The 5 Tool Analyzer

When being considered as a potential draft pick by MLB teams, several topics often come up. What are his future plans (especially if he’s currently in high school)? How’s his makeup? How well does he rate with the 5 tools? Certainly there are more questions as well. Just as a reminder, the 5 tools for a baseball player are the ability to hit for average, hit for power, fielding, throwing, and speed. If a player has all 5 tools, then he’s considered to be a sure lock as a future MLB player. Although it’s rare for all 5 tools to actually remain once reaching the big leagues, it’s not uncommon for a player to be very solid despite lacking one or more of these 5 tools.

I wish I could take the credit for this idea, but this was first proposed by Kevin Dame at The Hardball Times with much better looking graphs. Rather than simply sticking to scouting reports, he decided to make a graph that showed which percentile a major leaguer rated in these 5 categories. He described each of the 5 tools as being measured by:

Bat: wOBA (weighted on base average)

Power: ISO (isolated power)

Glove: Range runs above average (use link for Arm as well)

Arm: If an infielder, use double play runs above average. If an outfielder, use outfielder arm runs above average.

Speed: Speed score (average of stolen base %, stolen base frequency, % of triples, and runs scored %)

In case you’re curious, all of these statistics are available at FanGraphs. Of course, using all of these statistics to measure the 5 tools introduces a few flaws, but for the most part I (and Kevin Dame, at least according to his THT post) don’t mind.

In his post, Dame was taking suggestions from people on which players he should study. He started with a few on his own, finding that Ben Zobrist was 2009’s most complete player…

… and David Murphy of the Texas Rangers was the most average player:

Dame did a couple other players as well, and was taking suggestions for who he should use in the future. He received votes for various players like Ichiro, Albert Pujols, Elvis Andrus, even Carlos Zambrano, but I noticed that Minnesota Twins were distinctly missing from the list of requested players. This is where I got my idea. I’ve been looking for excuses to use Microsoft Excel* ever since I learned how to use it, and this was a great opportunity.

* It’s not the best and I know I could probably find and use other programs, but it’s what I have and for the most part, it does what I want.

Remember the flaws that I mentioned above? You may have noticed that for the Arm portion, we cannot measure the ability of a catcher. Also, range runs above average does not exist for catchers (it’s not possible is what I mean) so we cannot measure a catcher’s Glove either. So, I do not have completed graphs for Joe Mauer (unfortunately) and Jose Morales (meh).

Second, I ignored the former Twins, but included the newcomers. In other words, I omitted Orlando Cabrera and Mike Redmond, but included J.J. Hardy and Jim Thome. Lastly, Dame apparently has “5 Tool Analyzer” trademarked, so I suppose for legal reasons I cannot use that name. How about “5 Tool Radar” since “Radar” is the type of Excel graph I used for this? Oh well, I suppose if Dame or anyone else gets upset, I’ll just take this post down. It’s not like I’m stealing blueprints for the wheel anyway…

How does this work, by the way? By Dame’s suggestion, I found each player’s individual rating in each of the 5 statistics needed for the 5 tools, and then found which percentile each player ranked from the 2009 season. Therefore, since Jim Thome did not have a rating for his Glove or Arm, he ranked in the 0th (is that the correct way to say that?) percentile, or 100% of major leaguers ranked higher than Thome in those categories. Alexi Casilla’s Speed ranked in the 94th percentile last year, so only 6% of major leaguers ranked higher than Casilla’s Speed last year. Fairly simple to understand.

One problem I did see was if a player did not reach the minimum number of plate appearances or defensive innings played to qualify for the leaderboard. To address this, I tinkered with the leaderboards on FanGraphs until the limit on plate appearances or defensive innings was low enough that the specific player qualified (you can fiddle with it yourself here). It’s not entirely fair to judge someone like Matt Tolbert differently than Jason Kubel, but it has to do. Besides, would you really want me to claim that Player X with only 250 PA last year ranked better than Jason Kubel when he had over 500? This is a flaw, but once again, it has to do. As for defensive rankings, I only compared each player to other players at the same position.

For the outfielders, they were compared against all outfielders (since guys like Denard Span played all 3 positions last year and FanGraphs allows outfielders to be measured at all 3 positions together than than each one individually) while infielders were only compared against players at the same position. For infielders that played more than one position last year, I used the position at which they logged the most defensive innings. I’ve specified each position next to the player’s name when necessary. Offensively, each player was compared against everyone else in the league, provided that those players reached the same minimum of plate appearances as the Twin.

Finally, on to the graphs.


Thoughts: I think I might need to reconsider my tolerance of Brendan Harris. Jim Thome would not be recommended for a team coached by Dusty Baker, and you can clearly tell he was a DH last year. Despite looking rather small, Justin Morneau’s graph is one of the largest of all the Twins from 2009. Alexi Casilla was really only good for running the bases last year. Don’t worry though, the outfielders offer more hope.


Thoughts: Clearly each outfielder was deficient in at least one area last year (except Delmon Young, who apparently is missing all of his tools) but overall, this is a much more solid group than the infielders.

Once again, this would be a really cool thing to do with all MLB hitters, though it would be a rather long process to endure. I’m also very jealous of Kevin Dame’s graphs, considering how much more colorful and less pentagonish they were. At least in terms of the non-Mauer Twins last year, we see that no one excelled in all 5 areas last year, some didn’t excel in any, but most players rated as above average in 2-3 categories. I don’t know if this really serves any purpose, but I do know that I just like graphs and numbers. Thank you Kevin Dame, for feeding my insatiable desires.


9 Responses to “The 5 Tool Analyzer”

  1. Doctor_Teh Says:

    I’m sure you have seen it, but these remind me very much of the diamondview series being done on Beyond The Box Score right now. They haven’t yet gotten to the twins, but if you want, I can try to remember to post a link to it when they do.

    The main difference I see on there is that they consolidate Arm +Glove = Defense, which will change the ratings since they used just straight adjusted UZR.

    It is kind of sad to see how paltry our team looks without Mauer…I guess we have a few good people at each tool, but man, those are not big graphs, lol. Nice work on producing them though, they look good and convey the information easily.

  2. Josh Says:

    I’ve seen the Diamond View series as well and it is also very interesting. I believe they used JJ Hardy for one of their samples. Good work Andrew, however, I am a bit discouraged to see some of the small pentagons (right?) of our players.

    • Andrew Says:

      Yes, if evenly balanced a player should have a pentagon-like shape. Just as a primer until I get this figured out, Mauer obviously rated highly in Bat and above average in Power, while being below average in Speed. Jose Morales was good with his Bat, but rated horribly with Speed and Power.

    • Andrew Says:

      Correction: Every player’s graph displays a pentagon. Even Jim Thome’s is a pentagon because his Speed was actually in the 1st percentile, while his Arm and Glove were the 0th percentile. His appears to be a triangle, but if you zoomed in, you could tell that it’s actually 5-sided.

    • Andrew Says:

      *facepalm* Another correction: Thome’s is not a pentagon, it’s a quadrilateral. I quit.

    • Jon Peltier Says:

      I reviewed the diamond view approach in Composite Baseball Player Evaluation.

      These radar or spider charts are pretty, but Radar Charts are Ineffective. It is hard to compare two players, because they are plotted on separate plots. Small values are hard to distinguish, as they get lost in the center of the web. (The circular version is just silly.)

      If you use a parallel coordinates approach, as I suggested in my post, you can plot multiple players on a single chart, and make better comparisons.

  3. fishinbuddy Says:

    I really like the looks of the graph, but I have a quibble with the implementation. It looks like Delmon Young has no arm. I believe that scouts and other players agree that he has a strong arm. Perhaps the stat selected to measure arm strength is missing something. Even if you give him credit for a slightly above average arm his graph will look much different.

    • Andrew Says:

      Ah yes, I can understand how that is an issue. Clearly Delmon has a strong arm (2 years ago, he showed it off by throwing a ball all the way from the LF fence in the Metrodome to 2nd base on the fly), and the accuracy is about average to possibly above average. However, the metric used to measure a player’s arm is dependent on the results from the season. Yes, he has a strong arm, but it’s a matter of how he used it, and his low rating suggests that he didn’t use his arm very well last year.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: