2B and 3B
This afternoon, we touched on position adjustments and why they’re needed. In the case of positions such as 1B and SS, it doesn’t take a lot of explanation as to why there needs to be a bridge so that we’re not describing an average SS equally valuable to an average 1B. However, for a few other positions, the line is much less clear. The three positions where the respective defensive abilities is less clear are second base, third base, and center field.
In a lot of circles, you’ll hear defense described in two sections – “up the middle” and “the corners”. You’ll often hear the up the middle spots – C-SS-2B-CF – as being premium defensive positions, where teams will look for gloves first and bats second. The opposite is true in the corners – 1B-3B-LF-RF – where teams want offense and consider defensive value a bonus. By this definition, second base and center field are glove first positions, but it’s something of an afterthought at third base. If we bought into this notion, we’d believe that third baseman, as a group, are inferior defenders to second baseman and center fielders.
This notion is reinforced by the fact that third baseman are better hitters than second baseman. Major League 3B hit .266/.336/.436 last year, while Major League 2B hit .276/.338/.409. This has been a trend for a while, and since the statistical analysis revolution taught most of us to think in terms of player value in terms of above or below average in relation to the offensive average of that position, we’re used to drawing conclusions that worse hitting positions are better fielding positions. Since 2B hit worse, they field better. That’s the assumption. The problem, though, is that it lacks evidence. There is simply not a large fluctuation in defensive performance when a player moves between the two spots.
Take Akinori Iwamura, for example – he played a full season at 3B in 2007, posting a -2.6 UZR/150 at the hot corner. The Rays moved him to second base for 2008, and he posted a +0.6 UZR/150 at the keystone. He’s not the exception, but instead, performances like this are the rule. When a player moves to 2B from 3B, or vice versa, their defensive performance remains relatively stable. There’s just no data to support the idea that second base is a premium defensive position and third base is not.
While there are different skills necessary to succeed at each position (arm strength being more important at 3rd, range more important at second), the pool of players who can succeed at one is mostly made up of the same pool of players that can succeed at the other. This is because the population of both positions is made up almost exclusively of players who were deemed inadequate for shortstop. At one time, second baseman and third baseman were both called the same thing – bad defensive shortstops. From there, they were separated into 2B and 3B pools, but the evidence suggests that the crop of players who end up at 3B are better overall players than their 2B brethren. Why? I have a few theories, and we’ll get into those tomorrow.
Ryan Braun also made this pretty obvious with his move from 3B to LF netting a gain of some 30 or so runs.
It will be interesting to see what Iwamura does in coming years, since we only have one year of data at each position. I certainly want to and will look at a bunch more examples. It just seems that handling the double play and the range required makes 2B a much more valuable position.
Remember, though, both third baseman and second baseman are made up of former shortstops. Almost every major league third baseman has years of experience turning the double play from the SS side of the bag.
What is the separation criteria? Range vs. Reactions is what is appears to be to me. 3B is the “hot corner” many of the stud defenders we see are laying out behind the bag, or cutting off a ball before it gets into the hole. Second Baseman can be found in short right field sliding on the grass, or well behind second base. It seems they have significantly long period of time to react in addtion to more ground to cover than third baseman.
Just my two cents.
I’ll be interested to see where this goes. Whatever the skills are that make you a 3B rather than a 2B, they seem to be more readily available in players that can hit well. That, to me, suggests that the “bonus” should be greater for 2B.
Size? Increased bulk would usually lead to decreased range but better core hitting skills.
There are also issues of which players move between positions. Between 2B and 3B, it’s players with equal levels of arm/reactions and range/pivot, so they tend to field the two positions equally well. But of the people who don’t switch positions and the player pool as a whole, it might very well be that more of them have capable arm/reactions skills than range/pivot skills, making them more appropriate 3Bs.
Does UZR address double-play turning? If not, that’s something that might make 2B look harder.
However, smarter people than me (and most people here) have tackled position-changing studies like this, so it’s going to take proof to convince me 2B and 3B AREN’T pretty similar defensively (in terms of value).
UZR on FanGraphs currently does not include infield double plays. They’re currently treated as a regular out, but mgl is going to try and get them in before the season starts.
Obviously there is some defensive element seperating the two of them, since the market equilibrium is tilted towards better offense at third than at second. If there wasn’t something preventing 3-baggers from becoming 2-baggers, then the talent pool would be more evenly split than it currently is in terms of offense; there’s no reason to keep a power-hitting third baseman out of a job behind a weak-hitting second baseman unless the second baseman has some sort of defensive advantage over the third baseman. I think that logic and statistics have permeated MLB front offices to such a degree that we can no longer blame market imbalances on GM irrationality. I’d be curious to see how much data you have, because the list of guys with several seasons (and we know it takes numerous seasons for defensive stats to be reliable) at each position seems like it would be quite limited.
Teams are getting a lot smarter, but we certainly can’t justify every decision they make as rational. The closer is still used by every team in baseball despite being a huge waste of a relief ace.
I think they require different skiils though. For instance, David Eckstein would play a poor 3B because he has such a bad arm so it would be hard to make such a long throw and not have guys beat it out, on say a backhand play. But at second he is fine because he is relatively quick (need to cover more ground) and you don’t need a lot of arm to throw it from second to first. This can go vice-versa I think you don’t see guys like this try the conversion often and that is why the sample is skewed.
Bigger guys with good arms are better off at third than first but don’t have the range to play second, but they tend to hit better.
Smaller guys usually don’t make it if they aren’t quick and can handle either second or short so you see the lesser guys at second and because of their size they are usually not as good of hitters.
What percentage of balls in play are hit into each defensive player’s zone?
Check this out, Don…
http://www.hardballtimes.com/main/article/infield-defense-mdash-back-to-basics/
That’s an interesting link, thanks. Odd that so few balls are hit back up the middle as compared to straight at the 2B and SS… there must be some physics involved there.
I have a question about first base and how defensive skills are measured in most statistical systems… Do any of the formulas take into account a first baseman’s skills digging out throws in the dirt, coming off the bag and tagging runners coming down the line, being taller than average and having more reach to field throws while keeping their foot on the bag to record the out, and so on? I would call it the Olerud factor. He was known as a fine defensive first baseman but we all know he was not the fleetest of foot nor possessor of cat-like quickness. He was however 6’5″ and maybe the best I’ve ever seen at corralling poor throws from infielders. That has to count for something in terms of errors saved, i.e. plays that if fielded by an “average” first baseman would result in throwing errors on infielders. Sexson on the other hand was tall with great reach but almost seems afraid of scooping throws in the dirt, often waving his glove as if he was dodging a charging bull. Anyway, being fairly keen on the value of defense, this is something I’ve paid close attention to and haven’t ever seen broken down regarding the defensive value of first basemen. Any answers or links to articles?
No, for the most part, the defensive metrics available don’t take this into account. However, studies that have been done on the issue suggest that we’re not missing more than a couple of runs either way.
Will the real Dave Cameron please stand up?
Anyway I got a question Dave, since UZR is comparing guys against others who played his position this year, would there be any use for maybe setting the positional defensive baseline over the last 5 seasons, instead of just one? Just curious, if it would make positive difference at all in terms of evaluation.
I doubt it would make a difference, but perhaps we’ll add this to the list of potential research options down the line.
Alright, next question, is probably harder to answer, would there be any value to ratind defense for a player over 3 or 5 seasons, as in for ’07 you take ’06-’07-’08 and somehow average or weight the values to eliminate some noise? Or would that be little help as well?
Yea, that’s basically how I suggest looking at them. I talked about that a bit in the defense as inferential statistics post a few weeks ago. Due to the sample size issues, defensive statistics are best viewed through a two to three year lens.
I always thought of 2B being more important, run-prevention wise, than 3B because 2B handles the ball more. So if you had a great third baseman and mediocre second baseman would you be better served moving the great third baseman to second to take full advantage of those skills? Should Adrian Beltre be playing second?
As a co-founder of USSMariner.com, I’ve enjoyed writing about well run teams for a change.
If/when you decide to change this byline, you should do a post about what Z did to break the camels back: “Today Jack Zduriencik offered a minor league contract to Player X, a low-cost high-upside move that makes it official: The Mariners are now a well run team.” We could be there already?
Howdy! Excellent thought, but will this genuinely operate?
Perhaps it is an issue of pure durability. The average 2B (min. 1000 innings) over the past 3 seasons started 277.3 games and played in 290.1. The average 3B (min. 1000 innings) over the past 3 seasons, however, started 263.6 games and played in 277.1. Despite the frailties of Ian Kinsler, Aaron Hill, Chase Utley and Rickie Weeks, 2Bs, on average, saw more time on the field than did 3Bs. Then again, 3Bs have had their share of injured played over the past 3 seasons in Joe Crede, Adrian Beltre, Mike Lowell, Aramis Ramirez and Troy Glaus.
It’s surely not a distribution based on “glove or bat.” While MLB 3B hitters (min. 500 PA) over the past 3 seasons hit .270/.341/.438 (.779 OPS) and 2B hitters (min. 500 PA) over the same 3 season sample size hit .278/.341/.416 (.757 OPS), the average 2B (min. 1000 innings) had only a +0.6 UZR/150 compared to a +0.8 UZR/150 for 3B (min. 1000 innings).
Any other postulations as to the reason for the distribution? Is it, as you suggest, something cosmetic, or is there something not quantified in the data at play? Does playing 2B, which requires a player to “go both ways” in making outs, require more althletic, healthier types at the helm than does 3B, which generally only requires a player to go to his left in making 80% of his plays? Does a stronger bat equivocate a stronger throwing arm?
http://gameofinches.blogspot.com/2010/02/skill-and-health-distribution-of-2bs.html