Prospecting

[View this in its original article form here and here.]

A common refrain heard from close watchers of the prospect landscape is that the mainstream sources for prospect information–essentially Baseball America and Baseball Prospectus, along with flashes from other places–is that the information and the process is still behind the curtain. The more honest of the third-party list-making crowd regret certain rankings or just change them as information from these two heavyweights is reported. Both sources have become more open with their process, BA even explaining their process in depth and including a “just missed” list and several supplements to their standard Top 100. Both sites have essentially a daily upload of the buzz they’re hearing into article form, some in near real-time, others with a delay until the relevant topic is discussed.

In our view, the information isn’t behind the curtain at all, people just want the information given in a different way.

Saber-Scouting is not a substitute for these essential sources, but we feel our different style of content (mechanical breakdowns, first-hand scouting reports, etc.) will make us a necessary supplement. We feel that we are in a unique position having team-side experience, scouting experience, statistical inclinations, industry contacts, and having followed the prospect landscape closely for years.

I can’t help but think four things (in this order) when I look over the historical BA Top 100 Lists.

1. Seriously, they ranked him there?

2. If I was in their position would I have done that, too?

3. This is pretty unfair in general, breaking apart an old prospect list.

4. But is there a better way to do this?

Personally, I’ve thought about these four things a whole lot more than a human should. And the answer to the most important question of this group, number four, is “I think so.”

That isn’t quite as powerful as a resounding “YES WE CAN,” but I prefer to under-perform and over-deliver.

So how do we propose the “better way” of ranking prospects? First of all, we don’t think it’s better, we just think it’s our way, but it will consist of more than just “trying harder” or an amorphous “do it better.”

There is a section of Malcolm Gladwell’s masterpiece “Blink” (widely read in front offices and coaching staffs, by the way) where a hospital that has a poor rate of correctly identifying and treating heart conditions wanted to improve their diagnosis success rate. What would you do to accomplish this? Bring in experts and complicated machinery and computer programs? This hospital brought in the heaviest of hitters: the spreadsheet.

They took information on each patient and their symptoms, then recorded them and figured out which three symptoms occurred most often in patients with actual heart issues. If you didn’t have at least two of these three symptoms, you weren’t a priority to treat, or were just sent home, told you were fine (because you probably were). They became the best in the world in this field, at quickly diagnosing heart conditions, with this spreadsheet and ensuing checklist the only changes made from when they were among the worst. Essentially, they solved a complex problem by breaking it down into it’s components and simplifying the decision process so that a monkey could do it. Should analyzing baseball really be more complicated than diagnosing a heart condition?We don’t technically have prospect list “secret sauce” (although we’re working on it) to rival BP writer Nate Silver’s Playoff Secret Sauce, but if there was one example I could give to describe our method, it would be that of Franklin Morales.

Let me start by saying that Morales is a fine prospect and has already contributed to winning a pennant—he’s already a fine MLB player and has a huge ceiling and nasty stuff. But if we were to, off the top of our heads, make a reverse-Secret Sauce for pitching prospects, he has the top 3 in spades.

1. He doesn’t have command of his pitches, and not in an effectively wild way, in a “make sure your affairs are in order before you come to the plate” way—granted half of that is a function of how hard he throws. The indications from scouts I’ve talked to and the video I’ve seen is that this is a fundamental problem that will never be completely fixed.

2. He doesn’t have a third pitch. For relievers, this doesn’t really matter at all, but Morales is a starter, is valued as such by the Rockies and is ranked as such by those who do the rankings. Some say he doesn’t need it because his first two pitches are that good, some say it will come in time. I say he’s destined for the bullpen sooner rather than later, and he’ll never have the feel to be an elite closer. Goldstein says he, “is still coming around to the fact that 94 mph with command and sink is a better pitch than 98 and wild.” Many pitchers never figure that out. Low-end projections include Alan Embree, high end as a left-handed Brad Lidge. As said above (and again below), that’s nothing to sneeze at, but I’m not describing why Morales is worthless, but how he’s a classic example of our method differing from other methods.

3. He doesn’t consistently repeat his mechanics or arm slot for multiple innings. This problem can come in different flavors and each have different consequences. A big side effect of this is problem is lack of command (stated above) but can go as far as injury problems (doesn’t appear Morales’ is that flavor, though). His spring his velocity was down, and while this could be due to any number of factors or even just starting to stretch his arm out, it’s the type of thing that will always happen in conjunction with these problems.

To sum it up, he lacks command, consistent mechanics, feel and a third pitch. All the other stuff is fantastic (two 70 pitches, left-handed, tall, some MLB success), endangered species rare, but if I’m betting on a guy reaching his ceiling, Morales is probably the last.

Now, I’ve never seen Morales in person, and am not too familiar with his makeup. More goes into our assessment of a player than just those factors mentioned. This isn’t meant to replace a scouting report, but as a case study in our method, using an extreme example.

Baseball America ranks him 8th and Kevin Goldstein has him at 13th. Having not gone through the proper process to rank a bunch of prospects yet, but just eyeballing comparables in the lists, I’d probably have Morales between 50 and 75.

As the last in the cavalcade of caveats, Morales could “figure it out,” with his mechanics, feel, command, changeup, consistency, approach, etc. and take off into the stratosphere as a starter and be everything everyone thinks he’ll be and more. And I’ll be wrong, wouldn’t be the first time. I’ll also be happy for him. He also would be securely in our top 100 prospects, so were we really that wrong? Knee-jerk groupthink prospecting would tell you yes, but that’s not what were trying to do here.

We aim to give you a fresh insight and methods and make you think a little differently about baseball.

[The section below is from part two]

I was having dinner with a friend that works in baseball and we somehow got on the topic of what type of prospect you would pick if given the choice by your GM in a trade. If you knew (how you know or if you could is another question to be answered later) a guy would be a quick impact guy then fizzle (where you could then trade him, since you know everything in this hypothetical) or a later-peaking All-Star type. I illustrated the question with recent players to make the question one based in some reality: Ben Grieve or Aramis Ramirez?

Don’t get too worked up about the players themselves, just the type of players they represent. Grieve is the early-peaking player with old man skills in his early 20′s (power, patience, corner position) that peaks in his age-22 rookie season and is a fringe starter by 25 and out of baseball at 29. Ramirez is a quick-to-the-majors type as well, but doesn’t peak until his age-26 season (still the normal time to peak, just later than Grieve for this example) and has one above-average year in his first 4 seasons, then breaks out in years 5 and 6, at the end of his team’s contract control, then will be paid the market rate (translation: over-paid) the rest of his career. Like I said before, forget the exact players to a degree, focus on the profile: early-peaking fringe All-Star, or normal-peaking perennial All-Star. Now, from a trade standpoint, if you know what their career holds, all things being equal, which do you take?

I argued, in a devil’s advocate sense, given that I knew what the other guy would say, for Ben Grieve. He puts up $10 million or more of productivity while making the minimum, and once he hits arbitration, you trade him to a less-savvy team (preferably run by Chuck LaMar) for another fringe-All Star with a better immediate future. Economically, that’s probably the right answer. The team official argued for Aramis Ramirez due to his job description. He said, regardless of knowing the future or not, his job in a trade is to identify the best player to trade for, and Aramis Ramirez became the best player. Looking back 10 years on a trade for the definitive winner and loser, picking Ramirez is more right. Both answers are right in their own way. There are also many more factors in play than the ones discussed, but this is just a hypothetical.

That was assuming that you know the future, which I’d assume many people do not. Is there a way to know, with some certainty, which players will reach their peaks at what time?

You can make some good guesses, based on how much of their value is in performance versus tools when they reach the majors, and what type of skills they have (old man versus young man). But there isn’t really a way to know with enough certainty for it to seriously impact, for instance, a trade negotiation. But I think there are some players that fall into the extremes that do allow you to take it into consideration for prospect ranking. (Pardon me as I beat a dead horse). Franklin Morales will not reach his peak soon. Most people can agree on this, we definitely think this. If he does what most people think he will, the first sign of doing it won’t come probably until his arbitration years (years 4-6) at best. Should this effect how we rank him? Depends on the criteria you’re using for ranking prospects.

Our criteria will take some of this (but not that much) into consideration. There’s a small population of players we can tell that will be early or late peakers, let’s say 10-15 of the top 100 prospects. Our criteria will be most valuable to a team, and that takes into account (to a small degree) how much financial gain the team can make as a result of owning this player. Many later-peakers still have great trade value even if they aren’t performing as expected, and having later-peakers for their first few free agent years is usually a net positive (it’s in their decline that you don’t want to pay the market rate), but normally with a trade, some value is lost in translation and all things being equal, you want the now performer, not the trade chip.

Again, this won’t make a huge difference in our rankings, I just want to point out yet another facet of prospect evalution.

So what’s the point of this whole section? Mostly to point out that we think about this stuff in more in-depth ways that it appears, and also that this is the view an economist would take to ranking prospects, and that opinion is becoming more and more important each day.

As for the technical way this type of thing would manifest itself, take a few minutes and familiarize yourself with our Scouting Tutorial. As an example for this section we’ll go with Reds prospect Drew Stubbs. Stubbs is regarded to have four of the five tools at least above-average (55 on the 20-80 scale), except the most important one, hitting. So how would having every tool except the most important one affect how the raw OFP goes into an adjusted OFP, the number that drives his grouping and prospect ranking? Stubbs’ raw OFP would be in the neighborhood of 60. That would make him one of the top 10-15 prospects in baseball, and if you think the bat plays, then that’s the place he should be (and could be with a huge season in the Florida State League this year—I’ll be keeping a close watch on this). But we currently don’t think the bat plays and Stubbs’ (if we had to project it now) will hit about .250 in the big leagues at his peaks (a 40 or 45 future bat grade). Still a useful and everyday player in the big leagues, but not a 60 adjusted OFP player.

So how would we adjust that OFP? We wouldn’t just say we don’t like him and make him a different OFP, we’d essentially do a weighting of the tools (as many already do for the raw OFP to avoid this problem) and then swing it a few points either way to get him where we think he should be. The raw OFP (once weighted) is a guide for where he should be, the adjusted OFP is making your opinion of him into one number, so adjusting this number from the raw is what you’re supposed to do to be more useful than a robot and get your opinion across—it isn’t cheating—but moving it more than 2-3 points without a good reason is probably too extreme.

For the sake of the excercise, the 60 raw number becomes 54.5 and I’d actually bump that up a few a notch to 55 or 56, the B- to C+ area, basically slightly above-average everyday to above-average evverday player. As is, we think he’s a .250/.340/.450 type centerfielder with great defense and speed, that’s slightly-above average. But I’m also taking into account that he’s got the upside to be more, and we’ll denote these types of players in our team prospect lists. An overachieving college center fielder with a 55 grade that’s in AAA and has no upside and Stubbs’ at 55 with huge upside and some downside are different players and we’ll put a flag, or a different color text or something to indicate that difference, although the 55 number already takes some of that into account.

Another way that we’ll try to avoid huge busts aided by inaccurate raw OFPs leading us down the wrong road is an amateur hitting adjustment, and breaking hitting and power into components to pinpoint the type of player.

For amateurs, the present hitting grade is usually a throw-away for scouts. Is high school hitter X a current 20, 25, or 30 hitter? It’s impossible to know and makes no difference, the comments that are next to it are much more valuable. So, taking a cue from a team that uses this approach, we’ll use a peer grade for the current hit grade for amateur hitters. Basically, the 20-80 score the hitter’s bat has against his peers (say players of the same age taken in the same round). And the rule is that the future grade can be anywhere below that present grade, but can’t be more than 10 points above it. That way, players that can’t hit in HS (say, a 40 peer grade) can’t be projected to hit in the big leagues for more than average (in this case, no higher than 50 future grade).

Hitting, as many scouts and analysts already do, will be broken down (even if we don’t spell it out as such in the scouting reports) into plate discipline and hitting ability. Power will be broken down into raw power and power frequency. These distinctions are for how far a player can hit the ball, and how often he taps into that ability. This would be evident in players like Juan Uribe that can’t hit the bar especially far, but seem to tap into that ability all the time. Or, with players that have plus power but don’t have an approach that taps into it often, like Matt Murton.

This type of component approach is done for other tools, like breaking a fastball down into velocity, movement, and command, but those types of distinctions are more well-known, so we assumed if you’ve read that far you’re aware of it.