SI's Baseball Preview

[AngelAstro]

Originally posted by WiffleWOOD


(disclaimer: the following is not an idictment of all non-statheads. it is a general response to the question. don't hate me.)

and also, because people like to think that they are able to formulate their own opinion of the game, and will thus stick with what their "eyes" tell them, even if it flies in the face of quantifiable date like statistics.

nobody wants to think that the way they are seeing is wrong. hence, statistical data (like the data that shows jeter is poor defensively) is ignored.

Funny thing is that I've always thought that Jeter is poor defensively by eye, now there are stats to prove it.

I think the real knock against Sabermetrics right now is that it doesn't work because Beane's A's haven't won the World Series. The problem with this statement is that teams that follow sabermetric type principles have won the World Series - see Yankees of 96-00, with the exception that those teams weren't publicized has sabermetric teams. But the principles of Moneyball and Stick Michael's teams were the same. I wonder if a book like Moneyball was written about the mid-90's Yankees, how sabermetrics would be viewed today.

[WiffleWOOD]

Originally posted by AngelAstro


Funny thing is that I've always thought that Jeter is poor defensively by eye, now there are stats to prove it.

I think the real knock against Sabermetrics right now is that it doesn't work because Beane's A's haven't won the World Series. The problem with this statement is that teams that follow sabermetric type principles have won the World Series - see Yankees of 96-00, with the exception that those teams weren't publicized has sabermetric teams. But the principles of Moneyball and Stick Michael's teams were the same. I wonder if a book like Moneyball was written about the mid-90's Yankees, how sabermetrics would be viewed today.

agreed on all points.

[AngelAstro]

Originally posted by Monkeyman


Way to comepletely ignore what I said. I know I harped on for so long about, what was it, "style points?" My point was that Mussina appeared to have some insight as to how different people may respond in pressure situations. God forbid we listen to an actual baseball player about baseball once in a while.

I have no problem with quantifiable data, but I think there are flaws in Sabermetric analysis. For instance, things like K-rates dropping, they just assume a guy will get worse, because there is a correlation. But a correlation isn't the same as causation. Its like people who play violent video games and watch violent movies. There is a correlation that they are more violent people. So should we just assume that if someone plays violent video games or watches violent movies that they will become violent? Or should we find the actual cause of both? Maybe people who are more comfortable with violence would be more likely to watch violent movies, and also would be more likely to commit violence. Instead of saying "second baseman decline earlier" or "lowered k-rates lead to higher ERAs" find out WHY. That would be a hell of a lot more effective in putting together a team.

You just described exactly why sabermetricians use correlations. A sabermetrician sees that Barry Zito's K-rate has dropped over the past three years. This puts up the red flag that this pitcher may be in decline. Thus, Zito shouldn't be signed to a 10 year, 200 million dollar contract, because there is a good chance (the correlation that declining K-rates of pitchers is associated with increasing ERA from stats of the thousands of pitchers before him) that he won't be winning 20 games a season 2 or 3 seasons from now. Now, this does not mean that Zito will decline, just that there is a chance. The next thing that the GM has to decide is whether or not Zito would be worth the years and money on his next contract. Of course, any smart GM would not only use stats, but would also use scouting to really determine if Zito would be worth signing to a large deal.

[Monkeyman]

Originally posted by WiffleWOOD


(disclaimer: the following is not an idictment of all non-statheads. it is a general response to the question. don't hate me.)

and also, because people like to think that they are able to formulate their own opinion of the game, and will thus stick with what their "eyes" tell them, even if it flies in the face of quantifiable date like statistics.

nobody wants to think that the way they are seeing is wrong. hence, statistical data (like the data that shows jeter is poor defensively) is ignored.

My main problem with defensive statistics is its really hard to effectively and objectively quantify something as erratic and quirky as defense. That being said, when EVERY statistic consistently tells me Jeter is a poor SS, I'm willing to accept that he's not what I used to think. However, when I see a player who's RF and ZR vary from year to year greatly from one side of average to another, it makes it difficult to trust those numbers. I think offensive and pitching statistics are pretty much on target, though there are some inherent flaws in park-adjusted stats, which i discussed in another thread, but for the most part, they can pretty much be trusted. I'm on the fence about defensive stats, however, and probably will be for a long time.

[AngelAstro]

Originally posted by Monkeyman


Concerning clutch hitting, Mussina's quote tells it well: "As a pitcher I do know there's something to clutch hitting. I see some guys come up in pressure sitations, and you can tell they're just trying not to strike out. I'm trying to take advantage of a hitter's anxiety by getting him out on his front foot to get him to roll over and hit a grounder or pop it up. The good ones don't let that happen as easily. They don't expand the strike zone, and they're willing to hit the ball up the middle or to the opposite field. Magglio Ordonez is a good example." Try quantifying that. You can't. Mussina is a much better source of what a batter is actually doing in the real world than James is, because he's watching it happen, and he's figured out how to respond to it. Just observing the numbers isn't as useful as actually comprehending what causes those numbers.

I think the problem with Mussina's statement is that we don't know with whom he is comparing Maggs. If he is comparing Maggs to Adam Rowland (White Sox CFer) I am sure he would notice such a difference, because Maggs is a much, much, much better hitter than Adam Rowland, in fact Maggs is one of the elite hitters in the game. Just because Maggs is an elite hitter does not make him any more clutch than other hitters. Maggs is always going to hit well.

[AngelAstro]

Originally posted by Monkeyman


My main problem with defensive statistics is its really hard to effectively and objectively quantify something as erratic and quirky as defense. That being said, when EVERY statistic consistently tells me Jeter is a poor SS, I'm willing to accept that he's not what I used to think. However, when I see a player who's RF and ZR vary from year to year greatly from one side of average to another, it makes it difficult to trust those numbers. I think offensive and pitching statistics are pretty much on target, though there are some inherent flaws in park-adjusted stats, which i discussed in another thread, but for the most part, they can pretty much be trusted. I'm on the fence about defensive stats, however, and probably will be for a long time.

I agree about defensive stats, they are far from perfect. I hope that those smart enough to come up with the stats will be able to make a better defensive stat.

[Monkeyman]

Originally posted by AngelAstro


You just described exactly why sabermetricians use correlations. A sabermetrician sees that Barry Zito's K-rate has dropped over the past three years. This puts up the red flag that this pitcher may be in decline. Thus, Zito shouldn't be signed to a 10 year, 200 million dollar contract, because there is a good chance (the correlation that declining K-rates of pitchers is associated with increasing ERA from stats of the thousands of pitchers before him) that he won't be winning 20 games a season 2 or 3 seasons from now. Now, this does not mean that Zito will decline, just that there is a chance. The next thing that the GM has to decide is whether or not Zito would be worth the years and money on his next contract. Of course, any smart GM would not only use stats, but would also use scouting to really determine if Zito would be worth signing to a large deal.

But my point is it would be more valuable to find the actual cause. If Zito declines, it won't be BECAUSE of declining K-rates. It will be caused by whatever causes his K-rates to decline. What would be more helpful is to determine what that is, and base your decision on that, or maybe even fix the problem, if there is one. Maybe there isn't, as he pitched very well last year, despite K-rates and W/L record. If he continues his decline in strikeouts but continues to pitch very effectively, it would be further evidence that declining K-rates don't CAUSE rising ERAs, that they are just a symptom of overall worse pitching in MANY (not all) instances. There's a correlation between pitchers with good W/L records and teams that make the postseason, but sabermetric guys don't use that stat to stock up on guys with good records, and rightly so. But if they dig a little deeper, they might be able to accomplish a little more.

[markp]

has been shown to be the forerunner of a decline in overall pitching. It's one of the most predictive stats for pitchers.

[Dave in MD]

Originally posted by Monkeyman


But my point is it would be more valuable to find the actual cause. If Zito declines, it won't be BECAUSE of declining K-rates. It will be caused by whatever causes his K-rates to decline. What would be more helpful is to determine what that is, and base your decision on that, or maybe even fix the problem, if there is one. Maybe there isn't, as he pitched very well last year, despite K-rates and W/L record. If he continues his decline in strikeouts but continues to pitch very effectively, it would be further evidence that declining K-rates don't CAUSE rising ERAs, that they are just a symptom of overall worse pitching in MANY (not all) instances. There's a correlation between pitchers with good W/L records and teams that make the postseason, but sabermetric guys don't use that stat to stock up on guys with good records, and rightly so. But if they dig a little deeper, they might be able to accomplish a little more.

what you say may be true, except that in almost all cases declining k rates mean a declining pitcher. There are exceptions to every rule. And Zito may bounce back, but k rates are extremely important to look at when evaluating pitchers.

[AngelAstro]

Originally posted by Monkeyman


But my point is it would be more valuable to find the actual cause. If Zito declines, it won't be BECAUSE of declining K-rates. It will be caused by whatever causes his K-rates to decline. What would be more helpful is to determine what that is, and base your decision on that, or maybe even fix the problem, if there is one. Maybe there isn't, as he pitched very well last year, despite K-rates and W/L record. If he continues his decline in strikeouts but continues to pitch very effectively, it would be further evidence that declining K-rates don't CAUSE rising ERAs, that they are just a symptom of overall worse pitching in MANY (not all) instances. There's a correlation between pitchers with good W/L records and teams that make the postseason, but sabermetric guys don't use that stat to stock up on guys with good records, and rightly so. But if they dig a little deeper, they might be able to accomplish a little more.

I can agree with you here, it would be nice to determine why Zito is having a declining K-rate. This would be the job of the GM of the team trying to sign Zito. Maybe he's lost 2-3 MPH on his fastball, so now it is easier to hit his change, maybe his curveball doesn't have the same snap as 2 years ago, maybe he is pitching from a different arm angle, etc. Of course, stats aren't going to pick any of this up. However, stats will tell you to look to find out if there is something wrong. If a GM were just to depend on W-L record, he would never even know to look to see if there was some way to improve Zito's mechanics so that he won't decline. I don't think stats are the end-all-be-all, but I do think that they should be used to try to find trends that wouldn't be seen with the naked eye.

And of course, the trends aren't always right, there are pitchers, Moyer for example, who succeed even with low K-rates. So, obviously, statistics don't tell the whole story. But, more often than not, the correlations will be correct and a team could save itself some money and hastle by just looking at the numbers.

[Monkeyman]

Originally posted by Dave in MD


what you say may be true, except that in almost all cases declining k rates mean a declining pitcher. There are exceptions to every rule. And Zito may bounce back, but k rates are extremely important to look at when evaluating pitchers.

But he doesn't NEED to bounce back! By all accounts, he's a better pitcher than he was two years ago, when his k-rates were peaking. He's NOT being hit harder. His slugging average against was the best of his career, as was his OPS against. His OPS+ was worse than onle 2002, which is beginning to look out of context. Maybe he's a low to mid 3's ERA guy, who goes 230+ innings with a WHIP under 1.2 which is pretty damn good. The K rates are irrelevant. An out is an out. If he is effective despite lower K-rates, then he's effective period, and his K-rates are irrelevant.

[Dave in MD]

Originally posted by Monkeyman


But he doesn't NEED to bounce back! By all accounts, he's a better pitcher than he was two years ago, when his k-rates were peaking. He's NOT being hit harder. His slugging average against was the best of his career, as was his OPS against. His OPS+ was worse than onle 2002, which is beginning to look out of context. Maybe he's a low to mid 3's ERA guy, who goes 230+ innings with a WHIP under 1.2 which is pretty darn good. The K rates are irrelevant. An out is an out. If he is effective despite lower K-rates, then he's effective period, and his K-rates are irrelevant.

Or maybe he got lucky last year. Such stats usually don't hold up with a high K rate.

[Monkeyman]

Originally posted by Dave in MD


Or maybe he got lucky last year. Such stats usually don't hold up with a high K rate.

Citing luck in this situation is ridiculousness. A proven pitcher who continues to put up great numbers in every RELEVANT category is not just luck. Such stats don't usually, but Zito isn't your average pitcher.

[Dave in MD]

Originally posted by Monkeyman


Citing luck in this situation is ridiculousness. A proven pitcher who continues to put up great numbers in every RELEVANT category is not just luck. Such stats don't usually, but Zito isn't your average pitcher.

no its not. There's a lot of reseach that shows that pitchers really only have total control over strikeouts, walks and home runs. Everything else is often luck. There are exceptions. But I would bet that if Zito's k rate continues to decline, so will his other numbers.