The Value Of Below Average
“Average” is a word that contains negative connotations in our society. Nobody wants to be an average student. Nobody wants to be an average accountant. Nobody wants to be an average engineer, nor would they want an average mechanic fixing their car or an average carpenter building their house. Similarly, fans just don’t seem to get excited when they acquire an average or, even worse, below average player. Due to the limited supply of great or even good players, however, these players are absolutely necessary for a winning team. Let’s take a look at the distribution of talent as measured by WAR from last season. For simplicity’s sake, this discussion will be limited to the pool of position players.
The supply of above average players simply is nowhere near enough for every team. Even when we restrict ourselves to players receiving over 300 plate appearances, we still see 137 of 305 falling below average and 28 of those 137 below replacement level. Of course, given the technical definition of average, that’s not surprising at all.
The problem with below average players is that they’re typically not members of playoff teams. At least, this is the perception. We know that each team must start at least eight position players. All but four playoff teams since 2002 have had eight players reach 300 PAs, which means these players played a significant role . If average or below average players aren’t starters on playoff teams, we should expect the 8th best players on these teams, as defined by WAR per 600 plate appearances, to be above average players. Let’s take a look at these players.
The yellow line shows the mean for this sample, and the red line shows 2.0 WAR, or league average as defined by the statistic. Remarkably, the average for these 8th best players – regulars in many cases, important players at the very least – is 1.0 WAR per 600 plate appearances, or exactly 50% of league average. These players have been average or above 11 times; they have been below replacement.
There are few teams, if any at all, that wouldn’t benefit or wouldn’t benefit from the addition of a 1-1.5 WAR player, simply due to how many below replacement or marginal players have jobs. Of course, in some situations, the value of a player to one team is lower because of positional needs, but the data simply doesn’t support the notion of a below average player not deserving one of the 750 roster spots available in Major League Baseball. The next time you hear that your GM has picked up a player that projects as a 1-1.5 WAR player, don’t simply write him off as a guy to rot on the bench – he may be that last starter on the next playoff team.
0
“There are very, very few teams, if any at all, that wouldn’t benefit or wouldn’t benefit from the addition of a 1-1.5 WAR player, simply due to how many below replacement or marginal players have jobs. ”
This is kind of a silly argument. If there are more average players, then replacement level moves up, and those players become below average again.
Not necessarily.
One team acquiring an additional average player implies that another team is losing an average player.
Like when a team trades an average player in exchange for prospects.
The pool of average players stays the same and the team that acquired the average player takes a small step forward this season.
Of course, plenty of average major leaguers have been minor league free agents. Do your scouting homework and find a couple of these guys, then you raise the overall level of competition.
PA makes up a significant portion of value in terms of WAR. In this sort of analysis the player may indeed be above average, but is not played on a regular basis. Thus they are a below average contributor, but an above average player when given an opportunity. Adam Jones put up 1.8 WAR in 119 games (519 PA), if given a full season he would have cleared the “average” mark.
Playing time is a proxy for talent, and if you wanted to remove playing time from the discussion you should really normalize the data to 600 PA. For those players with less than 600 PA regress their numbers using their projections.
Yes, the problem with this post is that Jack used a counting stat (WAR) to measure ability instead of a rate stat.
You’re misinterpreting WAR. Plate appearances only contribute to WAR if you perform at a level above replacement level. Simply getting at bats does not help. This is obviously why you see players who have negative WAR; they’ve performed below replacement level.
Correct, but the problem is that Jack used the standard for league-average over 600 PA and then applied it to players who received less than 600 PA. If somebody accumulated 1.5 WAR in 300 PA, he was above-average when he played, but this graph would show him to be below the average line.
Gothca. That’s a good point.
—”but the data simply doesn’t support the notion of a below average player not deserving one of the 750 roster spots available in Major League Baseball.”
Well duh, of course there have to be lots of below average players in MLB, by definition.
wasn’t it eisenhower who was outraged that “half of our students read at a below average level?”
This seems obvious to me – to the point that I’m not sure I understand the post. Am I missing something?
I mean, if your worst regular hitter (NL, 2nd worst in AL) is still in the top 50% of players then you probably have an outstanding offense, don’t you? How many teams get more than 2 WAR from the 8 hitter? probably almost none.
I say hitter because WAR is affecte most by offense – less by defense and baserunning – and according to the post, only position players were involved in this analysis.
This matters because many average fans and lessor readers of this site dump on players who are average or less. They lack proper perspective.
I had a long debate with a serious baseball fans earlier this year that the Nats were hasty in designating Bergman for assignment. Bergman is a below average major league pitcher, but he is better than about 20% of all major league pitchers currently on rosters and thus deserves a regular role at the back end of some rotation. I could prove at that time that he was in fact better than roughly 20% last season and over his career. He was demoted with a tiny sample of stats this season. The others felt you would never win a pennant with Bergmans in your pen. The fact is teams win pennants all the time with Bergmans in the back end of the pen.
All true, but there is one interesting question here (which the author doesn’t really get into) about what the shape of this curve SHOULD look like if teams were perfectly efficient at replacing players with better substitutes from the minors.
naturally this curve is skew right toward higher WAR. In an ideal world, the left “tail” of this graph should be even tighter toward the median (more useful to look at the median in this case). This would arise in a perfect replacement market when teams could immediately identify and act on a player that goes below replacement level. The fact that this skew is a bit fatter points to the efficiency of baseball teams to do just that.
The high bar for ’02 might be Spiezio?
Here’s a good example of fans who don’t appreciate “average” players. A fan here complained about the Cubs signing of Marlon Byrd this offseason. When I pointed out that he’s been significantly above average the last three years (I said 2.8/year, so either my math was off or his UZR really got bumped down by the recent adjustment but regardless is now 2.3/year) he said that he would make a nice 4th OF but he certainly wouldn’t want him as a starting OF.
http://www.fangraphs.com/blogs/index.php/cubs-land-marlon-byrd/#comment-117794
Here a fan is shown that a player is better than 50% of starting OFs and he still only wants him as a 4th OF; and he claims to value 4th OFs; just think of what the average fan thinks of average and below average players.
I’m going to assume this is in response to Matt Murton and his value as a player.
Let’s take a look at the playoff teams from 2009 and who the 8th WAR position player (unadjusted for PA) are for each team as well as the OF positions.
Yankees – DH Hideki Matsui – OF Swisher, Damon, Gardner, Cabrera
Murton’s not starting for this team and he’s not the 5th OF as Hinske and Hairston Jr. each play the utility/OF role. Hairston and Hinske are no longer with the team.
Red Sox – SS Alex Gonzalez – OF Bay, Drew, Ellsbury, Baldelli, Kotsay
Murton might play the Baldelli role but he’s not a local boy, but Kotsay can play 1B. Both Baldelli and Kotsay were replaced after the season.
Angels – C Napoli – OF Hunter, Rivera, Abreu, Willits, Matthews Jr.
Willits and Matthews Jr were bad, but both are considered (or had been considered) good defensive CF. Quinlan played some OF but is a backup infielder as well. Murton might have had more value than Matthews Jr. but he doesn’t fit the role well
Twins – CF Carlos Gomez – OF Span, Kubel, Cuddyer, Gomez, Young
Would Murton have had more value than Young or Gomez? Yes on Young and maybe Gomez (if they move Span to CF), but Young and Gomez were considered better prospects. Gomez was shipped out and Young is playing better in 2010. I doubt Murton fits in here.
Phillies – 3B Feliz – OF Werth, Ibanez, Victorino, Francisco, Stairs
Francisco probably is a good comp for Murton. Stairs is a PH and veteran presence. With Francisco playing worse this year, I wonder if he’ll be back. If he’s dumped, maybe Fangraphs will write an article wondering why he’s looking for a job. In any case, $1.5 for a defensive replacement is a bit steep.
Cardinals – 3B DeRosa – OF Holliday, Ludwick, Ankiel, Duncan, Schumaker
Ankiel was good in 2008 and Duncan was ok, but both were dumped after the 2009 season. I doubt the Cardinals would have taken Murton over any of those two to start the year – especially with one guy’s dad as the pitching coach.
Dodgers – OF Pierre – OF Kemp, Ethier, Ramirez, Pierre (no real 5th OF)
Pierre was actually ok in 2009 and they didn’t go for a 5th OF so I don’t think Murton lands a job
Rockies – 2B/3B Stewart – OF Smith, Gonzalez, Hawpe, Fowler, Spilborghs
All were either better hitters historically entering the season than Murton or a much higher rated talent (Fowler). Maybe Murton rakes in Colorado but I doubt he’d offer more than what they had to enter 2009.
So based on this quick look, where does a Murton-type of player get a look? Boston? The Angels? In Strat-o-Matic, Murton could fill a 4th OF slot if he hits, but if he’s looking for a $1M+ salary, Japan’s the place to go. Quite frankly, Murton’s easily replaceable. He doesn’t offer much value unless he hits. And if he doesn’t hit, he gets dumped.
In any case, how many of those #8 WAR players are still with the team? 2 Napoli and Stewart. So you either produce or you’re gone.
Dude, what teams’ worst every-day player was worth ~4 WAR? That’s absurd.
I didn’t see anybody above 3 WAR on that graph. The ’09 Yanks had Matsui check in at 2.7 for the eighth-most valuable position player, and I think that high bar toward the right is the ’02 Giants (just by guessing the pattern – I know the one to its immediate left is the ’02 Angels), but I can’t figure out who their eighth guy is that would be that high.
“lessor readers of this site”
Thus providing a data point against the theory that PhD’s are either 1) more empathic or 2) better spellers than non-PhD’s.
Don’t you mean lessor spellors?