Originally posted by Phlabibit
This user is trying to say they use an engine to learn improved moves and they go into a database for later use. They even say they don't have many of these moves, but for some reason these moves are to blame for this user 'looking' like an engine user.
When it comes down to games being played, how often are you still in data in middle to end game play ...[text shortened]... The decision of the game mods are not based on a move here or there and a game here or there.
I can certainly agree that engine matches scored in the middle or end game are
(or at least should be) significantly more damning than matches scored whilst in the opening as in the former case there is a good chance those matches were achieved as a result of legitimate past analysis whilst that clearly is not possible in the middle and end game.
In the end game of course matches can come up that are simply bad, pointless moves, that either simply delay mate
(maybe even by pointlessly giving up material) or waste opportunity because the winning or drawing line is behond the computer horizon.
I can think of 2 games (both draws) I have played here where exactly that happened. In the first I had a K & 5 pawns against K, B & 3 pawns but I could not save my last pawns. Then I realised I could give up all 5 pawns just for 2 of his as that left him a B & RP (of the wrong colour) and I could get my King back. An easy draw and I could force it. Putting it on Fritz, after the game, and it is about 10 moves before Fritz changes its assessment from -5 to 0. In the second I had 2N and 2P (1 on the 6th rank and supported by my King) vs R & P. Fritz subsequently rated this about +3 but suggested lines that would have allowed my opponent to sacrifice his Rook for my last 2 pawns never spotting that if this happened it would immediately be drawn. My play was obviously determined by my need to use my Ns to shield those last pawns as I tried to shepherd them through (and failed).
If any person played the losing / drawing moves in those 2 games this would, in my opinion, be a strong indicator of engine use and I can think of a load of similar examples in the end game. In fact it is in endgame play where the most significant indicators of engine use are likely to be found either because the computer cannot grasp the correct plan or because (as a result of tablebases being used) it grasps it too perfectly.
In the middle game move perfect tactics that start with a difficult to see 1st move can also be a strong indicator. Move perfect tactics that start with obvious sacrifices prove nothing much as although a human may not have seen it all correctly at the start they may as the combination progresses (although over many games expect the occasional error).
In the opening matches mean far less. For example in the recent Grob tournament I decided I was going to play an obscure black defense (Nh3). My 3 books on the Grob had a little but not much on this so before the tournament started I sat down and aided by Fritz analysed out the first 12 moves of as many lines I could. Then I played them. After round 1 I analysed out new lines I had faced ready for round 2. This was legitimate engine use and it helped me get the first dozen moves of these games correct and the matches prove nothing.
Of course the mods may say to us that they realise all these things and that they don't make undue inferences from matches in the opening but do make those inferences when the same continues to happen in the middle or end game. If that is the case here then they have it 100% right as no one can "learn" middle game / end game play from an engine without actually using an engine.