In my chat last week, reader “dandycharger” asked me to name all of the genuine five-tool players in the majors, hoping to illustrate how loosely the term tends to be applied. Since I’m always happy to honor a reader request (especially when it supplies me with an article topic), I decided that I’d hold off on offering an incomplete, off-the-cuff, chat-length response in favor of writing up a semi-complete, Overthought™, column-length response, which brings us to today’s piece.
Tools are a slippery subject. Before we dive in, let me offer a quick refresher on which tools we’re talking about. (You can also look here for a lengthier take.) The wording may vary slightly depending on the source, but the proverbial five tools are as follows: hitting for average, hitting for power, foot speed, throwing ability, and fielding ability (the Dickson Baseball Dictionary calls those last two “throw” and “catch”). “But Ben,” you might ask, “isn’t foot speed linked to fielding and hitting for average? And shouldn’t throwing be considered a component of fielding ability?” To which I’d respond, hey, I didn’t decide what the tools were—this particular assortment of skills was handed down on stone tablets by the great scout in the sky long before I became a prospect.
As reader Bob in Seattle pointed out further down the chat stream, every player is a “five-tool player,” if you want to get technical about it; notwithstanding the incongruous sight of Jorge Posada on the basepaths or behind the plate, every major leaguer possesses at least some trace of talent in each of these areas. The question, then, is where to draw the line separating the haves from the have-nots, which can’t help but be an arbitrary distinction, since tools are distributed along a spectrum, not on a binary basis.
Think about it—at what point does it become possible to say that a player doesn’t “hit for average”? When his average falls—speaking of arbitrary thresholds—below .300? When it fails to match the league average, slips one standard deviation below the mean, or reaches the Mendoza line? When the player goes 0-for-the-season? There’s really no right answer, and attempting to come up with one is an exercise that can be resolved only by making a subjective judgment call, much like trying to pinpoint the exact magnitude of movement at which a cutter should be classified as a slider, or the angle at which a line drive should be considered a fly ball.
What’s more, tools are transient: It’s possible to possess one for a time, only to surrender it to age, injury, or overconsumption. It’s also worth asking why we should even bother to count how many tools a player has at his disposal; after all, isn’t all that really matters the bottom line? Still, while how many wins a player brings to the table is the most important thing, the manner in which a player puts his wins together has implications for his aging trajectory and likelihood of a repeat performance. Plus, talking tools is fun.
Much as we could quibble about whether these particular five tools provide the best representation of a player, the concept does have its uses, both in the scouting field and as an extension of the baseball lingua franca. At worst, the “tools” framework serves as a convenient shorthand, much like the notion of assigning theoretical rotation slots to starters or labeling a given pitcher an “ace.” (Even if the logical question that arises from those abstract terms is “Sure, but in whose rotation?”)
With that throat clearing out of the way, let’s get down to searching for the current crop of five-tool players. Since home-to-first times and FIELDf/x-like measurements of range and arm strength haven’t made it to the public domain, we’ll have to use statistical proxies for each of the tools. I settled on batting average (hitting for average), isolated power (hitting for, you guessed it, power), our rebuilt versions of BRR (speed) and FRAA (fielding), and the throwing component of Tom Tango’s Fans Scouting Report. Colin Wyers has developed a provisional arm rating that will soon be incorporated into FRAA (let’s call it FRAArm for now), but for a variety of reasons—most of them stemming from the metric’s incomplete assessments of infielders, whose arm strength is difficult to isolate from their other work with the leather—averaging the FSR’s three throwing components (release/footwork, throwing strength, and throwing accuracy) yields a more representative crop of five-tool players.
It’s also worth noting that this five-tool APB includes a position adjustment for FRAA intended to save us from judging all the candidates against the same baseline, since an average-fielding first baseman and an average-fielding shortstop probably can’t be said to have the same gifts when it comes to the “fielding” tool. This still isn’t a perfect method—for one thing, BRR is a measure of more than pure speed, since it accounts for the quality of a player’s instincts and decisions on the bases—but it works well for our purposes.
Based on a conversation with Kevin Goldstein, BP’s primary conduit to the scouting world, I confined the search to players who qualified as average or above in each statistical category; according to Kevin, a theoretical player who earned a rating of 50 or above for each tool on the 20-80 scouting scale would be considered a five-tooler. Since some of these stats take time to become reliable, I limited the pool to players with at least a thousand plate appearances over the last three seasons. As a result, this group won’t include anyone who debuted during the last two seasons and didn’t accumulate the necessary playing time to qualify, but the results should be fairly reliable. Two additional notes: FRAA totals in the table below are un-regressed, and FSR numbers aren’t included (thanks to my limited appetite for data entry and Excel’s SUMPRODUCT function), but I’ll vouch for the fact that all of these players’ arms rated as average or above—in fact, the only player who passed the test of the first four tools with flying colors and flunked throwing was Kelly Johnson. Without further ado, on to our 17 five-toolers, ranked in order of descending PA:
Name |
BRR |
POS_ADJ |
||||
2041 |
.331 |
0.304 |
2.6 |
120.3 |
-37.2 |
|
2004 |
.304 |
0.185 |
1.1 |
30.2 |
-2.8 |
|
1947 |
.278 |
0.188 |
5.6 |
56.7 |
-20.5 |
|
1940 |
.271 |
0.169 |
11.5 |
6.7 |
-2.5 |
|
1936 |
.275 |
0.164 |
1.2 |
41.8 |
-7.3 |
|
1905 |
.284 |
0.218 |
3.7 |
36.4 |
-2.7 |
|
1840 |
.283 |
0.237 |
7.8 |
67.3 |
6.9 |
|
1810 |
.279 |
0.234 |
8.3 |
29.0 |
-15.5 |
|
1730 |
.286 |
0.184 |
2.9 |
54.7 |
6.2 |
|
1724 |
.286 |
0.251 |
3.3 |
31.1 |
4.1 |
|
1707 |
.273 |
0.216 |
2.8 |
27.1 |
-16.6 |
|
1683 |
.285 |
0.193 |
11.7 |
34.6 |
-2.2 |
|
1539 |
.285 |
0.177 |
1.6 |
27.6 |
6.0 |
|
1532 |
.289 |
0.159 |
11.1 |
-18.7 |
20.1 |
|
1318 |
.289 |
0.197 |
3.4 |
-1.9 |
3.1 |
|
1269 |
.299 |
0.220 |
8.0 |
31.4 |
-3.2 |
|
1093 |
.292 |
0.263 |
0.9 |
75.2 |
-11.4 |
There are a number of related topics to investigate before we leave this subject behind: projected 2011 five-toolers, five-toolers still tooling around the minor leagues, and historical five-toolers, just to name a few. If you’d like to see any or all of those topics explored in an upcoming article, have a suggestion for refining the player selection process, or want to mention a player whom this method might have missed, leave a comment below.
Thanks to Colin Wyers for research assistance.
Thank you for reading
This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.
Subscribe now
I like the idea of seeing how this plays out for some of the youngsters who missed your years/PA threshold. Also, it might be interesting to see who the best non-5s are at each position, and what specifically holds them back from being a 5-tool.
I'd also be interested to see a thorough analysis of 'ace,' with input/opinion from a whole bunch of the BP authors. How do each of you define it, and who makes the cut by that definition.
One option you can consider is to use standard deviations in the 5 categories, and simply multiply them all (setting negatives to 0). This way, a guy who is barely below in one category (Tulo gets 0) and barely above in another category, will both come out low.
You can also try "standard deviations above replacement", and add 1 to each SD before multiplying them (and setting the rest to 0). Say for example Tulo is +2SD above the mean in 4 categories, and is -0.1 SD in the 5th. You give him 3,3,3,3,0.9. Multiply them all and you get 72.9. (If you want, put that to the power of one-fifth, and subtract 1 to get you to +1.36 "average" SD.) Someone who is +1 SD in each of the 5 categories will come out to an average of +1.00 average SD.
Something like that...
Have you done any sort of studies with the FSR data to see whether, for example, good hitters receive far better defensive ratings than poor hitters? Of course, maybe good hitters are better all-around athletes than poor hitters, in which case we'd expect to receive that result. But maybe if you compared the FSR defensive scores for those good hitters to their defensive ratings according to FRAA, for example (or some other system that's ignorant of the players' reputation or performance at the plate), you'd see some sort of bias show up. Knowing you, you've already looked into this and I've missed it.
Since Ben include footwork and accuracy, we're talking about "functional arm" as opposed to "arm strength". By that definition, David Eckstein at his peak could have been considered to have a plus "functional arm".
If readers don't like the answer, then they should change the definition!
***
Setting that aside, polls certain are subject to biases. And there are three, all of which can unfairly push Pujols up:
1. Halo effect
2. Past performance / memory (i.e., denial)
3. Positional
It's certainly possible how Pujols score of 57 in the arm strength category, where 50 is MLB average (neutral position) would be biased, and that it should maybe be a 40 or 45. Even so, he gets top marks in the other categories, so his "functional arm", as Ben defines it, would still seem to put him at least average.
But, like I said, sometimes it's hard to judge Pujols' arm accuracy, if he's only making 30 foot throws. Is Phil Simms a more accurate thrower than Warren Moon? Maybe on short passes? Put Jay Bruce or Evan Longoria at 1B, and who know, Pujols might out-accuracy them.
Also funny to see that he makes the cut for running. I wouldn't be surprised if he's the slowest runner on that list, in terms of actual foot speed.
No scout would give him passing grades for either tool, but he sure knows how to get the most out of what he has. Yet more proof of his amazing baseball IQ.
2) Would love to see the WORST tools list (those who had 1,000+ PAs last three years with the worst ratios as listed above).
3) Maybe Harold Richman would share Strat's "arm" ratings for OFs and Cs.
Also, as a Jays fan seeing Alex Rios so high on this list makes it clear that a couple more tools need to be added: work ethic and baseball intelligence.
As for your second point, I'm sure most scouts would agree that "makeup" would make a fine sixth tool.
As another Jays' fan, methinks that if Rios ran with his head like Rolen did, he would have garnered more in trade than simply salary relief (I know, he wasn't actually traded, but you get my point). Instead of having the 2nd worst BRR on this list, he would have been in the top 5 or so.
Second, there are obvious flaws with FRAA, but trying to compare the TOOLS of fielding when comparing different positions is problematic. Is Pujols REALLY have MUCH better fielding tools than Jose Reyes? I don't buy it.
What you can do instead is ask: what are the chances that this player really is above average in this category?
So, a player at -1 in baserunning runs might be at 40%, and someone at +1 in baserunning might be at 60%.
And you do that for every metric, and then multiply each of the 5.
So, Pujols might be 100% for average, and 100% for power, and 55% for baserunning and 55% for functional arm, and you multiply them all and say: he's 25% chance of being above average across all 5 tools.
Do this for everyone, and see what you get.