Posted by sjpoker on 6/21/2016 2:35:00 PM (view original):
Posted by bad_luck on 6/21/2016 1:15:00 PM (view original):
Posted by MikeT23 on 6/21/2016 1:10:00 PM (view original):
Posted by bad_luck on 6/21/2016 12:05:00 PM (view original):
Posted by MikeT23 on 6/21/2016 12:02:00 PM (view original):
Posted by bad_luck on 6/21/2016 11:58:00 AM (view original):
Posted by MikeT23 on 6/21/2016 11:47:00 AM (view original):
So, if I didn't watch the game because the stats would tell me who was productive, how would I know who reached on an error and who, you know, helped his team score runs?
What?
If you want to know what happened in an individual game, you should watch it.
If you want to know who was good over a span of time, you need stats.
Do you not care who helped their team win?
Yes, I do care. Are we taking about the storyline of one game or determining who helped their team win more games over the course of a season?
I've already answered this.
We could be talking about one game, one week, one month, one season or an entire career. You only seem to care about season/career. Is that because you don't like to watch games?
If your goal is to evaluate a major league player, one game doesn't tell you anything. Or do think 4 PAs is all you need to see?
If your goal is to know what happened in a certain game, watch the game and don't worry about this stuff.
Which is your BIGGEST flaw. Status alone cannot be used to evaluate a players worth. And next, you will tell me you know that. But what I will argue is that stats are a LOT farther off from evaluating a players worth than you think.
Based on the contracts that players get, I would argue that you're incredibly wrong. Overwhelmingly, the players with the best stats get the biggest contracts. You can fairly accurately predict contracts without knowing anything about the players involved - in fact, MLBTradeRumors has an algorithm that does this, and does a pretty good job. Sometimes there are perception-based adjustments. Samardzija is a reasonable example of a guy who people saw a lot of upside in whose stats didn't necessarily justify the contract the Giants gave him, and so far that decision has been working out fairly well (still very early going in the contract, but early returns are positive). But in general, the purely stats-based algorithm is within a few percentage points of AAV, and it gets more contract lengths right than any human evaluator I've seen. In the end, the computer can process more data, more historical comparisons, than any person. And that makes it more accurate.
Pretty strong evidence that stats absolutely
do play the overwhelmingly dominant role in the evaluation of a player's worth. At least to Major League front offices, which I will tend to trust over a guy posting on the WIS forums who, frankly, has been abjectly wrong about a lot of stuff.