THE BOOK--Playing The Percentages In Baseball

Any writer doing so would remove one of the barriers between them and actual accountability for their choice.

So no, none of them will.

What is the difference between “best” and “most valuable”?
How can someone be the best at his job and yet be less valuable than someone else doing the same job?

Someone can argue that I may be the best apple-picker in the world, but if I do all my apple-picking on one acre farm, I’m not going to generate much value compared to my competition that has access to 100 acres.

Value = talent X opportunities to leverage

So the question, really, is whether the Cy Young Award is treated like the MVP award (you’re not ‘valuable’ if you can’t carry your team to the playoffs) or the RotY award (simply: who’s the best?), right? Because I’ve never heard it argued that the record of the RotY’s team should be taken into consideration.

How did the RotY avoid falling into the same trap, anyway? Is it because rookies are more likely to get a chance with non-playoff teams, and so a disproportionate number of good rookies play on bad teams?

I have always operated on the basis that the Cy was for “best pitcher” or “pitcher who performed the best”.  Never have I seen it likened to the MVP. 

But, that’s what this thread is about, to find out exactly what the Cy voters are being asked to consider.

And, no one has stepped forward with that information.

I’ll ask for the Coke secret formula next… I’ll have better luck.

Voters hopefully fairly adjust for height of apple trees and skill level of those helping to catch the apples in the buckets for you. 

#3—Instead of “Talent,” would it better to call it “Performance”? 

And then, how do you quantify “opportunities to leverage”?

#7, what Tom is calling talent is really some combination of talent and performance based on the voters preferences. Presumably most voters have slightly differing preferences. It doesn’t really matter what we call it.

So the difference between best and most valuable is innings and leverage? Isn’t how many innings a pitcher can pitch is a skill and isn’t leverage similar for starting pitchers?

Leveraging your context to impact wins.  Some people will hold Mariners low offense against Felix.  Some people will hold Mariners low win total against Felix.  Those are all opinions of what goes into their definition of “value”.  Which is fine.

Someone asked me the difference between “best” and “most valuable”, and that’s what it is.  Best is best, and most valuable is a combination of “best” plus anything else that gives that performance more, or less, value than it would otherwise get in a neutral environment.

Tango/5 wrote: “Never have I seen it likened to the MVP.”

Isn’t that implicit, though, when voters give significant weight to Wins? They may not put it that plainly, but it’s an inescapable conclusion.

I think wins have been looked at primarily as a measure of individual performance, Neil. More a case of the voters being willfully ignorant of the team aspects of wins come voting time than something you could draw an implicit conclusion like that from.

Most of the (traditional/old school) voters think that pitchers have, to some extent at least, an ability to pitch to the score.  Even if they don’t explicitly think that, they somehow think that whether a pitcher wins or loses has something to do with his ability as a pitcher even if they can’t articulate (in their own minds or if someone were to ask them) why.

Certainly if two pitchers had the exact same stats in every category but wins, most of the voters would choose the one with more wins.  So surely no matter what the difference between the stats of two different pitchers are, most voters take wins into consideration independently of those other stats.  Some voters put more weight on wins than others.  I don’t think you can fault any voter who puts some weight on wins, independent of any other stat(s).  You CAN fault those who put TOO much emphasis on wins, especially for those pitchers who pitch for teams who have a bad offense, like Seattle, since it should be obvious to anyone with half a baseball brain that no matter how well they pitch, it is going to be tough to get wins.

That being said, to me, all retrospective awards (aren’t they all?) boil down to some aspect of actual performance.  I don’t EVER want to hear stats like FIP or worse, xFIP, in the discussion of an award like the CYA.  Ever.  If a pitcher allows a bunch of base hits, whether it be luck or not, he did in fact allow them.  Same thing with home runs of course (hence why I don’t want to hear about xFIP).

Now, if you want to adjust for defense, that’s fine.  But FIP or DIPS ERA does LITTLE in that regard.  If you wanted to adjust for defense, you would have to use something like PZR or simply adjust a pitcher’s stats for actual team UZR (or DE) or projected team UZR.

And of course you HAVE to adjust for park, since without it, you are NOT looking at how a pitcher actually pitches - you are looking at how he pitches plus the effect of his park.  If you didn’t do that, then lots of crappy pitchers would be winning the CYA if they happened to pitch in a fictitious park with a true PF of say, 40 (where 100 is league average).  Clearly that makes no sense.  And if a pitcher pitched great, but he played in a home park with a PF of 200, it would be almost impossible for him to accumulate good stats other than wins.

So really what are the only things that count as far as pitcher performance is concerned?  RA per game, IP, and that is about it.  Adjusted for park, as I said.  If you want to give some credit or demerits for w/l, without getting carried away, I have no problem with that.  If you want to give a LITTLE more credit for K than for batted ball outs, I also have no problem with that, again, as long as you don’t get carried away (a pitcher who gets a batted ball out pitches just about as “well” as one who gets a K, on the average, and as far as RA are concerned).

So, I really don’t see a whole lot of controversy over who pitched the “best” for the season.  And again, please don’t mention any “ability” or projective stats like FIP, xFIP, or DIPS ERA in the discussion.  That makes no sense.  The CYA award is NOT which pitcher has the most talent going forward, regardless of how they actually pitched.  That would be absurd.

So, if the CYA is the pitcher who pitched best, give it to the guy who had the best park adjusted RA (ERA is fine as a proxy) and IP (some reasonable combination, like WAR or WAA), with perhaps a small adjustment for wins and then for K’s.  That’s it.  Period.  End of story.

MGL/13, It seems like you’re making an arbitrary distinction in attempting to eliminate factors beyond a pitcher’s control: the difference between an out and a hit could have nothing at all to do with the “actual performance” of the pitcher. Imagine universe A and universe B. Universe A and universe B are exactly alike in every particular, except that in Universe B a butterfly flapping its wings pushes a baseball just out of reach of a right fielder and three runs score. What you mean is closer to “actual results,” not “actual performance,” and is the same sort of argument that proponents of wins and losses use.

One could just as easily claim that “of course you HAVE to adjust for luck, since without it, you are NOT looking at how a pitcher actually pitches - you are looking at how he pitches plus the effect of his luck. If you didn’t do that, then lots of crappy pitchers would be winning the CYA if they happened to pitch BABIP of say, .100.”

The real issue is that, in any given sample, it’s not at all clear to what extent things like BABIP and HR/FB% are a result of luck and to what extent they are a result of “actual performance.” How much performance is it OK to ignore in order to mitigate the effect of luck? Maybe your answer is “none"--which is fine--but I don’t think it’s unreasonable to reach some other conclusion.

#14, if you were to watch every game that every pitcher pitches, then sure, you might say or think, “So-and-so pitched a really great game, but got unlucky in allowing a few hits just out the reach of a few gloves, and consequently gave up 6 runs and lost the game.”

But, that is NOT the way these awards are decided.  The voters look at a pitcher’s stats.  They are NOT going to say, “Gee, pitcher A has a 4.30 ERA, but his BABIP is .389. He must have allowed a lot of flukey hits.  I think I’ll vote for him for the CYA because he probably pitched well and just got unlucky.”

That would be preposterous.

So, I guess I am trying to say that I am defining “performance” strictly as the results and not whether or not it appears that the pitcher got lucky or unlucky in the game.  And I think that is entirely appropriate, unless you want to watch every pitch of every game for every possible CYA candidate and note how lucky or unlucky they got in each and every game.  And if you did that, the fans and the rest of the media would not stand for that since they are either not watching every play of every game, or, if they are, they don’t care whether a pitcher got lucky or not - they only care about whether he gave up runs/hits and got the win or not.

I am afraid that no matter how you slice it, the only reasonable way to determine who wins an award like the CYA is how “well” he pitched for the season (and the amount of that pitching, in terms of IP), where “well” simply means how many runs he allows, whether by hook or by crook.

Now, as I said, if you want to make some small adjustments for K and for defense, in order to account somewhat for whether a pitcher pitched “lucky” or not, and to more closely reflect how he actually “pitched” independent of the results, that is fine by me.

MGL/15: great post.

And, what a great idea:

unless you want to watch every pitch of every game for every possible CYA candidate and note how lucky or unlucky they got in each and every game

We all saw all of Strasburg’s starts, and I’d bet we’d be able to come up with scoring guidelines to determine how well he pitched to each batter, results be damned.

And then we can apply those scoring guidelines to all the top pitchers (Felix, Weaver, Liriano, Lee, Price, CC, etc).  It would be fun, not so that we come to a conclusion as to which pitcher was indeed the most outstanding, but because we would be watching baseball with an intent to learn something.

Do we really need 10 million baseball fans looking at seasonal totals of IP, R, ER, K, BB, HR, H, W/L, and having them tell us the proper weighting for those aggregated numbers?  That’s a monumental waste of efforts, and it’s hardly any fun.

But, dish out scoring guidelines to those 10 million fans, have them each watch a couple of real baseball games, now that would be FUN.

#14, if you were to watch every game that every pitcher pitches, then sure, you might say or think, “So-and-so pitched a really great game, but got unlucky in allowing a few hits just out the reach of a few gloves, and consequently gave up 6 runs and lost the game.”

So if I read that correctly, are you saying it would be correct to do that in theory but too difficult to do that in reality?

Who is the most outstanding player in a poker tournament?  Is it always the winner, by definition?  Even though he got all-in as 4-1 underdog all those times, those river cards DID hit him.  We might not want to back him going forward though because our projective stats of him don’t look good (he made many -EV plays). Personally, I would not advocate this position, but I get the impression some would.

mickey: MGL is proposing exactly that. 

What we care about is what our player does against some baseline, giving each player the exact same context.

If Franklin Gutierrez is there to snag a line drive off Longoria’s bat, we want to credit Longoria because if there was some other CF out there, it might be a hit.

mickey, I think it is reasonable either way. 

Sure, if you could watch every game and say, “So-and-so pitched really well, but got unlucky because of x, y, and z, so even though he allowed 3 runs, he really pitched better than that..” then sure, you can give him an award for “most outstanding pitcher” over and above the guy who allowed zero runs, but got really “lucky” (and maybe it wasn’t luck, but who really knows?), by allowing a bunch of hard hit “at’em” balls, stranding runner, etc.

Again, you would probably not be able to sell that to all the fans and media who either did not watch the game, or watched it but don’t subscribe to that theory of “outstanding.” So, in many ways it is not practical or practicable.

On the other hand, as I said, some people don’t care if a pitcher got lucky or not - they only care about how many runs he allows, not “how” he allows them, even if they are watching the game.  I think that is a defensible position as well.

In the end though, people want to see the doling of the award being commensurate with traditional pitching stats, mostly ERA and wins.  That is the bottom line.

But, I’m really not advocating anything one way or another - I am just illuminating the issue a little, at least according to the way I see it.

I think that the most important point I was making was about stats like FIP, xFIP, and DIPS ERA.  As I said (emphatically), I don’t EVER want those to be uttered (other to eschew them of course) in a discussion about an award (which many analysts do), and I gave my reasons why…

Doesn’t it depend on your definition of most outstanding or performance? If someone’s definition is how well the pitcher pitched, without luck, defense or offensive support, what would you use to determine that?

Another interesting stat:
Felix Hernandez- 60 ER, 77 R
Francisco Liriano- 62 ER, 62 R

Definitely I would look at R, not ER.  Felix has given up 15 more R while facing 194 more batters.  He’s a bit ahead on that count.