Don’t give up on batting average just yet

If you ask an analyst to identify the statistical category that fantasy teams will find toughest to make up ground in roto leagues, chances are the reply will be “batting average.”

The premise behind such logic is typically based on the notion that average is a rate stat, and with a few thousand at-bats already accumulated, the opportunity to move AVG significantly becomes tougher as the season progresses. For example, a team that maintains a batting average of .272 at the half-way point and wants to get it up to .282, will need to accomplish a .292 average for the rest of the season.

That might seem daunting and we’re guessing that a lot of fantasy teams simply give up on chasing average thanks to the army of pundits who declare moving average upwards at this point of the season to be a Sisyphean task.

I have doubts about this logic. I think it’s quite foolish to assume that catching up in average is any bit more tough than making up ground in any other category. In some regards, I believe there’s benefit to chasing a high average despite what conventional wisdom might say.

The first problem with typical analysis on batting average is one of perception.

Making a .292 average the rest of the year might seem intimidating. But what if I told you that you only needed to get 30 more hits than your competitors? Would that change your mind?

After all, average is merely hits divided by at-bats. If we normalize the denominator by assuming that teams in a given league will achieve roughly the same amount of at-bats, all that’s left is hits. (There are some factors why teams won’t get the same number of at-bats, but the spread in a typical league isn’t that large.) If teams in your league each accomplish roughly 3000 at-bats from here until the end of the season, the difference between your competitors’ assumed .282 average and your desired .292 average over those remaining at-bats translates to 30 hits.

Is 30 hits more daunting than, say, a gap of 10 steals? I’ll leave that up to you do decide.

“But wait,” you say. “Doesn’t the fact that I’m stuck with a .272 average at this moment indicate that I don’t have the players to achieve a .292 average the rest of the way?”

Answer: It depends.

The second problem with conventional wisdom that states that moving AVG up at this point of year is a fool’s errand is that it ignores economics—specifically supply-and-demand curves concerning available player talent. Sure, moving your average up with your current roster might be tough, but how about all those players who might potentially help you in the free agent pool?

People who play fantasy baseball love batters who hit home runs and steal bases. Typically, they give less respect to high-average players who don’t contribute in the power and speed categories.

Look at every hitter in baseball this year with at least 15 HR so far. How many of them are owned in your league? All of them?

Look at every hitter in baseball this year with at least 15 SB so far. How many of them are owned in your league? All but one or two?

Now look at every hitter in baseball this year with at least a .300 BA in at least 200 plate appearances. Are players like Martin Prado, Nick Johnson, Skip Schumacher, Maicer Izturis, Cristian Guzman, Scott Podsednik, Alberto Callaspo owned in your league? Would they be that hard to attain in trade? Unless you play in a very deep league with a shallow player pool, I’m guessing there’s good supply and mediocre demand on a batter who makes good contact with the ball and can be expected to put up a high average.

(Bonus note: Alberto Callaspo has 31 more hits than Jay “Batting Average Killer” Bruce to date. Did someone say 30?)

A Hardball Times Update
Goodbye for now.

Often in fantasy leagues, we’re forced to make choices at this point of the season. Our teams might not be in position to dominate every category and finding a few extra points may be the difference between winning and coming in second place. We may choose to attack a certain category and give up on another category because that’s where we see the best opportunity for standings gain.

But be careful how decisions on punting one category can influence your team’s standing in the other categories.

The third and last problem with advice that tells teams that chasing AVG is a foolish endeavor at this point of the season is that it ignores the full ramifications and trade-offs of a team that elects to punt the category.

To drive this point home, I took the top 150 batters in the 2008 season. I wanted to determine the correlation between a batter’s success in an individual category and that batter’s overall value. The table below measures the degree of correlation on a scale of -1 to 1. The higher the number, the stronger relationship between a batter’s single category success and overall category success.

Category: Correlation with Overall Value
Runs: 0.79
RBIs: 0.66
Average: 0.54
Home Runs: 0.54
Steals: 0.26

As you see, average is roughly as important to a batter’s overall success as home runs, and certainly more important than stolen bases. Another way to look at this is to say that a fantasy team stands a better chance of giving up on steals without damaging their position in the other categories than to give up on average without hurting their team in categories such as runs and RBIs.

Ask most analysts to identify the statistical category that fantasy teams will find it easiest to make up ground in roto leagues, and many may answer, “steals,” because it’s a counting category that’s relatively scarce and having one good speed threat can make quite a difference. However, this advice ignores the fact that most batters who steal a lot of bases do little much else to help out.

The same can’t be said about average. A batter doing well in average has a better shot at doing well in other categories. This could be reason alone not to give up on the category.


12 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Mark Harrison
14 years ago

Brian, I agree with the overall idea in the article, I’m just still not sure that the numbers show that as it doesn’t show how steals are correlated with runs or RBI. Also as we’re getting towards the point in the season where people are looking at specific catagories to move up in overall value is a bit less meaningless i’d have thought.

Eriq Gardner
14 years ago

That’s right, Brian. You nailed it.

I could break it down further.

For instance, steals and runs have positive correlation (.34) but steals and RBIs have negative correlation (-.31).

If you’re chasing steals by adding a speedster off the waiver wire, chances are you might be hurting yourself in RBIs.

AVG has positive correlation to both runs and RBIs.

Of course, that’s just the macro take. Each player has his own attributes. The point is that the from a wide macro-economic view, chasing average seems to make sense considering the wide supply and fewer opportunity costs.

Nutlaw
14 years ago

Right. A correlation to “overall value” unfortunately means absolutely nothing unless we know what “overall value” is and how it is calculated.

Do those numbers indicate that steals are actually less valuable than other categories, or do they merely indicate that your “overall value” forumula values them less?

Joe T
14 years ago

Eriq,

I agree with you that making up distance in BAVG is no more difficult than closing ground in other categories.

One reason not mentioned is that the target to be achieved might not be as far away as one thinks.

This is because, unlike with the counting categories, the competitors you must catch may move down toward you. The guy with the HR’s will never have fewer HR’s than he currently shows, but a couple of rivals may have a significantly lower BAVG by season’s end.

Brandon Heikoop
14 years ago

I think much of the problem with chasing batting average is the cost associated with it. Yes a team could go out and grab Albert Pujols, Brad Hawpe, or another high BA, high caliber offensive player, but the cost of those players is quite high.

By comparison, much of the reason that these average starved teams are where they are is because they are bound to players such as Ryan Howard and Jay Bruce. Yes, these players could be exchanged for a player with superior average, but at what cost? Would a team truly gain by swapping Howard with Johnson? That is an extreme case, but you get what I am saying.

Further, your correlation assessment needs to be considered in context. As individual statistics there is a relationship, but what makes a player like Grady Sizemore or Ian Kinsler so valuable is their help in every category. Brian Roberts is another one of those players. Certainly you can find an equal amount of steals from another second basemen, but good luck matching the average, runs, rbis, and homers. So in a bubble, steals are the least tied to overall value, but peeling a few layers of the skin, we see how important speed is.

How many players are putting up the numbers of Big League Choo?

Mark Harrison
14 years ago

I’m not too sure about how meaningful the Correlation numbers you show are. Does the low correlation to steal not just reflect the fact that some of the bigger sources of stolen bases are guys whos value otherwise is pretty small while most big HR guys are also RBI and Run machines?

David in Toledo
14 years ago

By coincidence, Kerry Whisnant at Dugout Central has an article “How Good a Measure is Batting Average,” introducing a simulation test.

Vince
14 years ago

To the Ryan Howard Point. Hes not saying trade Ryan Howard for Nick Johnson, but rather apply what he is saying to the situation. Meaning this: If youre trying to make up average and Ryan Howard is holding you down, you could attempt to move Howard for a more balanced power threat such as Carlos Lee, Shin Soo-Choo, Justin Upton, or anyone else hitting roughly .280 the rest of the way with pop. Then pick up Nick Johnson and slide him into 1B. Bam… +30 hits without a significant drop in any other categories.

Option two if guys like Johnson are already rostered. You could move Ryan Howard for packages like Votto + Morgan/Span/McCutcheon and achieve the same result. Basically, all he is saying is that batting average is not that hard to make up and that people are inlove with power and overvalue it.

Hector
14 years ago

There is an extra reason batting average is hard to make up.  The bottom teams in any league tend not try as hard once it’s clear they cannot win.  That typically means it’s easier for a good team to make up points on bottom teams late in the season—but this applies primarily to the counting stats.  The average of those bottom teams still tends to stay about where it was and there is little relative advantage gained by their reduced activity level.

To be honest, I think this is one of the biggest reasons people notice it being easier to move up in the other categories.  Bad teams can stand still in average but not in the other stats.

Eriq Gardner
14 years ago

Not sure I’d say that batting average isn’t hard to make up. But I do think people overestimate the difficulty. They focus on moving the average, which seems hard when several thousand at bats have been played. They tend to not look at the actual gap, which may be a lot more manageable.
I don’t think there’s much difference between an average gap and a gap in other categories. AVG is perceived as a rate stat, but if at bats are mostly the same throughout the league, really what we’re talking about is a hits deficit. Plus, we can turn any category into a rate stat. Your gap in HRs is also a deficit in HRs/AB, in all probability. Just a difference in perception.
I see the argument about teams giving up, but perhaps they are leaving batters with low averages in lineup. Not totally sure the real impact of teams giving up on the chase for AVG. I’ll consider it more, though.

Hector
14 years ago

I think you are right “if at bats are mostly the same throughout the league” then it’s really just a counting stat.  I also think that’s a good framework to look at a lot of people’s experience that it seems harder to make up average.

I think one reason that seems to happen in leagues is that ABs aren’t the same and bad teams that don’t rotate as much will end up with fewer ABs (relatively) in the second half.  Obviously, this depends on how active your league is and what the settings are, but even in very competitive leagues with prizes, managers in the last few spots are often unmotivated (competitive managers probably manage multiple leagues and would rather spend their time on their good teams).  It would be interesting to see if some fantasy provider had numbers on ABs in the second half by ranking, but i suspect you’d find a strong correlation between a team’s position in the standings and the numbers of ABs.  If so, that does mark a departure from the counting stats.

I think the other part of it is purely strategic.  Certain stats are easier to attain from the free agent pool than others.  For instance, gaining runs is going to be very hard because the best run rates are all taken.  However, you might be able to find a call-up late in the season who’s batting 6th and racking up RBIs or a speedster for steals.  Those guys are FAs but you won’t find that many guys who will have the same expected impact on average available.  So strategically, it’s a lot easier to make sure you have average first and then adjust for the other categories later. 

There are also some other practical aspects about average which make it tough to make up as well which involve the circumstances of having a low average in the first place.  First, there are some guys with really bad averages that are going to bring you down so long as you keep the guy, for instance, Adam Dunn or in the extreme look to basketball – Shaq.  So then it’s not a matter of the category, but rather that certain players can poison the category.  If you have a few of those guys on your team, it’s hard to makeup because they keep sucking you down.  Second, even if you have a guy who just slumps for a while but gets other stats (say BJ Upton earlier this season), those slumps in average can dig a pretty deep hole.  Now, of course, slumps in other categories could dig holes as well, but then that circles around to the point earlier which is that it’s harder to find batting average specialists in free agency.  There’s always guys who in retrospect did well in given month, but they are much harder to identify than guys that will give you steals or rbis.  Of course, if your league is active in trading and ichiro is always available for trade, then there you go.  But if you are relying on free agency to make up batting average, then that’s a whole different ballgame.

Anyway, that’s my thoughts, I like the correlation stats, but I would be interested to see another set of numbers which reflect how easy it is to find a guy to makeup ground in a specific stat for you.  And not just retrospect, but prospectively (i.e. I think it’s a lot easier to identify situations where some guy is going to get a new spot in the roster, etc., where he will get RBIs or steals, but that doesn’t mean he’ll get average).

Hector
14 years ago

Continuing a bit – average suffers from a prospective/retrospective look issue similar to hitters vs pitchers.  Prospectively, managers tend to draft hitters earlier than pitchers for a number of reasons (stat variance/predictability, availability of guys later on as free agents), although retrospectively this means pitchers are likely systematically undervalued on draft day.  But since we can only draft prospectively, that’s how we strategize.

And I think for some similar, some different reasons, batting average is similar.  Prospectively, the best strategy is to get guys on your team that have average and use your shopping around skills on others.  It’s not just the stat implications such as bad teams getting fewer ABs, it’s also just harder to find guys that boost your average as it is to find guys to get SBs or RBIs.

(technically runs are also hard to make up, but practically it’s often moot because it’s hard to be way behind in runs and still be in competition—if you’re behind in runs you’re probably behind in a bunch of other stats as well.  I actually wouldn’t be surprised to see that runs are harder to make up, but we just don’t hear people talk about it because only teams in contention talk about these things and if you’re out runs you’re probably out of contention.).