Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / Rybka 2.3.2a module in Fritz 11 (cloning)
- - By Pia (****) Date 2007-12-12 16:30
6r1/4k3/3p4/2nP3P/p1P1P3/4N3/5p2/K6R w - - 0 1


Analysis by Rybka 2.3.2a 32-bit : (64MB)

49.Rf1 Rg1
  µ  (-1.35)   Depth: 5   00:00:00
49.Rf1 Rg1 50.Kb2
  µ  (-1.13)   Depth: 6   00:00:00  3kN
49.Rf1 Rg1 50.Kb2 Rxf1
  µ  (-1.35)   Depth: 7   00:00:00  4kN
49.Rf1 Rg1 50.h6 Kf6 51.e5+ dxe5
  -+  (-1.44)   Depth: 8   00:00:00  7kN
49.Rf1 Rg1 50.e5 Rxf1+ 51.Nxf1 dxe5 52.Kb2 Kf6 53.h6
  µ  (-1.13)   Depth: 9   00:00:00  15kN
49.Rf1 Rg1 50.e5 Rxf1+ 51.Nxf1 dxe5 52.Kb2 Kf6 53.Ka3
  µ  (-0.90)   Depth: 10   00:00:00  27kN
49.Rf1 Rg1 50.e5 Rxf1+ 51.Nxf1 dxe5 52.Kb2 e4 53.h6 Kf6
  µ  (-1.18)   Depth: 11   00:00:00  35kN
49.Rf1 Rg1 50.e5 dxe5 51.Kb2 a3+ 52.Kxa3 Rg3 53.Rxf2 Rxe3+ 54.Ka2 Rc3
  µ  (-1.12)   Depth: 12   00:00:01  60kN
49.Rf1 Nxe4 50.h6 Kf6 51.h7 Rf8 52.h8Q+ Rxh8 53.Ng4+ Kf5 54.Nxf2 Nd2 55.Re1 Rh2
  µ  (-1.33)   Depth: 13   00:00:02  128kN

Analysis by Fritz 11: (64MB)

49.Rf1 Rg1 50.Kb2 Nxe4 51.h6 Kf6 52.Kc2 Rxf1 53.Nxf1 a3
  µ  (-1.18)   Depth: 9/23   00:00:00  68kN
49.Rf1 Rg1 50.Kb2 Rxf1 51.Nxf1 Kf6 52.h6 Kg6 53.Kc3 Kxh6 54.Kd4
  µ  (-1.18)   Depth: 10/18   00:00:00  108kN
49.Rf1 Rg1 50.Kb2 Rxf1 51.Nxf1 Kf6 52.h6 Kg6 53.e5 dxe5 54.Ne3 e4
  µ  (-1.29)   Depth: 11/21   00:00:00  163kN
49.Rf1 Rg1 50.Kb2 Rxf1 51.Nxf1 Kf6 52.Ka3 Kg5 53.e5 dxe5 54.Kb4 Nd3+ 55.Kxa4 Kxh5
  µ  (-1.17)   Depth: 12/21   00:00:00  363kN
49.Rf1 Rg1 50.Kb2 Rxf1 51.Nxf1 Kf6 52.Ka3 Nxe4 53.Kxa4 Kg5 54.Ne3 Kxh5 55.Kb4 Nd2
  µ  (-1.35)   Depth: 13/27   00:00:00  734kN
49.Rf1 Rg1 50.e5 dxe5 51.Kb2 Ne4 52.h6 Kf6
  -+  (-1.41)   Depth: 14/31   00:00:01  1672kN
49.Rf1 Rg1 50.e5 dxe5 51.h6 Kf6 52.Ka2 Rxf1 53.Nxf1 Kg6 54.Ne3 Ne4 55.Ka3 Kxh6 56.Kxa4 Kg5
  -+  (-1.59)   Depth: 15/31   00:00:02  2521kN
49.Rf1 Rg1 50.e5 dxe5 51.h6 Kf6 52.d6 Kg6 53.Kb2 Rxf1 54.Nxf1 Kxh6 55.Ka3 Kg5 56.Kb4 Nd3+ 57.Kxa4 e4
  -+  (-1.59)   Depth: 16/30   00:00:04  4708kN
49.Rf1 Nxe4 50.h6 Kf6 51.h7 Rf8 52.h8Q+ Rxh8 53.Ng4+ Kf5 54.Nxf2 Nd2 55.Re1 Rh2 56.Re2 Nxc4 57.Ka2 Rg2 58.Rc2 Ne5
  µ  (-1.18)   Depth: 17/33   00:00:09  10812kN

There are many positions where both engines displays very similar lines. But even this one example is mathematically enough to raise questions.

Do Vasik sell Rabka parts?
Or it is a world-wide plot against Rybka?
Parent - - By turbojuice1122 (Gold) Date 2007-12-12 16:55
No, this is quite different from what we were seeing with Strelka, where in a decent percentage of positions, not only was there one line in the eval that matched, as you have here, but all of the lines match, as do all of the evaluations--here, the evaluations are pretty different.  This is nothing like what we have with the Rybka clone called Strelka.  Also, note that after Nxe4, the rest of the moves are forced (from a computer standpoint), so the fact that they share the same lines (at different depths, no less, quite unlike the situation with Strelka and Rybka 1.0 beta) is pretty much expected.
Parent - - By Pia (****) Date 2007-12-12 17:56
Yes, it's not Strelka case, it's different but also very suspicious.

r1b2rk1/5pbp/p2p4/1p3P2/2P1p3/N2B1P2/PP3P1P/R4RK1 w - - 0 1


Analysis by Rybka 2.3.2a 32-bit : (512 MB, Display PV Tips)

2.Bxe4 Rb8 3.Rab1 Re8 4.Rfd1 Be5 5.Kg2 Bd7
  ±  (0.79)   Depth: 5   00:00:00
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.c5 Bxa2
  ²  (0.57)   Depth: 6   00:00:00  3kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.b3 f5 7.Bd5
  ²  (0.60)   Depth: 7   00:00:00  4kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.b3 Rfd8 7.Rbd1 Bd4
  ²  (0.57)   Depth: 8   00:00:00  6kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.b3 Rfd8 7.Rfd1 f5 8.Bc6
  ²  (0.61)   Depth: 9   00:00:00  10kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.b3 Rfd8 7.Rfd1 Bf8 8.Rxd8 Rxd8 9.cxb5 Bxa3
  ²  (0.62)   Depth: 10   00:00:00  16kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.b3 Rfd8 7.Rfd1 Bf8 8.Rxd8 Rxd8 9.cxb5 Bxa3 10.bxa6 Rd2
  ²  (0.48)   Depth: 11   00:00:00  26kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.b3 Rfd8 7.Rfd1 Bf8 8.Rxd8 Rxd8 9.cxb5 f5 10.Bc6 Bxa3 11.bxa6
  ²  (0.61)   Depth: 12   00:00:01  45kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.b3 Kh8 7.f4 Rg8 8.Kh1 Bd4 9.Nc2 Bc5 10.Ne3
  ²  (0.61)   Depth: 13   00:00:01  85kN
2.Bxe4 d5 3.Bxd5 Bxf5 4.Bxa8 Rxa8 5.Rae1 Bd3 6.b3 Bf8 7.Nb1 bxc4 8.bxc4 Bxc4 9.Nc3 Bg7 10.Ne4 Bxf1 11.Kxf1 Be5 12.Kg2 Rd8 13.Rc1
  ²  (0.48)   Depth: 14   00:00:04  245kN
2.Bxe4 d5 3.Bxd5 Bxf5 4.Bxa8 Rxa8 5.Rae1 Bd3 6.b3 Bf8 7.Nb1 bxc4 8.bxc4 Bxc4 9.Nc3 Bg7 10.Ne4 Bxf1 11.Kxf1 Be5 12.Kg2 Rc8 13.Ng5 Rc5
  ²  (0.47)   Depth: 15   00:00:08  455kN
2.Bxe4 d5 3.Bxd5 Bxf5 4.Bxa8 Rxa8 5.Rae1 Bd3 6.b3 Bf8 7.Nb1 bxc4 8.bxc4 Bxc4 9.Nd2 Bxf1 10.Nxf1 Rb8 11.Ng3 Kg7 12.Kg2 Ba3 13.f4 Rb2 14.Nf5+ Kg6
  ²  (0.45)   Depth: 16   00:00:14  774kN
2.Bxe4 d5 3.Bxd5 Bxf5 4.Bxa8 Rxa8 5.Rae1 Bd3 6.b3 Bf8 7.Nb1 bxc4 8.bxc4 Bxc4 9.Nd2 Bxf1 10.Nxf1 Rb8 11.Re4 Kg7 12.Rg4+ Kf6 13.Ra4 Rb6 14.Ne3 Ke6 15.Re4+ Kd7
  ²  (0.42)   Depth: 17   00:00:24  1359kN
2.Bxe4 d5 3.Bxd5 Bxf5 4.Bxa8 Rxa8 5.Rae1 Bd3 6.b3 Bf8 7.Nb1 bxc4 8.bxc4 Bxc4 9.Nd2 Bxf1 10.Nxf1 Kg7 11.Rb1 Bc5 12.Ng3 Rc8 13.Kg2 Kg6 14.Ne4 Bd4 15.Rb7
  ²  (0.45)   Depth: 18   00:00:37  2050kN
2.Bxe4 d5 3.Bxd5 Bxf5 4.Bxa8 Rxa8 5.Rae1 Bd3 6.b3 Bf8 7.Nb1 bxc4 8.bxc4 Bxc4 9.Nd2 Bxf1 10.Nxf1 Rb8 11.Ng3 Rb2 12.Re2 Rxe2 13.Nxe2 Bc5 14.Kg2 Kg7 15.Nf4 Kf6 16.Nd3 Bd6
  ²  (0.44)   Depth: 19   00:01:01  3382kN

Analysis by Fritz 11: (512 MB)

2.Bxe4 d5 3.Bxd5 Rb8 4.Rfb1 b4 5.f6 Bxf6 6.Nc2
  ±  (0.95)   Depth: 9/19   00:00:00  43kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.c5 Bxa2
  ±  (0.87)   Depth: 10/19   00:00:00  63kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rfb1 bxc4 5.f6 Bxf6 6.Nxc4 Bf5 7.Be4
  ²  (0.69)   Depth: 11/19   00:00:00  93kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.b3 Rfd8 7.Rbd1 Bd4
  ±  (0.72)   Depth: 12/21   00:00:00  153kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.b3 Rfd8 7.Rbd1 Bb2 8.Rxd8+
  ±  (0.75)   Depth: 13/26   00:00:00  280kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.b3 Rfc8 7.Bd5 Re8 8.Bc6 Rec8
  ±  (0.84)   Depth: 14/24   00:00:00  567kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.b3 bxc4 7.Nxc4 f5 8.Bd3 Bd5 9.Nd6 Be5
  ±  (0.75)   Depth: 15/25   00:00:01  989kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.b3 Rfe8 7.cxb5 axb5 8.Rbd1 b4 9.Nc4 Rbc8 10.Nd6
  ±  (0.78)   Depth: 16/27   00:00:02  1850kN
2.Bxe4 d5 3.Bxd5 Rb8 4.Rab1 Bxf5 5.Be4 Be6 6.b3 Kh8 7.cxb5 axb5 8.Rfd1 Be5 9.Kh1 Rg8 10.Rbc1 b4
  ±  (0.78)   Depth: 17/28   00:00:03  3180kN
2.Bxe4 d5 3.Bxd5 Bxf5 4.Bxa8 Rxa8 5.Rae1 Bd3 6.b3 Bb2 7.Nc2 bxc4 8.bxc4 Bxc4 9.Re7 Bf6 10.Rd7 Bxf1 11.Kxf1 Re8
  ²  (0.60)   Depth: 18/29   00:00:08  6911kN
2.Bxe4 d5 3.Bxd5 Bxf5 4.Bxa8 Rxa8 5.Rae1 Bd3 6.b3 Bb2 7.Nc2 bxc4 8.bxc4 Bxc4 9.Re7 Bf6 10.Rd7 Bxf1 11.Kxf1 Rc8 12.Ne3
  ²  (0.54)   Depth: 19/32   00:00:15  13858kN
2.Bxe4 d5 3.Bxd5 Bxf5 4.Bxa8 Rxa8 5.Rae1 Bd3 6.b3 Bb2 7.Nc2 bxc4 8.bxc4 Bxc4 9.Re7 Bf6 10.Rd7 Bb5 11.Rd6 Bxf1 12.Rxf6 Kg7
  ²  (0.63)   Depth: 20/31   00:00:25  22452kN
2.Bxe4 d5 3.Bxd5 Bxf5 4.Bxa8 Rxa8 5.Rae1 Bd3 6.b3 Bf8 7.Nb1 bxc4 8.bxc4 Bxc4 9.Nd2 Bxf1 10.Kxf1 Bb4 11.Rd1 Rc8 12.Ke2 Kg7 13.Rg1+
  ²  (0.43)   Depth: 21/35   00:00:46  42214kN
2.Bxe4 d5 3.Bxd5 Bxf5 4.Bxa8 Rxa8 5.Rae1 Bd3 6.b3 Bf8 7.Nb1 bxc4 8.bxc4 Bxc4 9.Nd2 Bxf1 10.Kxf1 Bb4 11.Rd1 Rc8 12.Ne4 Rc2 13.Rd4 Be7
  ²  (0.42)   Depth: 22/36   00:01:24  78101kN
2.Bxe4 d5 3.Bxd5 Bxf5 4.Bxa8 Rxa8 5.Rae1 Bd3 6.b3 Bf8 7.Nb1 bxc4 8.bxc4 Bxc4 9.Nd2 Bxf1 10.Nxf1 Rb8 11.Ng3 Rb2 12.Re2 Rxe2 13.Nxe2 Kg7
  ²  (0.45)   Depth: 23/35   00:02:29  140mN

Long line, don't you think? Most likely, it's too long.
Parent - By Vempele (Silver) Date 2007-12-12 18:49
That's 5 non-forced moves. Why don't you try a long mate problem next? See if both engines play the exact same mate in 14 (with identical scores)?
Parent - By turbojuice1122 (Gold) Date 2007-12-12 19:07
Neither the previous one nor this one are suspicious in the least bit (as much as I would like to look for an excuse to harp on a Chessbase product)--while this one has similar scores toward the end, it's at very, very different depths, and the scores on the depths leading up aren't identical at all.  Of course, also most of the moves in your bold are forced.  Furthermore, when engine moves aren't forced, they're going to tend to agree about three-quarters of the time from engine to engine.  Thus, to get, say, six identical non-forced moves in a particular comparison is around 0.75^6 = 18%.  Thus, if you try 6 different positions, you will probably find at least one in which you have a situation such as what you've given here.  This is different from what we've been talking about previously, where we tend to be dealing with lots of non-forced lines giving same evals at same depths, or even more damning, the exact same behavior from depth to depth in both the evaluations and the given lines (for example, both engines oscillating back and forth between, say, 0.55 and 0.52 in the evaluations, something that I've seen lots of times with Strelka and Rybka 1.0).
Parent - - By Gaмßito (****) Date 2007-12-12 19:34
Yes, it's a long line by Fritz, but this is one of the new features in Fritz 11, to displays its thinking as ' Rybka PV Tips' style always. (it's really their default option now). This is why you will see that it always will have very long lines in comparison with the previous versions. In my opinion, this is the best way for a program to display its thinking. Very long lines and in a very constant way. I will be happy if Rybka 3 will have this as default too. 
 
About your position, I also think that it is not suspicious at all. Like Turbo said, it is a position where from the computer point of view, there are many forced moves. You can also find other programs where they will coincide with almost the same thoughts of Rybka and Fritz.  

What is true and good is that Rybka has been the bank of tests in the Chessbase team. Thanks to Vas, we can hope a much better competence by the rivals. Many are trying to hunt Rybka now and they are coming closer. Rybka 3.0 should arrive and put again the things in their place.

Regards,
Gambito.
Parent - By Pia (****) Date 2007-12-12 20:53
OK, I was just paranoid. Thanks.
Up Topic Rybka Support & Discussion / Rybka Discussion / Rybka 2.3.2a module in Fritz 11 (cloning)

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill