Fan vs. PECOTA Projections

[RShack]

I don't really think there are many LH fans out there, SG. I just don't think there are as many people who object to him as strongly as you do. The way you talk about him, you'd think he'd offended your mother or something

IMO, this team isn't going to do anything competitive in 2008 so I don't really care if LH is out there or not... I hate to chalk up 2008 as a failed year but I Can't see it as anything but that right now. We've stomached a decade of pathetic baseball already, I'm sure we'll be able to stomach Luis Hernandez for a season.

What some people seem to be missing about this whole issue is that there *are* interesting things about it. I don't know why people have to jump to arbitrary black-and-white answers about it. The real issue isn't LH, the real issue is "What kind of value does a SS have?" It gives us is a chance to stop and pay attention to what "SS value" really means, and what different profiles of SS-value might be valueable. The real issue is "SS value", and that kind of topic is a very interesting topic to some of us who like to watch baseball. And here we are, facing a possible transition from a great-hitting-SS to a great-fielding-no-hit SS, which gives us the perfect chance to ponder all the implications. "What if he hit X and played Y-quality D? What would that really mean?" That's a "what if" discussion to learn stuff about what SS means. This is a topic about which trades, and the relative improvement different trades might bring, are a secondary matter.

I just wish that having a "what if" discussion about SS-value (or other things besides trades) was better tolerated and nurtured around here. It's a way to learn stuff about baseball. What's wrong with that? Do we have to reduce every single thing to a discussion about trades? About 85% of the traffic around here is having "what if" discussions about all kinds of trades that won't happen. It's certainly fine to play "what if" about 3-way and 4-way trades schemes if you want to do that. But most of the trades that get talked about here have as much chance of happening as LH has of winning the dang Triple Crown, so why isn't it OK to discuss "what if" scenarios about LH as a SS? Why is it so bad to look at the different sides of SS-value? Can't it be fun to learn about how a great glove might have some significant Actual Value, and how to relate that to other things?

I don't see why we're not having fun trying to figure out how well he'd have to field to justify a .578 OPS. What would it take for him to justify that? This should be an opportunity for thinking these things through and trying to learn something about baseball. I don't think there's anything "pathetic" about playing "what if" games to learn about baseball. I think it's more interesting than discussing this week's 136th trade idea (which is just a re-hash of last week's 87th trade idea). So, the real question isn't whether or not LH is "pathetic". The real question is: If it turns out that LH hits more-or-less what Bill James thinks he's gonna, then how do we figure out what kind of D-year he needs to have to justify it? And how would such a D-year compare to what other notable good-glove SS's have done. We need some way to get a more-or-less fair handle on that before we even know how to say what the likely net cost of playing him is. So, how do we figure that out? I don't know how, but I hope we can somehow ID the best cheap way to do it. If people wanna talk about how to do that, I wonder if it should go back in the other thread where we were trying to talk about this kinda stuff. That way, people who wanna complain about LH can stay here, and the discussion about what different aspects of SS-value mean could be over in that place, which is: http://forum.orioleshangout.com/forums/showthread.php?t=56907&page=14

[Frobby]

What some people seem to be missing about this whole issue is that there *are* interesting things about it. I don't know why people have to jump to arbitrary black-and-white answers about it. The real issue isn't LH, the real issue is "What kind of value does a SS have?" It gives us is a chance to stop and pay attention to what "SS value" really means, and what different profiles of SS-value might be valueable. The real issue is "SS value", and that kind of topic is a very interesting topic to some of us who like to watch baseball. And here we are, facing a possible transition from a great-hitting-SS to a great-fielding-no-hit SS, which gives us the perfect chance to ponder all the implications. "What if he hit X and played Y-quality D? What would that really mean?" That's a "what if" discussion to learn stuff about what SS means. This is a topic about which trades, and the relative improvement different trades might bring, are a secondary matter.

I think the people who are impatient with these discussions are people who have read some of the sabermetric discussions of these issues and who think that those analyses have answered these questions a long time ago.

Now, I don't put myself in that category. However, if I did, it would be pretty simple for me to look at two stats that are publicly available that frame offense and defense into one combined stat: Bill James' win shares and BP's WARP3. And those would provide basic answers to these questions if one believed those stats.

Let's take Win Shares. In 2006, Adam Everett had the most defensive win shares of any SS (8.6), and yet he ranked 16th in overall win shares, with 12. There were 11 shortstops who earned more than 12 win shares just with their bat. Derek Jeter ranked 18th in defensive win shares but still had the most win shares of any SS, with 33. 3 win shares = 1 win, so according to this system, replacing Jeter with Everett in 2006 would cost a team 7 wins. That's about as extreme an example as one could find.

Now look at WARP3. Jeter had a WARP3 of 11.7, Everett was 5.7, so the difference is 6 wins.

So there you have it: replacing a great hitting, poor fielding SS with a great fielding, poor hitting SS would cost 6-7 wins. Make the players a little less extreme and the difference will be less -- if you believe in these stats.

I think it is sabermetric gospel that good hit, no field is better than good field, no hit pretty much every time. So if you believe that gospel, there isn't much to discuss. And if you don't believe it, all you can really do is argue your gut feeling, because there aren't statistical analyses that back you up.

With that said, defensive stats are getting better all the time, and I think Tom Tango's new method (described in my thread about The Hardball Times Annual) may be the best yet.

[RShack]

I think the people who are impatient with these discussions are people who have read some of the sabermetric discussions of these issues and who think that those analyses have answered these questions a long time ago.

Now, I don't put myself in that category. However, if I did, it would be pretty simple for me to look at two stats that are publicly available that frame offense and defense into one combined stat: Bill James' win shares and BP's WARP3. And those would provide basic answers to these questions if one believed those stats.

Let's take Win Shares. In 2006, Adam Everett had the most defensive win shares of any SS (8.6), and yet he ranked 16th in overall win shares, with 12. There were 11 shortstops who earned more than 12 win shares just with their bat. Derek Jeter ranked 18th in defensive win shares but still had the most win shares of any SS, with 33. 3 win shares = 1 win, so according to this system, replacing Jeter with Everett in 2006 would cost a team 7 wins. That's about as extreme an example as one could find.

Now look at WARP3. Jeter had a WARP3 of 11.7, Everett was 5.7, so the difference is 6 wins.

So there you have it: replacing a great hitting, poor fielding SS with a great fielding, poor hitting SS would cost 6-7 wins. Make the players a little less extreme and the difference will be less -- if you believe in these stats.

I think it is sabermetric gospel that good hit, no field is better than good field, no hit pretty much every time. So if you believe that gospel, there isn't much to discuss. And if you don't believe it, all you can really do is argue your gut feeling, because there aren't statistical analyses that back you up.

With that said, defensive stats are getting better all the time, and I think Tom Tango's new method (described in my thread about The Hardball Times Annual) may be the best yet.

I guess I was more focused on trying to flesh out what good-D SS really means, and that's not at all the same thing as just looking up somebody else's number when we're not even quite sure exactly what that number is really telling us. The problem is that neither Win Shares nor WARP3 tells me anything meaningful about the ways in which SS D-contributions differ. I have no idea what WARP3 actually does. Some years ago (during a time when I was working with stats a bunch in my non-baseball profession, so I was more in tune with them), I spent some time reading up on the ins and outs of Win Shares, and it provided nothing in the way of enlightenment. I do not wish to diss Win Shares. I have no particular opinion about its strengths and weaknesses about how it does its main job, which is to divvy up credit for a team's performance vs. the league. I am perfectly prepared to believe that it works as an insightful accounting system that takes the big-picture result of a team's season and allots little bits of credit for it here and there. I certainly don't wish to fight with anybody about it. My only point is that it does nothing to illuminate anything whatsoever about the actual differential value of a great-glove SS from a crappy-glove SS. If anybody has found a way to study Win Shares and become enlightened about defense, I would love to hear their insights. Because I sure tried, I really did, and I couldn't actually find anything to learn about that. Nothing at all.

At least simplistic bottom-up guestimation formulas about hitting make some sense. If you say "AB * OBP * SLG = Runs", at least you know what your doing. And you can have discussions about whether it makes sense to weigh OBP and SLG equally or not (I kinda think it doesn't, but I don't really know). But with the way that Win Shares deal with actual-D performance, all kinds of things get blurred-up together, to the point where you don't even know whose attributes are really being counted. It's definitely a top-down system (which is fine), but once it gets to a certain level, it just starts blurring lots of things together (which I think is not at all fine). I was kinda hoping there would be some insightful bottom-up way to look at it (sorta like the simple RC formula, above) based on things like range and arm, that would provide some Useful Perspective to back up whatever value gets calculated. That's the kind of thing that 1970 used in the other thread, but that doesn't help us project MiL guys because evidently we don't have the data. Win Shares may well have good value for certain tasks, but when it comes to providing a good study of D-impact, it seems to involve a whole lot of vagueness and hand-waving and blurring things together. I'm not trying to poke holes in a sacred cow, it might be great for certain purposes, but I don't see how it illuminates anything at all about defensive contribution. All it does is give it a number through top-down means that ignore the aspects of good-D SS's that would seem to matter. So, because of that, it doesn't really help any.

As for what seems to be an evident belief that good-D is not a terribly important attribute, I agree that it seems to be the current claim. There may well be a very compelling justification for that, but Win Shares sure ain't it.

[Rob H]

As far as arguing defensive metrics goes, I pretty much stick to Dewan's +/- system.

[Gurgi]

Interesting arguements.

Bill Ripken played 12 years not being able to hit much. Was he worth putting out there? I think so. I think Hernandez might be able to do a decent job until either he gets stronger and hits or we need to upgrade SS because we are becoming competitive again.

I also belive the expectations on players to produce certain standards of offense is/was in large parts because of PED's. I think if we ever really get a handle on PED's offensive numbers will fall back close to mid eighties levels. Thus making Defense a little more important.

[RShack]

As far as arguing defensive metrics goes, I pretty much stick to Dewan's +/- system.

OK. I was just googling around about that, just to find out more about it... and I ran across something that does not make sense to me. The thing I was reading was this, and the specific thing that puzzled me was this part of what it said there:

One thing we want to do in the future is translate these plus/minus numbers and all the other defensive metrics we have in this book into one number. That might be a number similar to Runs Created, but for defense not offense. Maybe it’s called Runs Prevented. But between you and me, you can use the rule of thumb that Bill James used in his article on Derek Jeter and Adam Everett.

That is, use a number a little less than half of the plus/minus number as an estimate of runs prevented. Since the value of a single is a little less than half a run, you can use a “little less than half” of the plus/minus figure to estimate runs prevented. Adam Everett’s plus/minus figure of +33 could be estimated as preventing about 15 runs. Then using another rule of thumb that estimates the value of a win at 10 runs, Everett’s defense generates an extra 1½ wins for the Astros in 2005. ... I would suggest using half and rounding down...

For the most part, that makes sense to me. It's exactly what I want them to do. But the bolded-and-underlined part doesn't make sense to me. I understand that a single is worth a little less than half a run. That part is fine. But when the SS gets a guy out instead of letting him have a single, the SS is doing more than just stopping a single. He's also getting an out too. Isn't an out worth anything? If you get the first 27 batters to make outs, then you haven't given up any runs, and not giving up any runs is way better than average. So, when a SS prevents what should have been a single, isn't that worth more than just half-a-run? Shouldn't it be worth half-a-run plus whatever an out is worth? How much is an out worth? It's gotta be worth something, doesn't it?

[RShack]

Another question... about Win Shares. As I understand it, Win Shares says that a run-saved is worth more than a run-scored. The whole Win Shares calculation is predicated on that, to the tune of giving 8.33% more weight (.52/.48 = 1.08333) to runs-prevented, right?

So, if people think Bill James basic Win Shares idea is plausible, then shouldn't we be giving 8.33% extra weight to D-runs vs. O-runs?

[DrungoHazewood]

Interesting arguements.

Bill Ripken played 12 years not being able to hit much. Was he worth putting out there? I think so. I think Hernandez might be able to do a decent job until either he gets stronger and hits or we need to upgrade SS because we are becoming competitive again.

That argument hinges on who the Orioles (or his other teams) could have replaced him with. But he set the bar very low most years. He was a truly adequate player only one year - in 1990. Besides that he never got much above 2.0 wins over replacement. He was a really awful hitter. He played much of his career in the 1990s but has several seasons with an OPS around .500, and combined that with almost no baserunning pluses.

In modern baseball you have to be a spectacular fielder to make a .600 OPS acceptable. That's the whole point of this thread. Bill Ripken (admittedly in a slightly smaller run environment than the late 90s) didn't match that most years. He was only worth it if:

- His fielding was legendary. I'm not inclined to say it was.

or

- The O's only had more ridiculously bad options behind him. That may have been true, even if it's not really defensible.

I also belive the expectations on players to produce certain standards of offense is/was in large parts because of PED's. I think if we ever really get a handle on PED's offensive numbers will fall back close to mid eighties levels. Thus making Defense a little more important.

That's almost certainly not true. PEDs were one of many, many reasons for the offensive explosions in the 1990s. Even if PEDs are completely eliminated you'll still be left with bandbox parks, players who hit the weights for five hours a day, thin whip-handled bats, and a strike zone the size of a postage stamp.

[Frobby]

Another question... about Win Shares. As I understand it, Win Shares says that a run-saved is worth more than a run-scored. The whole Win Shares calculation is predicated on that, to the tune of giving 8.33% more weight (.52/.48 = 1.08333) to runs-prevented, right?

So, if people think Bill James basic Win Shares idea is plausible, then shouldn't we be giving 8.33% extra weight to D-runs vs. O-runs?

Have you read the Win Shares book? I did, but it's been 3-4 years from whenever the book came out and I forget nuances like this. In any event, the book is quite good, even if in the end you don't buy into the way he calculates Win Shares.

As to your question about the value of a single versus an out, I think the estimate that a single = .5 runs already takes that into account.

[Sports Guy]

I don't really think there are many LH fans out there, SG. I just don't think there are as many people who object to him as strongly as you do. The way you talk about him, you'd think he'd offended your mother or something

IMO, this team isn't going to do anything competitive in 2008 so I don't really care if LH is out there or not. Sure, I'd like us to deal Bedard and Roberts and net us a real SS prospect, but I don't know if Bedard will be traded at all (I don't think he will) and I don't know if a Roberts trade will give us that.

I hate to chalk up 2008 as a failed year but I Can't see it as anything but that right now. We've stomached a decade of pathetic baseball already, I'm sure we'll be able to stomach Luis Hernandez for a season.

IMO, he's not worth debating much, anyway.

The idea of going with LH is less about him and more about it just being a poor decision by the FO and laziness.

That is my bigger issue....LH himself is one problem...The FO portion of it is a larger problem.

[tangotiger]

I guess I was more focused on trying to flesh out what good-D SS really means, and that's not at all the same thing as just looking up somebody else's number when we're not even quite sure exactly what that number is really telling us.

In my With Or Without You method, we know what the number is telling us. I suggest that your thoughts on the other methods (Win Shares, BP, etc), however well-founded, don't apply to my method. Please read the Catchers and Jeter articles in the annual.

[tangotiger]

So, when a SS prevents what should have been a single, isn't that worth more than just half-a-run? Shouldn't it be worth half-a-run plus whatever an out is worth? How much is an out worth? It's gotta be worth something, doesn't it?

Right. It's around 0.75-0.80 runs. Anyone who tells you otherwise: don't listen.

[RShack]

Have you read the Win Shares book? I did, but it's been 3-4 years from whenever the book came out and I forget nuances like this. In any event, the book is quite good, even if in the end you don't buy into the way he calculates Win Shares.

I can't remember exactly what I did or didn't read back-when. But I know for sure that I jumped way into it, simply because I was excited about Bill James in general, and I was trying to see what the ballyhoo'd Win Shares told us about the value of good-D guys. I didn't just give it a glance, I read all kinds of stuff, and took the trouble to work out various scenarios. I was left with the conclusion that it was a top-down accounting method that got all blurry and hand-wavy when you tried to get to the bottom of a specific guy's D-contribution.

Once again, I'm not saying that Win Shares isn't useful, I'm just saying it's not useful for what we're talking about here. Just because a number might be good for some things, that doesn't mean it's good for everything. Trying to use Win Shares for this is like using a garden rake when what you need is a comb. The fact that you need a comb doesn't mean that garden rakes are a bad idea, it just means that they aren't useful for combing your hair, that's all.

In my With Or Without You method, we know what the number is telling us. I suggest that your thoughts on the other methods (Win Shares, BP, etc), however well-founded, don't apply to my method. Please read the Catchers and Jeter articles in the annual.

I wasn't trying to diss your method (or any of the alternate methods). I was just responding to the idea that Win Shares addresses this topic in a meaningful way, that's all. It doesn't. As for the various flavors of WARP, AFAIK whatever they do or don't do is stamped "Top Secret", so we just can't tell either way.

As for reading the article you just referred to, I look forward to doing that. I recently ordered the annual but, through no real fault of my own, I can't actually read it until it gets here ;-)

[El Gordo]

Right. It's around 0.75-0.80 runs. Anyone who tells you otherwise: don't listen.

That would mean that Everetts + 33 FB rating would translate into 2.3 wins and Jeters -33 would translate into -2.3 wins?

[RShack]

I think it is sabermetric gospel that good hit, no field is better than good field, no hit pretty much every time. So if you believe that gospel, there isn't much to discuss. And if you don't believe it, all you can really do is argue your gut feeling, because there aren't statistical analyses that back you up.

I understand what you're saying, but I disagree with your conclusion. The problem with your conclusion is that it's the same circular argument that gets used a lot around here. It's circular because it starts from the premise that says "stats = baseball" and then insists that any effort to refute it must be based on baseball stats, and that anything else is somehow wrong. Now, regardless of your opinion about the value of D, you gotta admit that that's circular reasoning.

Fortunately, if we back up just a step and look at stats, we can see a very compelling reason to challenge the current claims about the low value of baseball-D. All we need to do is look at what stats tell us about things over time, especially when it comes to measuring complex performance issues. If you do that, then I think it's inevitable that you reach the following conclusion: "Whenever people start trying to be scientific about measuring complex performance issues, it naturally happens that whatever they can measure easiest-and-first always gets way-overrated." Just to annoy some people, let's call this "Shack's Law" ;-)

Here are a couple examples of Shack's Law in practice:

• IQ Tests: When they first started, the mainly measured how good somebody was with numbers and words, simply because that's what was easy to measure. You could make up tests and get numbers back about those things, so that's what IQ measured. But then they realized that it's not that simple, so they started adding different things, like stuff about spacial relations and other fuzzy perceptual-problem-solving tasks. At which point the weight given to just manipulating numbers and words went down, which means they were overrated at first. More recently, as we see that computers can do number-stuff and some-word-stuff better than people, but are otherwise colossally stupid, there's been an explosion in research showing all the ways that intelligence goes way beyond just ability at numbers and words. At the moment, IQ tests have still not caught up to what science knows is wrong with them. The consistent theme is that the weight given to how people do at manipulating numbers and words gets reduced as other things get measurable. Which shows that the stuff they could measure at first was way-overrated.
• Same thing with GRE tests. It used to be that there were just 2 parts to them: math and verbal. Each one of those two components carried 50% of the total weight. Then they realized that grad school success also involved a big component of analytical thinking. So they added a 3rd part about analytical thinking. That's harder to do, so it took longer to come up with a test about it. But once they did, then it got added to the mix, such that math and verbal both got demoted from 50% to 33.3% of the total. Which means that they used to be overrated by fully 50% (the diff between 33.3 and 50).
• Same thing with how the better car magazines test car performance. The easiest thing to measure was 0-60 acceleration times, so that's what everybody did at first. But then people realized that handling was important too, so they added various means of trying to measure that. When they did that, the role of 0-60 acceleration got demoted to carry less weight than it originally did. At present, even those 2 very measurable things get overrated by stats. Everybody who's into serious driving knows that the real measure is not just how fast a car can go (either in a straight line or on a handling course) but also comes down to how stable and comfortable the car is while doing it. But they don't know how to measure that, except subjectively. So, here's what routinely happens: they do tests that measure what they can measure, but then adjust the numbers based on stuff they *know* is true but that they can't "objectively" measure. So, for example, a hot-rod Mustang might beat the BMW on the numbers, but the BMW wins the test anyway because it goes like stink while staying stable and comfortable, and the Mustang can't do that as well. (This causes angry letters-to-the-editor from huffed Mustang-owners who are mad 'cause the stats "prove" that the Mustang got unfairly dissed ;-) The distinguishing feature between the BMW and the Mustang is the complex hard-to-measure part. Which is what baseball-D is: it's the more-complex phenomena that's harder to measure.

You can even see the same phenomena at work in baseball stats themselves. Everybody used to go by BA, simply because it was easiest. Over time, people decided that OPS is better, and concluded that BA had been way overrated. It's the exact same principle at work. It's not that anybody was *trying* to overrate BA, just like nobody is *trying* to underrate baseball-D. Nobody is *trying* to be unfair about it. To the contrary, all the serious people are trying very hard to be fair about it. Rather, it's just another example of Shack's Law: it's a perfectly natural and inevitable consequence that what we know how to measure gets overrated compared to stuff we don't know how to measure properly yet. So, while it's nobody's fault, I think it is entirely fair to conclude that the "state of the art" about baseball stats virtually guarantees that good-D is being underrated (by accident). I'll bet you anything that 10 or 15 years from now, people will conclude that way-back in 2007 we were consistently (but accidentally) underrating the value of good-baseball-D. It happens every time. I don't know of any case where it *hasn't* happened (which is why I'm calling it a "law" rather than a "theorem" ;-)