A very interesting thread. In 960 blitz I bet that blunders are more frequent than in classical blitz.
A new topic to study related with yours.
I was there in Freestyle Open at Karlsruhe and I asked myself as a curosity in what move all games diverged, I mean in what moves all games were different and see the evolution, in the first move 100% of games have the same position, and in move x each position is unique.
I saw around me and in very few moves all games were different
@justiniano565 said in #11:
> A very interesting thread. In 960 blitz I bet that blunders are more frequent than in classical blitz.
> A new topic to study related with yours.
> I was there in Freestyle Open at Karlsruhe and I asked myself as a curosity in what move all games diverged, I mean in what moves all games were different and see the evolution, in the first move 100% of games have the same position, and in move x each position is unique.
> I saw around me and in very few moves all games were different
That's an interesting question, the Lichess opening explorer also works for 960 games which should give a decent sample size for each position, even though the games will be of lower quality.
An extract of a GM's chess insights for the Pieces moved by Variant. In the movement across phases, the largest range difference was for Pawns, with a 6% point difference between Blitz (27.9%) and Chess960 (33.9%).
... /piece/variant/variant:blitz
... /piece/variant/variant:chess960
Variant Pawn Knight Bishop Rook Queen King
Chess960 29.1% 18.7% 13.7% 14.8% 13.7% 10%
Blitz 26.4% 16.3% 15.5% 16.6% 12.2% 12.9%
Compare pieces moved by phase:
... /piece/phase/variant:blitz
Game phase Pawn Knight Bishop Rook Queen King
Opening 45.9% 24.9% 17.3% 2.1% 4.9% 4.8%
Middlegame 22.6% 17.6% 18.5% 17.3% 17.3% 6.7%
Endgame 18% 9.2% 10.9% 25.4% 11.1% 25.5%
... /piece/phase/variant:chess960
Game phase Pawn Knight Bishop Rook Queen King
Opening 51.1% 24.4% 13.5% 2.1% 6.3% 2.5%
Middlegame 22.5% 19.8% 16.8% 16% 16.6% 8.4%
Endgame 17.2% 10.5% 8.6% 26.3% 16.6% 20.9%
Compute the difference in range per piece:
Piece Blitz Range (%) Chess960 Range (%) Difference (%)
Pawn 27.9 33.9 6.0
Knight 15.7 13.9 1.8
Bishop 7.6 8.2 0.6
Rook 23.3 24.2 0.9
Queen 12.4 10.3 2.1
King 20.7 18.4 2.3
This suggests that Pawn usage varies more dramatically across phases in Chess960, likely due to its shuffled starting positions requiring more early structural clarity. I wonder if a Stockfish dev version would have 6% pawn difference.
https://lichess.org/@/YoBot_v2