http://www.drivelinemechanics.com/20...ntioned-on-the
Whenever these debates come up on MLB Network they always refer to the teams that base all their rosters off of stats and act like it's terrible but they never name the teams :(
Printable View
http://www.drivelinemechanics.com/20...ntioned-on-the
Whenever these debates come up on MLB Network they always refer to the teams that base all their rosters off of stats and act like it's terrible but they never name the teams :(
That's because they don't exist. Ironically enough, just like the "Straw Hat Man", it's a strawman argument.
This was a terrible segment. Harold Reynolds couldn't put together a coherent sentence at all. It expressed no fact. It constructed strawmen left and right. Nobody takes PECOTA as "gospel". That's just ridiculous. It's a projection system. It doesn't take ITSELF as "gospel." It's "projection" is a weighted mean of a whole slew of different percentile projections for each player. It's a tool, not gospel, and everybody with half a brain knows that.
Any team that attempted to build itself purely on stats these days would collapse. Any team that attempted to build itself while completely ignoring stats would crash and burn as well. This "Scout vs. Stats" nonsense is a media construct that never existed. Anybody serious about player evaluation know that both statistical analysis and scouting are complementary to each other.
Quote:
Originally Posted by HoustonGM
All that's probably true, but I did have a couple of the posters at Fangraphs tell me that the Royals were the worst team in baseball..and then when I asked them to explain the only reason they offered was because CHONE projected them at 72 wins, which was the worst record projected.
I'd say that's a pretty good example of somebody just basing their opinions entirely on a projection system..which is pretty foolish. I do think there are quite a few who simply base everything off of what a projection says....and like you said, that's foolish.
I'd like to read those comments. Mind linking me to the article if possible?
I don't doubt it. There are plenty of foolish people, but the way this segment made it out to be is that anybody who looks at PECOTA thinks its "gospel", and that's just not true.
I don't know where to find it, to be honest...I don't even go on that site that often. Fangraphs isn't my favorite site, I think they are a good example of a site that DOES go too far into the stats and draws entire conclusions from them. I prefer baseballprospectus by a WIDE margin.Quote:
Originally Posted by HoustonGM
The comment wasn't a whole lot...just one of them making a smart-ass comment about the Royals being the worst team in the league because CHONE projects them to be so. I think there are many examples of it on fangraphs site, I don't think they do a very good job of looking at anything else other than stats.
I am sure that segment is awful. Most mainstream media have zero clue about what sabremetrics is about.
...and it's not just "oh, Fangraphs hates the Royals," that's just an example I can recall off the top of my head. They actually write good stuff about the Royals too, but even that is just completely about what some projection says and entirely based off of that.
If it was a commenter that said what you read, that's not Fangraphs.
Fangraphs is a statistical site. That's what they do, so yes, they're going to be heavily statistics based. I think BP is even moreso, personally, but I find no problem with it. I don't think...actually, I KNOW...that neither site denies the importance of scouting/observational data. There's a reason that when a Fangraphs writers analyzes a signing, for example, they put a "range" based on projections, saying something "Player X should be worth 1.5-2.5 wins next year", and that's because projections are just that...projections...not statements of what will happen in the future, and they all have a margin of error.
It was the guy who wrote an article, but in the comments section of an article.Quote:
Originally Posted by HoustonGM
A lot of people who comment on there are similar too though..and that is an example of people who put TOO much importance in the stats.
I ain't saying fangraphs completely sucks, I just like BP better. Fangraphs has interesting stuff, I think they just get a little too deep in the numbers when analyzing the game.
JoePo's take on the MLB Network segment
Quote:
3. But my absolute favorite part is that after saying that stats are ridiculous and that we have it all mixed up, Harold promptly proceeds to make HIS OWN STATISTICAL PREDICTIONS. He says that PECOTA is underselling Youkilis’ power, his ability to hit for average and overstating Albert Pujols’ RBI number. I am constantly amazed how often this happens, how often someone will bash the use of baseball statistics by quoting other baseball statistics, or picking some baseball statistic that doesn’t even exist (a statistic that, apparently, does not count any of the things that matter) and then ripping apart this imaginary statistic.
Posnanski is, as usual, right on the money. I like the fact that he says that stats don't predict the future AS WELL as they explain the past...that's one of the big things that I think some "statheads" miss....but yes, that segment was a complete abomination.
Whew! Good thing we're not, y'know, defensive about it or anything. :)Quote:
Originally Posted by HoustonGM
Yeah, I'm defensive, because I prefer actual facts instead of babbling nonsense in my baseball analysis. :)Quote:
Originally Posted by oriole^
Excellent article on this. It's subscriber only, but here's two quotes:
Quote:
The other side of this, however, is that players absolutely value non-performance characteristics much more highly than is sensible. This is understandable, because to them the guys around them are not just baseball players, but co-workers, travel partners, drinking buddies, and confidantes. They are around these people 12 hours a day for two-thirds of the year. They're using a completely different set of criteria to evaluate their worth. It's not an additional set; it's a different one. To the extent that these people evaluate each other on their work product, it's on an extremely micro level that doesn't translate well, if at all, to an overall assessment of performance.
Quote:
Players bring a lot to a studio show; Baseball Tonight used Reynolds to great effect, getting him out from behind the desk and having him demonstrate and explain the physical aspects of making plays around the bag at second. He knows the game as he played it, he's extremely engaging and entertaining, and you can build a show around him. I want his voice, and Larkin's, and Casey's, and Al Leiter's, but I want them telling me things that they know that I don't. You wouldn't ask Joe Sheehan about the footwork around second base, or Clay Davenport about the effects of different finger pressures on breaking balls. So don't ask Harold Reynolds what he thinks about a PECOTA projection. You need other perspectives.
Good quotes HGM and I actually do like watching the diamond demos, that's one of my favorite parts about the show.