Introducing the 5 Tool Analyzer
I’ve always been intrigued with the concept of the “5 tool player” and have been experimenting with ways to visualize what such a player looks like. The 5 Tool Analyzer maps a player’s skill (relative to his peers) across 5 offensive and defensive measures. The resulting map reveals a few things at a glance. First, the overall size of the map indicates how good the player is. Big glob of color = good. Small glob = bad. Second, it shows where they are strong and where they are weak. And finally, it also shows how balanced a player is. Circular glob = well-rounded (literally and figuratively). Weird-shaped glob = some strengths and some weaknesses. Let’s start with a quick visual introduction:
Now that we’re clear on what we’re measuring, let’s take the 5 Tool Analyzer for a spin. To start with, I was curious to see if I could find a player who was “perfectly average” in 2009 and I think I found him in Dave Murphy (he’s roughly at the 50th percentile across all 5 dimensions).
By now you’re probably wondering if there is a “perfect player” in baseball (a player who is the best across all 5 dimensions). Using the 5 Tool Analyzer, the perfect player would be represented by a perfect circle filling out to the 100th percentile. What I discovered was that most all-star players have at least one weak (or average) area. But there were a few players who came close to perfection, and you’ll be surprised to see the one who came closest:
Chone Figgins was an intriguing example of a player who truly excels in all but one category (in his case power):
Adam Dunn’s visual map paints a classic portrait of an “all bat” player. He has DH written all over him:
Jack Wilson represents the inverse of Dunn – an “all glove” player.
So that’s a quick tour of the 5 Tool Analyzer. if you like it and want to see more players visualized, name your player!
I’d like to see these used to represent a players development over the seasons in 3D. No small task I know!
Wow. I love the “blob/color” approach, Kevin. I think this is the best example of this sort of graphic I’ve seen.
Having said that, I’ll raise the usual issue I have with these sorts of graphs: they graphically imply that all five dimensions are of equal worth, which is clearly not true. I don’t know of any easy way to fix this issue, but it’s important to remember.
While the results might not be as aesthetic, I think the analytical value of this visualization might be improved if the “arm” and “glove” scales were adjusted to better reflect their relative importance. There is no position for which these are as important as bat and power. And the “arm” component for 1B-men is much much less important than, say, C or SS or RF.
You’d wind up with ovoids instead of circles, but the overall size of the “blob” would better reflect the player’s value.
A-Rod
Jeter
Longoria
Derrek Lee
Pujols
Jeter
Ryan Braun
Yuniesky Betancourt
Jeff Francoeur
wOBA already includes SLG. You would be better off putting in OBP in its place.
I’ll also agree with the others that the five dimensions as presented makes it seem as if they are all equally valuable.
In order to do what you want, and keep the diameter the same for each dimension, you need to change the “slice” of each to not be equals. For example, if the spread in SLG (in terms of runs) is twice that of the glove, then the slice of the SLG should be twice the area of the glove.
I can try to work out those numbers if you like.
matt kemp
andre ethier
james loney
casey blake
Dave had said: ‘they graphically imply that all five dimensions are of equal worth, which is clearly not true’
I am not sure why this is clearly untrue. At times a players arm may be as valuable as anthers speed or power. I think players having a big blob in one area of the graph shows their strengths and weaknesses and is useful for valuating that player. Just as this graph would indicate the value of Albert Pujols it would indicate the value of Jack Wilson. I think all 5 dimensions are of equal value.
So do we call players like Figgins “Pac Man” players?
Are the percentiles position adjusted? ie. Kendry Morales might have the highest Double Play Runs Above Average for a first baseman, but I don’t really see him as having the arm of a 5-tool player. I’m not sure how you compare infield and outfield arms but it seems that some calculation is necessary.
Similarly, DPR is on average 0 for second-basemen and shortstops, but shortstops should get a higher rating for the Arm category – is it possible to adjust for that?
Following Studes, Jackie Gleason, and Tango, I would put league average at the 50% line and move in and out from there based on runs above or below average. Whether to take absolute runs or adjust for playing time, (runs per 150 games, or 600 (650) plate appearences) is something I could go either way on. It would make a difference with part time players such as Nyjer Morgan.
I like Tango’s suggestion a lot. I hadn’t thought of that before.
This is awesome. I think “everyone” is my only request.
These are absolutely fantastic. Reminds me of the old Little League Baseball game ratings on NES. And I second Mitch, everyone will do nicely.
Pujols
Michael Bourn and Willy Taveras
Ichiro
Adam Jones
Matt Kemp
Elvis Andrus
Carlos Zambrano
Chase Utley
Kevin—
These are awesome.
Showdown between The Upton Brothers.
Vernon Wells. Just want to see what 21 million dollars get you…..
I agree that these 5 tools are not equal in value, but I don’t think they need to be changed at all. These are merely ranking players in relation to the rest of the league.
By the way, how do you plan on judging arm and glove for catchers and 1st basemen?
If you incorporate Tango’s suggestion of altering the size of the different elements, you should also convert all 5 tools to a Wins (or Runs) Above Replacement metric. Then the area inside the curve would have the interpretation of total WAR so that you would have a visual interpretation of total ability.
I don’t see why you have to convert each metric to runs above/below replacement. Tango’s idea to change the slice of the pie effectively does that for you. No need to change anything else.
Hey Kevin, nice work, these graphics are by far my favorite of what you’ve posted here.
How bout someone like Lastings Milledge, who hasn’t fully developed yet?
pedroia
Jayson Werth
Very nice. A question: for the rate stats (bat, power, speed), who gets included in the percentile calculation? Do you have some PA requirement?
Any thought of replacing the circle with a fractal?
Those are quite beautiful.
The “growth” visualization via the evolution of the graph over time would also be quite a site. As would a rendering of a lifetime “blob” in 3D.
And you forgot the “intangible” stat … gotta have that in there. A full text search of baseball “human interest” columns coordinating positive and negative press, with park factors of course … for the inevitable “intangibly above average” stat!
Not adding much here, but damn those are some sexy graphs.
Would Yuniesky Betancourt’s graph look like a black hole? This is a great way to visualize the Vortex of Suck.
Adam Jones
Nick Markakis
Brian Roberts
I was going to say it, but Tango beat me to the punch: wOBA includes power numbers, so wouldn’t OBP make more sense there? Adam Dunn’s chart for example probably wouldn’t change, but I can see this overrating a hacktastic power hitter like Miguel Olivo or Mark Reynolds, or someone else whose wOBA is heavily dependent on his power (since he has twice the area to fill up)
Robinson Cano, Bengie Molina
I’m actually not really surprised that Zobrist was closest to perfect, to be honest.
Tango, I think you’re wrong here with the suggestion to use OBP instead of wOBA. Walks are not one of the five tools. As much as it may hurt to say it, I think batting average works best here.
You can’t use OBP in looking at the “bat” tool because such a big part of OBP is walks. What’s OBP minus BB? Batting average.
Dan, maybe BB isn’t one of the “traditional” tools, but these days, most GMs and scouts value both the ability to hit for average AND walk. So I think OBP > BA for this exercise.
Tango, I think you’re wrong here with the suggestion to use OBP instead of wOBA. Walks are not one of the five tools.
But they clearly should be.
First let me say I absolutely love your idea.
I would like to see comparisons by position and league.
Pujols, Gonzalez, Fielder, Lee, Helton, Berkman, Howard, Cantu, etc… for the NL for example.
And I really like the idea of weighting the graphic by size of wedge as Tango was talking about. That would give a better representation of overall talent.
Thanks!
JOE MAUER!!!
Justin Morneau!!!