THE BOOK--Playing The Percentages In Baseball

My fantasy league uses OPS instead of BA. We have some VERY stubborn traditionalists, but even they agree it’s a better stat, at least for fantasy baseball purposes. The commissioner is pushing for replacing Wins with Quality Starts next year.

It seems to me that the batting average issue is a little more fundamental.  At it’s core, baseball is a game of a batter trying to hit a ball as well as he can so that he may run around the basepath and score a run before being put out.  And it’s a game played for fun. Everything else flows from that. 

Batting average, it follows, is an attempt to boil down that fundamental confrontation of pitcher and batter to a single, easily digestible number.  Player A is better than player B at getting hits.

Now, as the game has developed around that basic core, as management has looked for competitive advantage, and as fans have looked for the next frontier of enjoyment, we’ve grown more nuanced in describing the ways a player can impact his team’s run scoring.  In accurately assessing that performance, we know now we need to include walks and bases acquired to get the full picture.  So we use OBP, we use SLG and OBP.  We use Eqa and VORP.

However, a kid doesn’t pick up a stick and a ball and think about how fun it would be to jog to 1B after watching 4 pitches outside of the strike zone.  The joy comes from the crack of the bat and the batter-turned-runner spring towards first looking to make a turn towards second.  Batting average captures that in a way OBP, SLG, and OPS simply don’t.

Unlike measurement systems, baseball isn’t so utilitarian.  It exists for the joy it produces. 
I certainly understand the need to use better measures whenever we’re actually trying to measure performance in a tangible sense.  But performance and enjoyment aren’t perfectly correlated and batting average lives in that inefficiency.  Until you can replace that visceral, emotional connection between stat and enjoyment, the congregation at the church of batting average will be strong.

I advocate OBP instead of AVG and SLG instead of HR in all my 5x5 leagues—it never flies.  Fantasy baseball isn’t about putting together the best team of the most productive players.  It’s about the fusion of the most enjoyable moments of the game with the fun and challenge of roster management.

All of that said, you might be right.  While it may never be more fun to watch a walk than a high pop fly (even if it is more productive), it’s more likely that fantasy baseball players will adopt OBP, as their bean-counter instincts compel them to the more meaningful statistic.  It would be very interesting to see a wide-spread movement, though I fear that, like the sport, the supporting infrastructure may be too well developed to allow for it.

There are two separate issues here—the first is whether OBP is a more meaningful stat than BA.  Of course it is.

The second is whether the scale that BA produces is more familiar to the vast majority of fans than the scale than OBP does.  Again, the answer is, “of course.” To most fans (even, I submit, readers of this blog), a .300 average “feels” like a “good” result. 

So, the suggestion is simply to scale wOBA to BA rather than OBP.  The decision to adjust the raw wOBA numbers by multiplying by 1.15 is an arbitrary one—why not adjust the numbers so they resemble the more familiar scale? 

There’s no “reason” why a .300 average is good, a .350 average is “great”, a .255 is “average”, etc.—but it’s come to be common usage.  Leveraging that would increase the likelihood that wOBA would be understood at an intuitive level, which would be a good thing.

Just my 2 cents, but I think wOBA is better on its own intrinsic scale (avg = .300 or so, IIRC). That’s because the stat is pretty fundamental. (Of course, then you wouldn’t call it wOBA.)

There are other stats where this is not true. For example, 20 years ago I made up a stat called the Pitching Index, which is (K-BB)/HR. (It’s the forerunner of DIPS and FIP ) The PI, even today, is a good indicator of pitching ability, but it doesn’t have any intrinsic meaning. Therefore, converting to an ERA scale would not be uncalled for, if I wanted to bother.

To me, wOBA is not the type of stat which should the victim of needless manipulation.

I agree with David that wOBA does not need a fudge (although I understand why you did it for the Book, using it in binomial style).

However, I disagree with Rick/2 and Craig/3 that the BA scale is better than the OBA scale.  For one thing, it would make it seem like a tawdry knockoff of EqA.  For another, as I think you guys mentioned in the Book, it allows you to compare to a valuable stat (OBA), and get some sense for how good the player is in the areas of offense not considered by OBA.  If you have a guy with a better OBA than wOBA, then you know he was weaker in power.

With a BA scale, you can make a similar comparison, but it is a little more broad as divergent EqA and BA indicates something about secondary skills.

Actually, wOBA gives you two very important characteristics:

1. When wOBA = OBP, then we know that the SLG is proportionate to OBP the way the league’s SLG is proportionate to the league OBP is.  wOBA starts with the idea of safe and out.  The denominator is safe+out, or PA, the same as OBP.  The numerator is also centered around safe=1.  A walk and single are a bit less than 1, and the extra base hits are alot more than 1.  So, it has that centering aspect whereby if you have a typical profile of hits, walks, and HR, then when wOBA = OBP, the SLG will follow in the expected proportion.

2. By making it at the same level of OBP, you can now use the binomial properties that we are used to (slightly adjusted).  OBP is a binomial that uses PA as its opportunities and so does wOBA.

I understand the appeal to scale wOBA, and you can have a wBA and wSLG if you like by multiplying by .90 and 1.40 (or so) respectively, instead of 1.15. 

However, I don’t see any need for mainstream acceptance of wOBA.  EqA is a pretty good stat, and the BP machine is 100x bigger than whatever I’m involved in, and EqA is not really that prevalent.  So, even if mainstream acceptance were a goal, I think it’d be way hard to achieve, if not impossible.

I like wOBA for the niche it’s set out for itself.

Patriot/5 has a better post and more succinct than what I posted in Tango/6.

I think in fantasy baseball there is much concern about spreading the fantasy value to all different kinds of players, thus making managing a team more interesting.  Of course stats like OPS and OBP are better than AVE, but OPS tends to hike the value of sluggers, as home runs are now double counted.  OBP isn’t as bad as OPS in this regard, but many of the big sluggers tend to draw a lot of walks (have a high OBP compared to their BA).  Same with pitching, Saves and ERA are not relevant stats when it comes to “real” baseball, but they give fantasy value to a whole other set of players.  If you are concerned about sabermetrics in fantasy baseball, then you should set up a points league and weight the stats in accordance with linear weights or RC.  I like the fantasy value equilibrium that the standard 5x5 stats provide.  I wouldn’t mind switching from BA to OBP, but also don’t mind the way it currently is.
vr, Xeifrank

If you really want to mimic OBA, then why not constrict wOBA to between 0 and 1?  Would that ruin the binomial thing or anything else? You could do it by the following :

(unadj. wOBA/1.7) ^.63

"In 30 years, BA may not even exist, while OBP will always exist. Eventually, OBP will supplant BA because the old stupid generation will die and the young people who favor OBP will thrive. In any case, BA is a horrible stat, and I’m never going to perpetuate its use.”

Can you please state whose quote that is (I can guess) so I am attributing it to the correct person?

The above comment is one of the “stupidest” I have ever read by someone who is considered intelligent. - “Everyone is stupid, just in different things”. (Will Rogers)

No one is going to make Ted Williams’s .406 or Ty Cobb’s .367 not exist.

BA has a correlation to runs scored of .70 to .75 - that is not “horrible”, many studies done on sites like these have lower correlations.
There are obviously better stats, and people have known that longer than you have been alive, and some people have known it before your father was born.

I know/believe the metric system is better than English, but I use the English system.
I know/believe Lotus 123 is better than Excel, but I use Excel. Some people believed Beta was better than VHS, 8 Track was better than Casette.

How Microsoft got everyone to use Excel instead of Lotus is an incredible example of how to get people to change from one way of doing things to another.

How the US did it with the metric system is the example of what not to do.

Calling people an “old stupid generation” is beyond belief.

Given the single choice between OBP or BA, which do you choose?

If all you knew at the beginning was OBP, and then BA came along in 1980, would you then switch from OBP to BA?

By the way, I shouldn’t have said stupid.  I should have said ignorant.

So, I agree.  It was stupid of me to call people stupid, when they were ignorant of anything better than BA, were brought up as a kid on BA, and when they may have finally learned of OBP, they were too old to do anything about changing their spots.

In order to get change on this matter, those who are willfully ignorant of OBP and still hold BA to a higher regard, and continue to perpetuate the use of BA through their media power need to go away.

Then, we can let the fairer fight go on, and OBP will supplant BA.

I agree that BA’s cache is the cool sounding name “batting average”, as opposed to “on base” something.  One looks more passive than the other.

#11 - obviously i would choose obp over ba -

I have been doing that prior to 1980.
I was in a number of board baseball leagues (Strat-O-Matic), and would always do real well because I chose players based on obp. My teams didnt do as well as I thought, because I underestimated (Ignorant?) the importance of slugging.

Isnt obp + slg becoming much more accepted by the general public? And isnt that better than obp?
Should you try to convert people to obp + slg?

Those who perpetuate ba would be espn and usa today. Print newspapers are dying. If I google 2008 baseball statistics, they come up first, and the 1st stat that give is ba. HR & RBI are next. ESPN does not have obp on its top page of stats.

No, I don’t think that OPS is better than OBA.  If you could only look at one number for any given hitter, then yes, I would rather have OPS.

But that’s an artificial constraint.  Since we are free to consider more than one piece of information, OBA is more valuable than OPS.  OBA is a fundamental measure that you are going to want to know regardless of how good your overall measure is.  OPS is redundant and thus utterly useless if you can use RC/G or wOBA or ...  Why perpetuate its existence, let alone its scale?

I agree.  OPS is nothing real.  OPS is probably as good as TA.  Let’s look:

OPS TA event
0.45 0.39 1b
0.83 0.78 2b
1.22 1.17 3b
1.60 1.56 hr
0.23 0.39 bb
(0.28) (0.28) out

TA undervalues singles by 6 more points than OPS.

TA values doubles correctly, and OPS is 5 points over. 

HR are both overvalued, but OPS over-overvalues them compared to TA by 4 points.  Same for triples.

And as much as OPS undervalues walks, TA overvalues walks by around the same amount.

In short, OPS overvalues extrabase hits by 4 or 5 points, while TA undervalues singles by 6 points, and the walks are a wash.

Overall, OPS and TA are around the same in terms of telling you about run creation by players.

TA at least tells you something real (batting bases per out), as opposed to OPS which is just adding whatever rates it finds.

Whoever is critical of TA (and you should be) should be equally critical of OPS.

***

dq: shocked about the OBP thing about ESPN.  Those guys put out such new-age stats, I can’t believe they wouldn’t put OBP right up there.

Also, I actually take as a compliment that you think what I originally said earlier was the stupidest thing you may have heard from me.  I was quite sure I have said much stupider things, but if that’s as stupid as you think I get, I’m doing ok.

#17 but dont you get

OPS OBP
.45 .38 1b
.83 .38 2b
1.22 .38 3b
1.60 .38 hr
.38 .38 bb

based on r/ (h +bb) for obp, I may have done it wrong, but based on that logic, isnt ops better than obp? Doesnt ops correlate to runs bette than obp?

The intent of OBP is pretty clear isn’t it?  The comparison level is OBP to BA, not OBP to OPS.  Of course OPS is better to know overall, but that’s only because it uses OBP plus extra information.  But it adds that extra information in a clumsy way.

BA is a bit clumsy.  OBP is not clumsy and it has extra information.

Aaron B. from Novato, CA (see link to Neyer chat) is a big fan:

Aaron B. (Novato, Calif.): “OBP plus Slugging % (OPS) is the most accurate indicator of how well a hitter has been hitting.” Please tell Ryan how wrong he is (OPS undervalues OBA, etc.)Weighted On-Base Average (wOBA) is the ONLY way to accurately combine on-base average and slugging.

Rob Neyer: Well, there are other ways. But yeah, wOBA is solid. We’re just not quite there yet, popularity-wise.

Thanks for playing today. I’m still not sure who the MVP is, but I know a lot more now than I did an hour ago.