on D3.
Chaxx: 12 ply search, 2131. seconds. I'm not an expert at the game, but intuitively it looks to me like your pieces could overpower the limited defenses and win. Also there might be some way to sweep around the two main defending pieces. I'm not sure though. This game is really instructive for me. I've already learned quite a lot while running the program and studying the game. I've come up with numerous ideas for improving the program.
MrWCF: This is very interesting to me, because I think that I'd be torn between advancing and retreating with the K9,J7,I6 Men if they were mine, so I naturally wonder what CHAXX will decide. CHAXX might even decide to retreat at 10 ply but advance at 12 ply.
Chaxx: 12 ply search, 1070 seconds.
MrWCF: This will be very difficult for CHAXX to get right, I think. CHAXX must be able to see 20 ply ahead. For example, one possible (best?) line is 28.D2-D3 D7-E7, 29.C2-E4xG4 E5-F4, 30.G4xE4 E7-E5xE3xC3, 31.I6-K8-K10 E6-F5, 32.K9-K11 F5-E4, 33.K10-K12 E4-E3, 34.K12-J12 C3-D4, 35.K11-I13 D4-F2, 36.J12-H14 E3-G1, 37.I13-G15 F2-F1 and CHAXX loses.
Chaxx: 12 ply, 4101 seconds. I need to overhaul Chaxx's static position evaluation logic to better deal with positions such as this one. I think there's no doubt Black wins from this position. Chaxx didn't see the danger, it was thrown off by it's own pieces having a nearly clear shot toward it's scoring Castle.
MrWCF: But CHAXX did see the danger after my last move, and moved 28.I6-H5. That shows, I think, that CHAXX is successfully looking very far ahead, else it would have incorrectly continued 28. D2-D3 or 28.I6-K8-K10. I think that CHAXX is playing very strongly, indeed. I think that the outcome of the game is still in doubt! And don't forget that I am not limiting myself to only three or four minutes of thought on each move.
Chaxx: 13 ply search, 9948. seconds.
MrWCF: Not 29....E5-G3, 30.H4xF2 F5-F3xF1, 31.D2-E3 F4xD2, 32.C2xE2 and CHAXX wins!
Chaxx: 13 ply, 500 seconds.
MrWCF: Well, now we're going to be back to material equality after thirty-one moves by both of us. But this position, I'm pretty sure, is an easy win for me.