Interesting observation on the latest Scorpio

[Dann Corbit]

I have been watching the latest TCEC bout between Scorpio and Defenchess.
I see something very interesting.
Scorpio seems to evaluate optimistically like Crystal, which is very unusual.
Usually Crystal' evaluation trace is all by itself, but Scorpio seems to have a greater scale to its evaluation now.
Crystal still has more drastic swings to the evaluation (Scorpio is more even keeled)
But I don't remember the scale of Scorpio's eval being that high before

I wonder if Daniel Shawul changed something.

[Daniel Shawul]

I have added something that could be the cause of that. When one side is winning by 90 percent or more only one of the winning variations is searched. This inflates winning scores and is able to find mates better than previous version. But the first two games against defenchess probably due to end game net overestimating R vs B exchanges but haven’t yet looked at the games carefully.

[Chessqueen]

I have been watching the latest TCEC bout between Scorpio and Defenchess.
I see something very interesting.
Scorpio seems to evaluate optimistically like Crystal, which is very unusual.
Usually Crystal' evaluation trace is all by itself, but Scorpio seems to have a greater scale to its evaluation now.
Crystal still has more drastic swings to the evaluation (Scorpio is more even keeled)
But I don't remember the scale of Scorpio's eval being that high before

I wonder if Daniel Shawul changed something.

I also watched a few of the latest game game 7 when the score was 3.5 to 3.5 between Scorpio Vs Defenchess and noticed that Defenchess held Scorpio to an even score up to game 8. As I was watching I thought if the author of Defenchess get a NET NNUE it will reach at least another 20 rating points placing Defenchess as one of the highest rated engines. I Believe version 2.3 is almost a la par with Scorpio
https://tcec-chess.com/live.html