Csdrhrt

I'm using stockfish-9-bmi2 on my laptop to provide me move by move evaluation of games. I've set Stockfish to evaluate to depth = 25 plys. I find that the evaluation provided oscillates move-on-move after a certain number of moves. Below is the plot of the evaluation:

In the above plot, the ply-count is on the x-axis, the evaluation from Stockfish is on the primary y-axis (left), the Mate-In number is on the secondary y-axis (right). The Blue line with white circles is the plot of the evaluation and the red circles depict the mate-in number.

However, for the same game, when I jot down the evaluation from lichess.com's analysis board, which is also using Stockfish 9, I get a much more stable evaluation move-on-move. Lichess.com's analysis board also uses depth = 25 for most moves but can use a different number for certain moves. Below is the plot for the same:

In the above plot, the Blue line depicts the evaluation. The break in the trend is because of a Mate-In-3 on ply count = 63.

Considering that both methods use Stockfish 9 and is set to depth = 25, can someone help me with why the results on my laptop oscillates so much?

My Laptop's configuration:

• OS: MacOS High Sierra


• Processor: Intel core i7 2.2 GHz


• RAM: 16GB


• Stockfish: stockfish-9-bmi2

The game used is Praggnanandhaa vs Vishy Anand - Tata Steel Chess India Blitz 2018

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6

5. Nc3 a6 6. a3 g6 7. Be2 Bg7 8. Be3 O-O

9. Qd2 b5 10. O-O-O Bb7 11. f3 Nbd7 12. h4 Ne5

13. Bh6 Bxh6 14. Qxh6 Rc8 15. g4 Rxc3 16. bxc3 Qa5

17. Nf5 gxf5 18. gxf5 Qxc3 19. Rdg1+ Ng6 20. fxg6 fxg6

21. h5 Bxe4 22. fxe4 Qa1+ 23. Kd2 Nxe4+ 24. Ke3 Qc3+

25. Bd3 Qd2+ 26. Kxe4 Qxh6 27. hxg6 Qf4+ 28. Kd5 h6

29. g7 Rc8 30. Ke6 d5 31. Rg6 Rc6+ 32. Kxe7 Rxg6

33. Bxg6 Kxg7 34. Bd3 Qe5+ 35. Kd7 h5 36. Rg1+ Kf6

37. Rg6+ Kf7 38. Rh6 a5 39. Kc6 b4 40. axb4 axb4

41. Kd7 Kg7 42. Re6 Qd4 43. Ke8 Qg4 44. Ke7 h4

45. Be2 Qg5+ 46. Kd6 h3 47. Re7+ Kf8 48. Re5 Qf6+

49. Kxd5 h2

P.S.: I have also tried using stockfish-9-64 instead of stockfish-9-bmi2 and the results are similar

If I understand correctly your x axis is the moves. Your script implied Anand and Prag were blundering every time they make a move,

Clearly you have a bug in your script. Stockfish engine always give you a score relative to the player making the move, NOT white. You need to multiply the score by minus one if it’s Black to move.

Please do this and you will be fine.