Why isn't 5. ... Qc7 marked as an inaccuracy here?

[dryguy]

I'm trying to fully understand how inaccuracies, mistakes, and blunders are determined. I can't read Scala, but from looking at the code for advice.Scala, it seems like moves are marked as inaccuracies when the winning chances drop by between 0.1 and 0.2.

However, looking at this game, on move 5, I calculate a drop of 0.108 based on the evals, yet the move was not marked as an inaccuracy in the analysis.

# I used R for the calculation
# 5. Nf3 { [%eval 2.02] } 5... Qc7 { [%eval 2.69] }
cp <- c(202, 269)
winning_chances <- 2 / (1 + exp(-0.004 * cp)) - 1
cat(winning_chances)
0.3833663 0.4914725

winning_chances_loss <- diff(winning_chances)
cat(winning_chances_loss)
0.1081062

Am I doing the calculation wrong?

[dryguy]

Answered on reddit.