|
< Earlier Kibitzing · PAGE 1 OF 3 ·
Later Kibitzing> |
| Feb-12-05 | | Albertan: Hiarcs is a great chess program. I remember when I had Hiarcs 5.32 years ago and I was watching the Linares tournament live on the internet. Even back then Hiarcs was picking more of the GM's moves beforehand than any of my other chess programs (Fritz,Chess Tiger, and Junior) The fact that Hiarcs drew a match with GM Bareev is supportive evidence to my claims about it:http://www.chessevents.nl/bar_match... http://www.hiarcs.com/hiarcs_milest...
http://www.chessbase.com/newsdetail...
http://www.chessbase.com/newsdetail... |
|
| Feb-12-05 | | csmath: Based on more than 100,000 games played between computers with the same hardware, SSDF rated Hiarcs 9 (the newest, I actually used it) only 11th: http://web.telia.com/~u85924109/ssd...
behind Shredder 7.04 and Shredder 8, Fritz 8, and Junior 8. I personally believe that Hiarcs 9 is stronger than Fritz 8, but these are irefutable statistics. Shredder 9 has just been released so that program is probably the best right now.
Hydra is not available in retail and it is heavily based on cluster computing so I doubt it is as good as to make the top here. |
|
| Feb-12-05 | | Albertan: CSMath, I have had some discussion with one of the friends of the designers of Hiarcs. He told me that the reason that Hiarcs 8 does not perform as well against other programs is due to the fact that Hiarcs is not as fast at calculating as some of the other programs you mentioned. This information was found on the internet:
" With the improved search function (found in Hiarcs 9), it's now possible to assess positions using new positional criteria such as “typical pawn structures”, “long term plans” and above all “safety of the king”. A better selection of variations provides a greater search depth, enabling Hiarcs 9 to impress by its sharp powerful play against the enemy king as well as its better understanding of endgames, even with unbalanced material." source:http://www.chessbase.com/shop/produ... Csmath I agree with you that Hiarcs 9 is stronger than Fritz 8. I played a match using Hiarcs 8 and Fritz 7 and Hiarcs crushed Fritz 7. Fritz may be good at attacking but the improvements Developer Mark Uniacke has made to the Hiarcs 8 program are impressive (and I can only imagine how much stronger Hiarcs 9 is).Chess knowledge has always been the program’s main strength. The improved search algorithms made it possible to implement even more advanced positional functions, especially in the areas of "typical pawn structures" and "dynamic evaluation of the centre". HIARCS 8 and HIARCS 9 both evaluate important factors like weak squares, king’s safety, piece exchanges, initiative and king attack much more reliably than the predecessor versions. Steve Lopez in an article previewing Hiarcs 9 stated:
"The new version is Hiarcs9; Mark Uniacke has done it once again, bringing us a new version of the program that (in my opinion, anyway) most closely approximates a human's style of chessplay."
source:http://www.chessbase.com/newsdetail... |
|
| Feb-12-05 | | square dance: are there any plans for a fritz 9? |
|
| Feb-12-05 | | bishopmate: I don't know why mine's any different that yours Albertan... but my deep fritz 7 beats my hiarcs 8 regularly, and not just my hiarcs 8 but others' too---mostly, unless they have a super fast computer |
|
| Feb-12-05 | | csmath: Fritz 9 is about to be released as well.
I think currently Shredder 9 is the best thing around. For the second choice I would truly opt for Hiarcs 9. What I like about Hiarcs is a common sense evaluations. This program does not propose "stupid" moves which Fritz and Junior do. Believe me I know what I am talking about, I played dozens high quality standard games using advanced approach with Hiarcs. For competetive chess I would still use Shredder first. |
|
| Feb-12-05 | | square dance: <csmath> do you know what new features fritz 9 will have? any details? |
|
| Feb-12-05 | | Albertan: Bishopmate maybe your Deep Fritz 7 beats Hiarcs all the time because Deep Fritz sees deeper into the position than Hiarcs 8. Do you start the game using an opening book like Powerbook2005? |
|
| Feb-12-05 | | csmath: <<do you know what new features fritz 9 will have? any details?>> No, I don't. There will be probably algorithmic improvements I guess. Fritz 8 introduced new engine (Bilbao) and it was a tangible improvement so I guess you can expect the same this time. You'll get a new opening book and probably some multimedia enhancements as usual. This is just a common sense guess. |
|
| Feb-13-05 | | csmath: <<Bishopmate maybe your Deep Fritz 7 beats Hiarcs all the time because Deep Fritz sees deeper into the position than Hiarcs 8. Do you start the game using an opening book like Powerbook2005?>> I don't see that Fritz goes any deeper than Hiarcs. Shredder does that but the cost is too much prunning and Hiarcs has peculiar advantage over Shredder that it can calculate forced positions better than Shredder. For example most of the time Hiarcs finds problematic mate positions faster than Shredder.
Fritz looks to me rather similar except that it does not have Hiarcs' positional strength. Hiarcs is probably positionally #2 retail engine around. It wouldn't hurt to get a little bigger opening book as well.
Overall Hiarcs is probably one of the two best programs, perhaps only an inch weaker than Shredder.
I just played (and won ;-))) another high quality game in Petroff on ICC using them both. Essential breakthrough in the middlegame was made by Shredder but all the moves were verified by Hiarcs before they were played. |
|
| Feb-14-05 | | Albertan: Oh csmath I always have watched the ply depth on my chess programs and Hiarcs seemsto calculate less ply than Fritz or Shredder 8 or Junior 9. I read something about Hiarcs and how it could find mates and key moves better than any other program. Do any of the other programs play the Cochrane Gambit against your program when you play on ICC and use the Petroff's? |
|
| Feb-14-05 | | csmath: I've never seen that played by machine.
I never play Petroff as black, I play Sicilian exclusively. I am only forced to play against Petroff occassionally. I hate that defence almost as much as I hate French. It is pretty solid though.
Anyway Cochrane Gambit is not very good idea against a machine though it makes a lot of sense against human since people that play Petroff as black probably expect to have a solid game without sudden tactics, this throws them off course completely.Hiarcs is better than Shredder in calculating forced mates, trust me, I have seen that many times. |
|
| Feb-14-05 | | RisingChamp: Actually IMHO the best way to get tactics against Petroff is play something absolutely bizzare which neither chucks material nor makles great concessions e.g like a3 or b3 or something and just get a nonstandard position. |
|
| Feb-14-05 | | Albertan: Rising Champion have you ever played through this amazing game (A Petoff's )? Morozevich vs Topalov, 2001 |
|
| Feb-15-05 | | csmath: <<Actually IMHO the best way to get tactics against Petroff is play something absolutely bizzare which neither chucks material nor makles great concessions e.g like a3 or b3 or something and just get a nonstandard position.>> I think you are right about nonstandard positions, it is always possible to make one, only one has to be very carefull against strong machines, you don't want to hand over initiative. It is very dangerous to allow black to regroup and get the initiative on the king's side, opening files and pushing f-pawn. I think Petroff is very solid defence but nothing more than that. People that play it, play for a draw, and that is sometimes aggravating if you are white. |
|
| Jun-30-05 | | csmath: Palm (Pocket Computer) Hiarcs 9.5 trashes Sergey Volkov +2 =2 -0: Game #2:
1. e4 5 1... e6 2 2. d4 3 2... d5 1 3. Nc3 3 3... Nf6 1 4. Bg5 3 4... Be7 3 5. e5 3 5... Nfd7 1 6. Bxe7 3 6... Qxe7 1 7. Nf3 2 7... a6 3 8. Bd3 12 8... c5 1 9. dxc5 14 9... Nc6 6 10. Na4 26 10... Ndxe5 10 11. Nxe5 11 11... Nxe5 4 12. Be2 4 12... Bd7 6 13. Nb6 7 13... Rd8 2 14. Qd4 5 14... Qf6 15 15. O-O-O 10 15... Bc6 15 16. Rhe1 16 16... Nd7 2 17. Nxd7 7 17... Qxd4 5 18. Rxd4 3 18... Rxd7 2 19. f4 13 19... g6 1 20. g4 19 20... Re7 35 21. c4 12 21... dxc4 12 22. Bxc4 11 22... h5 2 23. g5 12 23... O-O 2 24. b4 8 24... Kg7 6 25. Kb2 7 25... Rc8 10 26. Red1 3 26... e5 13 27. fxe5 7 27... Rxe5 3 28. h4 3 28... Rf5 9 29. Re1 Rf2+ 98 30. Kc3 12 30... Rf3+ 1 31. Kd2 19 31... Rf2+ 2 32. Kd3 5 32... Rf3+ 2 33. Re3 7 33... Rxe3+ 16 34. Kxe3 3 34... Re8+ 1 35. Kd3 4 35... Re1 3 36. Rf4 6 36... Re7 10 37. Rf6 3 37... Rc7 3 38. Kc3 7 38... Rc8 8 39. Rxf7+
1-0
Game 4:
1. e4 0 1... e6 1 2. d4 3 2... d5 1 3. Nc3 2 3... Nf6 1 4. Bg5 3 4... Be7 1 5. Bxf6 23 5... Bxf6 2 6. e5 2 6... Be7 2 7. Qg4 3 7... O-O 11 8. Bd3 4 8... f5 1 9. Qg3 3 9... c5 6 10. dxc5 4 10... Nd7 9 11. c6 23 11... bxc6 3 12. Nf3 10 12... c5 20 13. O-O 12 13... Nb6 43 14. Rad1 17 14... Bd7 3 15. b3 11 15... Qc7 10 16. Nb5 9 16... Bxb5 15 17. Bxb5 1 17... a6 20 18. Bd3 3 18... a5 1 19. c4 13 19... dxc4 7 20. bxc4 16 20... Rad8 10 21. Qf4 9 21... Rd7 14 22. Bc2 11 22... Rfd8 4 23. Rxd7 4 23... Qxd7 6 24. g4 10 24... Rf8 16 25. Rd1 22 25... Qc6 2 26. gxf5 5 26... exf5 1 27. Rb1 5 27... Kh8 7 28. Bxf5 11 28... Nxc4 14 29. Qe4 3 29... Qxe4 33 30. Bxe4 4 30... g6 33 31. Rb7 9 31... Bd8 4 32. Kg2 4 32... Nb6 18 33. e6 9 33... Nc8 14 34. Ne5 3 34... Nd6 6 35. e7 9 35... Bxe7 3 36. Rxe7 1 36... Nxe4 2 37. Nf7+ 5 37... Rxf7 16 38. Rxf7 4 38... Nc3 1 39. Rc7 4 39... Nxa2 3 40. Rxc5 1 40... Nb4 7 41. Rxa5 5 41... Kg7 2 42. Kf3 4 42... h5 14 43. Ra7+ 8 43... Kf6 10 44. Ke4 1 44... Kg5 10 45. f4+ 8 45... Kh6 10 46. Rb7 2 46... Nc6 4 47. Kd5
1-0
The time was 5+3. Sergey Volkov has ELO 2682.
|
|
| Jun-30-05 | | csmath: Games played by Palm HIARCS 9.5 on a Palm Zire 72 in the internet match against Russian GM Sergey Volkov in June 2005. This is probably the first pocket PC program of a superGM strength. |
|
| Dec-20-05 | | jcr2001: So has anyone checked out the new HIARCS 10 yet? |
|
| Dec-20-05 | | karlth: It is supposed to arrive in the mail in a few days. Looking forward to it. |
|
| Dec-23-05 | | karlth: I am watching Hiarcs 10 against Crafty 19.19 and while Crafty is averaging around 2800kN/s Hiarcs is barely managing 10% of that! It is not just that Hiarcs evaluation function must be much more complex but also that the engine must be doing very selective searching as both programs are reaching approx. the same depth in the tree. The Chessbase interface is nearly identical to the one from Fritz 8, including the same old bugs, and of course identical to Fritz 9. They really should begin updating that ancient interface. |
|
| Dec-27-05 | | karlth: Hiarcs pounded both Toga II and Crafty in a 5 minute blitz match. Hiarcs was using the Hiarcs10 tuned opening book while its opponents were using a more basic book, this might well have influenced the result. |
|
| Apr-04-06 | | blingice: Referring to <aw1988>'s post on Spassky vs Fischer, 1972, <aw1988: In any case, for all you kibitzers out there who use machines religiously, do not trust the evaluation. Play out the moves with proper thinking time. One, this is beyond a computer's horizon-- they probably will not think like a human and count Bobby's pawn structure (unless of course maybe HIARCS- but you get the idea).> Why is HIARCS considered special in this sense? I have heard that it is even more human-like in the method that it calculates. |
|
| Jun-05-06 | | Jilted Rook: I was playing HIARCS 10 the other day on "rated mode" @ELO 2200, 4mins + 2secs and it made a fairly basic blunder no master would likely make. Anyone else experienced something similar? |
|
| Jun-06-06 | | notyetagm: Anyone know any details on the newest Hiarcs engine, Hiarcs X50? Thanks. |
|
| Jun-06-06 | | cotdt: Hiarcs is the slowest because it is built in with the most chess knowledge (with the exception of Diep which is not commercial). That is why it has trouble against other chess engines, but in long evaluations it is the best, and I've seen it spit out great moves that other chess engines don't even analyze. |
|
 |
|
|
< Earlier Kibitzing · PAGE 1 OF 3 ·
Later Kibitzing> |
Kibitzing
