Posted by tampafla on 5/6/2018 7:33:00 PM (view original):
Posted by mullycj on 5/6/2018 3:48:00 PM (view original):
I did a study a few seasons ago and found the 3pt shooting in the last minute was actually higher against a +5 defense than the other defenses played during the rest of the game. Sure the sample size was still smallish (50 shots or so). But I also have a suspicion they purposely increase the chance of making a 3 at the end of the game no matter what you do anyway.
To answer your question. If a team needs to make a 3 to tie the game I am going to instinctively send out the defense that is supposed to stop a 3 whether it actually works or not. If they make it while I run a -2 then I would blame myself.
Well the end of game logic you mention makes sense. It’s exactly the way it works in the GD game. They force unrealistic upsets and outcomes into the game. But gotta take the good with the bad. I do agree that bad defenders shouldn’t impact shots AS MUCH AS a good defender. But if the effectiveness is tied just to defensive skill ability then that’s unrealistic. Even more so then the CS agent saying that was the fix that was implemented. That is an awful fix to the engine if true. Formation IQ should mean something.
I don’t see any reason to suspect there is end of game logic “forcing” upsets. Although, it can sometimes work to “force” close games, but that’s not the intention. The engine just works to compensate for early runs, which is why if you go on an uncommonly good run against an opponent, you can be almost assured the engine will snap you back; either you’ll go cold for a spell, or your opponent will get hot, or some combination. It regresses to the mean, which usually works *against* upsets. Different concept than what mully is talking about, although there is probably some cross reference in many instances (ie, the late game hot streak is really the engine regressing to the mean).