[rjf72]

They were looking at 3 0 games (3 minutes for the entire game with 0 seconds added per move) with a total of 6 or fewer pieces, including kings. So stuff like King + Rook + Bishop vs King + Rook + Knight. There are a countless number of confounding variables in this study that do not seem have to been given sufficient consideration and this phenomena is likely a product of that. These sort of 3 0 endgames are generally time scrambles. The vast majority of 6 man endings reached without one player having already resigned are going to be technical draws, so the players are generally playing to win on time - not on the board.

A critical factor that the study did not consider at all was your opponent's remaining time. The stronger player is generally going to be the player that has more time left, often much more time. If you have 20 seconds and your opponent has 2 seconds left the easiest way to wrap up the game is to just make a move, any move. This means the quality of move is going to go down in a vacuum. By contrast you won't find this situation as frequently from the weaker player since they'll rarely be meaningfully up on time.

Another related pattern is that in a dead drawn time scramble a weaker player is going to be more likely to make a blunder, such as hanging a piece. And not exploiting that move (such as by taking the piece) would itself be a blunder. But in a time scramble with just 2 pieces vs 2 pieces you're often just making semi-random moves as quickly as possible - meaning you will miss the hanging piece. Again for similar reasons this is not an issue you'll find as commonly from the perspective of the weaker player both because they'll less often be substantially up on time and because the stronger player is less likely to randomly blunder, meaning just making a random move is itself less likely to be a blunder.