How Much Risk Is Worth Additional Upside?
Keith Law’s list of the Top 100 prospects in baseball came out today, though you have to be an ESPN Insider subscriber in order to see the rankings. Since the content is behind a paywall, I’m not going to give away too many of the rankings, but there were a few things on the list that caused me to do some thinking, and those thoughts inspired this post.
The first thing I noticed was how heavy the list was on pitchers. There’s only two pitching prospects in the top nine, but then 12 of the next 16 spots go to hurlers, and overall, 49 of the 100 spots on the list are occupied by pitchers. It’s not exactly breaking news that young pitching prospects get hurt and flame out at rates much higher than comparable hitting prospects, so in order to compensate for the extra risk they bring to the table, their placement has to be justified through additional upside. Keith’s a smart guy and understands all this, and I’m sure he’d be able to make a valid argument that each pitcher on the list has enough potential to justify their ranking even with the understanding of greater risk.
This ties into the other thing I noticed while perusing the list, however – the top prospects, in most cases, have very little minor league experience. Bryce Harper, Manny Machado, and Jurickson Profar all rate in the top 7 after spending last year in low-A ball, and while they’re all big time talents, they’re not really the exception. Dylan Bundy and Bubba Starling are ranked highly despite having never played a professional baseball game, while Archie Bradley has them beat by just two innings in rookie ball last summer. With the exception of Harper – a special talent and unique prospect in many regards – these guys are all several years from the big leagues. While their physical talents are real, we also have less information about them than some older prospects who have spent a few years climbing the ladder. The less you know about a player, the higher your uncertainty around that player’s value has to be. Lack of knowledge translates to risk, and risk weighs down value.
I don’t know enough about prospects to say that Law’s rankings of these very young kids are too aggressive – he knows their upside better than I do, and I’m sure he spent a lot of time deciding how to balance the increased risk that goes along with players we know relatively little about. But, I do wonder if prospect valuation in general focuses too much on a player’s perceived potential and not enough on minimizing the chances of a player being a bust who never contributes a meaningful thing to a big league club.
For example, Yonder Alonso is the only 24-year-old on Law’s list this year, and he comes in at #69. Alonso’s been in the spotlight for a while, and people have had plenty of time to break down his weaknesses. We know he’s not a great defender, he doesn’t pull the ball with authority all that regularly, and he projects to be more of a doubles hitter than a home run guy despite playing a position where power is essentially mandatory. The three years Alonso has spent in the minor leagues have given us a lot of information about his abilities, and not all of that information has helped his stock as a prospect. We’ve had time to identify his flaws, and he’s had time to show us that they’re not things he can easily improve upon, or at least he hasn’t been able to as of yet. It’s hard to imagine a 24-year-old suddenly learning how to rip the baseball over the fence in the same way that you can with a raw 21-year-old. Alonso’s stagnated power development has lowered our expectations of his future potential, and in turn, has hurt his stock as a prospect.
However, there’s a flip side to this coin – we also know that Alonso’s contact skills are good enough to translate against high level pitching, that his approach at the plate is basically Major League ready, and that there’s enough juice in his bat that he can drive the ball the other way with regularity. His performance against both Triple-A and Major League pitching last year solidified our understanding of his strengths as well, and our confidence in our ability to project his future is higher than with pretty much any other prospect on the list. He’s the guy we know the most about, and that reduces risk, which in turn should raise his value.
However, I know it’s tough to get excited about a meh defensive first baseman with average power. What we’ve learned about Alonso leads us to think that he’s probably going to be a +2 to +3 win player, and he has limited star potential unless there’s an unexpected power surge coming. A scout I talked to recently compared Alonso to Wally Joyner with a bit less defensive value. Looking at their minor league numbers, the comparison fits pretty well.
| Â | BA | OBP | SLG | OPS | BB% | K% | ISO |
|---|---|---|---|---|---|---|---|
| Joyner | Â | Â | Â | Â | Â | Â | Â |
| AAA (1985) | 0.283 | 0.363 | 0.440 | 0.807 | 11.0% | 11.7% | 0.157 |
| 1985 PCL Average | 0.272 | 0.343 | 0.403 | 0.746 | 9.5% | 15.1% | 0.131 |
| % Better Than Average | 4% | 6% | 9% | 8% | 15% | 23% | 20% |
| Alonso | Â | Â | Â | Â | Â | Â | Â |
| AAA (2010/2011) | 0.296 | 0.364 | 0.478 | 0.842 | 9.7% | 15.9% | 0.182 |
| 2010/2011 IL Average | 0.262 | 0.330 | 0.405 | 0.735 | 8.4% | 19.3% | 0.143 |
| % Better Than Average | 13% | 10% | 18% | 15% | 15% | 17% | 27% |
The comparison isn’t perfect, as Joyner spent his full age 24 season in the Majors, so we’re comparing his just age 23 season to Alonso’s 23/24 seasons, but it’s not an issue that eliminates the usefulness of the comparison, especially considering how well Alonso hit in the big leagues when he was promoted.
Now, Joyner wasn’t exactly a star in his own time. He made one all-star team in 16 seasons and only finished in the top 20 in MVP voting twice. His best season was +4.2 WAR, and for his career, he averaged +2.9 WAR per 600 PA. And that was with some pretty nifty glovework – take his defensive value away, and you’re looking at a guy who was more like a +2.5 win player in most years. The Joyner comparison wasn’t meant to be a flattering one to Alonso – the scout was trying to imply how limited his upside actually was.
But, let’s keep in mind a realistic baseline for most prospects. Joyner produced +40.4 WAR over 8,000 big league plate appearances. He had seven seasons where he finished with +3 WAR or higher, and he posted a wRC+ of 110 or better in every season from 1986 to 1998. Yes, he had limited power, only averaging 15 home runs per 600 plate appearances over his career, but his ability to make contact, draw walks, and rack up doubles made him a pretty nice player for a long time.
Maybe Joyner is something close to the best case outcome for this kind of player type. After all, he stayed extremely healthy and hit better in the Majors than his minor league numbers suggested was likely, so rather than relying on just one comparison, I grabbed all first baseman over the last 30 years who have racked up 2,000+ PA in the big leagues and have posted an ISO between .130 and .170. This gives 45 players ranging from Greg Norton to John Olerud. The median offensive line from those 45 careers? 3,765 PA, .277/.344/.421, 105 wRC+, +10.5 WAR.
Joyner had the fourth best career of the bunch, coming in behind Olerud, Mark Grace, and Don Mattingly, all of whom were much better defenders than Alonso is. Perhaps a more realistic comparison for Alonso is Jeff Conine, who hit .285/.347/.443, good for a 107 wRC+. He hung around long enough to post +24.3 WAR. Even if we think he’s just going to turn into James Loney (+8.6 WAR to date) or Lyle Overbay (+13.0 WAR), there’s still some real expected production to be had, and there’s obviously upside beyond those guys.
I’m not suggesting that Alonso is a top 10 prospect in the game. I fully acknowledge that teams should and do prefer a reduced chance at getting a premium player over a safer guy who probably tops out as a solid everyday guy. But, it’s all a balancing act – you will trade some security for greater upside, but you shouldn’t just be willing to accept an unlimited amount of additional risk in pursuit of slightly higher potential returns. And, in looking at Law’s Top 100 list this morning, I just wonder if we’re getting that balance right. How many of these 18 or 19 year old pitching prospects are going to put up +10 WAR in the big leagues? History tells us most of these guys will never make it, and the ones that do will still likely fail to live up to expectations.
Is a 10% chance of becoming a quality starting pitcher for a few years before your arm starts to hurt really better than a 50% chance at being a league average player, especially when you can start producing that value immediately while the upside play requires a few years of deferred value? Have we just focused too heavily on upside when it comes to prospects, to the point where we’re now valuing lottery tickets over relatively sure things?
I can’t say that I have the answer to these questions right now, but I think they’re at least worth asking.
2
Another reason lists could be so pitcher heavy is that apparently pitcher, for some analysts, are easier to project. Sickles just did a few articles on this. Top pitchers, almost without exception, he rated as A or B+ prospects. The position players were all over the place, with several good position players being B- or C level prospects.
Just food for thought.
Interesting. So the theory would be that it’s just more common for a “low upside” position player (say, Ben Zobrist) to develop into a star player out of nowhere than it is for a guy throwing 87 to suddenly start throwing 95 and turn into an ace. Basically, we can easily identify guys like Felix and Verlander as high-reward guys at a young age, and then just have to let attrition sort through who is going to make it or not, while with position players, we might not be able to spot elite hitting skills at a young age.
This makes a decent amount of sense. Sounds like an issue that needs more research, though.
Exactly.
Given what we think we know about pitch speed and pitch type effectiveness, it does seem like we think we can project them better than an 18 year old OF who can run to first in 4.2 seconds. In addition, if you want to look good at making these types of list, if you have more pitchers they may be more likely to fall off the list due to injury than skills. so when we look back in 3 – 5 years at this list it won’t be because Law is bad at projecting the skills, just that injuries caused a fair amount of these guys not to achieve their potential.
Also, where did Law have Alonso ranked the past couple years?
A quick look I did on how BA has done better at predicting the best pitchers then hitters:
http://www.beyondtheboxscore.com/2011/4/6/2094227/2010-batting-leaders-on-ba-top-100-prospect-list
http://www.beyondtheboxscore.com/2011/4/14/2111317/2010-pitching-leaders-on-ba-top-100-prospect-list
First, the OP summary of Sickels’ rankings are simply wrong. (see my post below and ignore my grammatical error).
Second, there had been research that shows top ranked hitters succeed at a great rate than do top-ranked pitchers. That would suggest that it’s easier to project hitters.
t — see Michael Pineda, whom you know quite well, as a pitcher
It’s a neat theory, but is it really right?
I mean, it seems to me like there are tons of “live arm” hard throwers in the minors who are pegged as future relievers for one reason or another.
Cliff Lee is a great example of a guy never made anyone’s list of future aces. And, indeed, anyone who called Cliff Lee a future Cy Young in 2005 would rightly have been laughed at. But it happens.
Plus, think of all the Homer Baileys: guys who everyone thought was an ace-unless-injured, who actually didn’t turn into an ace.
without going further, i would be careful in quoting sickels on anything :-)
Why?
You summary of Sickels’ ranking are simply wrong.
Sickels publishes top 50 lists for both hitters and pitchers. This year 38 of 50 pitchers were given a grade of B+ or higher and 35 hitters. I don’t have the past year books in front of me but I believe in some years there were more hitters than pitchers at B+ ot better.
I’m guessing you don’t buy his book. You should.
John Sickels was quoted saying he “was probably better at analyzing/ranking pitchers than hitters” in a monster thread at his site.
I’m guessing you didn’t read his article. You should.
-C
He’s talking about top 50 pitchers and hitters currently in the Majors. Here’s his posts on the two:
Pitchers
Hitters
My post was in response to a false claim by “Hunter fan says” who wrote that Sickels gives many more top grade to pitchers than he does to hitters. That is simply wrong. This year they were nearly even, in 2010 there were more top grades for hitters, and I believe in 2011 there were more top grades for hitters too. In generalm Sickels gives many more top grades to hitters, about 50% more, than he does to pitchers, and by top grades I mean A or A-.
What Sickels wrote in his recent thread was this:
“I used to think that position players were more predictable than pitchers, but over the last few years my thinking on that has changed. Nowadays I am more confident in our ability to judge pitchers. ABSENT INJURIES, I think it is easier to judge pitchers both traditionally (stuff, mechanics, projection, etc) and sabermetrically.”
I added the emphasis to ABSENT INJURIES because that is a huge caveat, the subject of Cameron’s thread and yet cthatbeerman managed to read the article and miss it entirely.
Rotofan: no, that’s not at all what Hunter said.
You raise some very interesting points. Still, I think a large part of the reason why we see so many low-minors upside guys in Law’s rankings is simply the state that baseball happens to be in right now. Law himself has acknowledged that most of the most exciting talent is in the lower minors this year, to an unusual extent. I certainly think your questions about the valuation of low-ceiling, high-floor guys are worth asking, I think in most years the top of lists like these won’t be so flooded with guys with little professional experience.
Law also mentioned at the top that he adjusted his rankings to favor upside in accordance with how front offices currently value prospects. So it seems to be an industry thing, not a Law thing.
…illustrating yet again why AA is absolutely murdering it. the world is zigging–valuing low A prospects too highly–and he zags by valuing ML-ready prospects/players (santos, lawrie) instead.
That’s one lengthy follow-up on my previous comment regarding the Joyner/Alonso comparison. Thanks.
As far as the question at the end of the article is concerned: I would imagine team composition would play a part, so making an aggregate ‘Top 100′ list without either contextualizing for the team the prospect is on, a direct study of what’s more valuable to have will probably have some loose ends.
Ultimately it’s all about aligning enough WAR on the same roster at the same time. In that vein, it probably makes more sense to have two high risk/rewards and one lesser sure-thing than it does to have three high risk/rewards as the odds of aligning the WAR on the roster at the same time is easier optimized. But, again, it all depends on the baseline WAR for the roster in which these players will be placed, bringing me back to my previous point.
My grammar above replacement was negative in that second paragraph. Should read:
As far as the question at the end of the article is concerned: I would imagine team composition would play a part, so making a study on the proper way to assemble ‘Top 100? lists without contextualizing for the team the prospect is on will probably always have some loose ends.
This is a topic I’m surprised not to see discussed more in the sabermetric community. There’s been a ton of work done in mathematical finance on how to price options and the value of risk (start with Black-Scholes and work your way in from there) and while many of the assumptions aren’t directly parallel, it wouldn’t be too complex to build new related mathematical models that more accurately reflect the ‘physics’ of baseball value and risk rather than financial markets. We have more than enough data on how prospects pan out by this point that we should be able to assign reasonable probabilities to most players (or at least most ‘player types’) for model-building purposes…
Basically the problem is that you do not have enough information on the lower level or prospects without any professional information. Because you do not have any statistical information, all of your information is based on scouts, and upper first-round picks tend to get glowing reports from scouts, with nothing to balance it out. Once we see Bubba Sterling strike out in 32% of his at bats next year, or Bundy walk 6.8 batters per 9, their rankings will fall.
BA tends to put every first round pick from the previous draft in their Top 100, but simple math tells you that if the average first round minor league career is longer than 3 years, they’re not all going to stay on that list.
I have to agree with the sentiment – why so much emphasis on “upside”, but not on “downside”?
I was thinking this earlier today when reading the article on Bauer and Bradley. Sounded to me like Bauer is sooo much more likely to pan out as a premium SP, and is already ready. Bradley is 19, developing, and isn’t close to ready. Sounded like he could end up in the bullpen.
(Bradley ranked higher than Bauer on Law’s list)
Isn’t downside always the same? That is, the downside for every prospect is that they’ll flame out and never make it to the big leagues. What changes is their likelihood of hitting their downside vs. the likelihood they hit their upside.
The real downside is if they turn into Neifi Perez.
Keith Law did mention on Baseball Today that this list is different from other years in that there are a lot of high-ceiling, high-risk young players since the top levels of the minor leagues are fairly thin.
Also, one prospect list is a small sample size.
This is a key point… there is a lot of “new, high-ceiling talent” in the minors… most of the big ticket guys are new to the minors over the last few years because so many teams are now going to younger, cheaper in-house options… In the past two years we’ve seen what seems like a lot more top prospects graduate to big leaguers than in recent memory (although I have not done a study to confirm this but it might be interesting to look into it…). Once a player hits double-A and proves himself the clubs are rather quick to get them to the Majors and a lot of orgs seem to view triple-A seasoning as less important these days.
Most lists this year will be topped with lots of high risk / high ceiling prospects. I always figured that much of this has to do with the “state” of the minor leagues. We’ve seen strings of young prospect graduations for several years running now, and there haven’t been many older visibly projectable prospects in the minors lately because of it…
Many of these “ranking gurus” use very different sets of “criteria” from list to list and even year to year. Also, many claim to express how they assemble these rankings, but they never actually go into detail to describe what the system is, and too often they’ll swerve outside that criteria to fulfill a hunch let’s say. So in many respects, it’s impossible to critique what they assemble, because we don’t know how to accurately relate to their judgements, which is fine, but it’s not helpful for any kind of analysis.
This article is also based on Keith Law and not, say… Baseball America. Anytime a guy says he “usually uses his own scouting reports”, but will talk to “others as needed”, I’ll look the other way. When you consider the monster that is Baseball America, and the amount of resources it takes for them to be as thorough as they are… I don’t buy it Keith Law… It’s gravy without the potatoes.
With exception to a few sources, I think lists nowadays are being influenced more and more by fantasy baseballing.
I think the view in some front offices – and it aligns with Law’s rankings – is that stars are exceptionally hard to acquire unless you develop them yourself. You have to drasctically overpay to sign a star free agent, while mid-tier guys can often be had at reasonable prices. So, when it comes to prospects you’d rather have a few guys that might be a star rather than a bunch of guys who are low-upside/high-probability, because it’s easier to fill in a few gaps with available veterans than it is to find the stars to build around.
As for pitchers being easier to assess I think the opposite is true. You find more pitchers coming out of nowhere to be top prospects than you do with hitters, they’re just generally more volatile. I’d bet if you reviewed the top 100 lists (whether Law or BA) there’d be more pitchers that were drafted in low rounds (Matt Moore) or were low budget internat’l signs (Martin Perez) than you’ll find position players who came from nowhere.
Exactly. Guys like Overbay, Joyner, Loney are not hard to acquire. Developing them has some value, but not the kind of value that can’t be easily replicated in other ways.
As I see it, the reason why, as you put it “lottery tickets are valued above relatively sure things” is because when the lottery ticket hits, he is way more valuable than the sure thing. The star is always worth more than the average player. While nobody who makes prospect lists assumes that all players will reach their ultimate peak upside, the lists themselves assume they do.
An entire franchise can be built around luckily timed lottery hits. See Philadelphia and Rollins, Utley, Howard, Victorino, Werth, and Ruiz. They all had varying amounts of risk, but if the Phillies had Wally Joyner quality players (let’s try to avoid the Howard debate here) instead, they would not be the powerhouse they are now.
I’m really surprised that this doesn’t come up more often when we discuss prospect rankings.
Everyone talks about upside all the time because nobody want to be the guy who called a future Hall of Famer the 102nd best prospect in baseball. This goes double when the guy was a high draft pick.
But for many teams in win now mode, a sub-All Star talent who is major league ready is worth much more than a potential ace in short season ball.
I’d love to see the sabermetric community develop a matrix for prospect evaluation that includes a sliding scale based on the organization’s overall development.
Use the Phillies as an example.
They have Halladay, Lee, Papelbon, Rollins, and Howard signed to contracts paying about 60% of their current payroll in 2014. If prospect X can’t be ready to play in 2014 (providing budget wiggle room by making the league minimum), then prospect X must be a less valuable prospect to the Phillies than he would be in another organization whose window is still opening.
So we start by grading the Phillies based on age and perceived budget constraints, and we find they are a type 3 organization. prospect X gets a C grade on the Phillies (because he won’t be ready while their window is at its peak) but if traded to the Royals (a type 2 organization whose window is just opening) prospect X gets a B grade.
I really enjoyed this article because it’s very thought provoking. One thing I have noticed is that high ceiling players get too much attention to the detriment of perceived low floor guys and that’s a shame because scouting is a very art. When ever I hear that reports say prospect xxxx has a limited upside I cringe a little. The examples of supposed low ceiling guys include Kevin Youkilis and Edgar Martinez and those two are very good. Also, Albert Pujols was a 13th round pick and probably was initially perceived as a “C”prospect (i’m guessing this) and yet, he is one of the greatest players ever.
I think we forget that many of these high ceiling players become the mid talent guys we are talking about. It’s really hard to determine which low ceiling players have the ability to succeed. We don’t give much attention to these types of players because there are so many that fit into this category, and it’s much less likely to guess which mid level prospects will be able to succeed. Your rate of high ceiling successes are still far greater.
Your chances of succeeding with Dylan Bundy are still much greater than say Clayton Richard. A Clayton type is the lottery ticket, not Dylan. Having a Dylan type is like having maybe a few thousand lottery tickets in comparison.
Show me the guy who can identify the Clayton Richards of the baseball world as being more future productive than the Phil Hughes types before they actually get tested and you have a miracle worker.
Also, Yonder has always had “high ceiling/significant upside” aspects. Using him as part of an argument like this is too easy. Even while he was still only in Miami, you could guess that he and Billy Butler might be similarly projectable.
I know there is not much difference between #5 and #7 but I don’t quite understand the rationale to have Profar behind Machado? Profar is younger, had a better 2011 and by most accounts has a better shot sticking at SS… Just a minor quibble…
I’d assume that for Law the difference between how their respective bats project is large enough to outweigh Profar’s better defense, slightly better 2011 production, and youth. Also, a fair amount of Machado’s 2011 was spent in high-A, where he didn’t exactly embarrass himself; Profar spent the whole year in low-A.
I was thinking about a similar topic a while back, and I think we need probability distributions for prospects. They would obviously be somewhat subjective, but right now when someone says “high upside” or “high risk”, it’s only a vague notion. If you said something like “Player X has a 50% chance of peaking at 2 WAR/season, and a 10% chance of peaking at 4 WAR/season” we would have a better basis for discussing his value. Prospect rankings should then be correlated with the expected value of the WAR/season distribution, perhaps weighted by a function that skews toward the high upside end of the spectrum to account for the rarity of such players.
The problem is obviously with coming up with numbers to put into the distributions, but as Dave discusses, the distance from the Majors would fit into the equation by flattening the distribution and shifting it toward 0 WAR. Draft pedigree, MLEs where available, and scouting reports would help to establish the player’s ceiling, which might be defined as the 80th percentile of his distribution.
Perhaps something like this isn’t feasible for baseball prospects, but for a field that is so probabilistic in nature it seems like a good approach for quantifying what is now a vague concept.
I don’t remember the details, but Tango linked to a study on BA’s Top 100 awhile back. The result of that study was that once you get past the top 20, there’s little difference in the rankings. The average WAR produced by those guys is less than 1/ year during their first 6 seasons.
I guess the better question would be why would anyone PAY for Keith Law’s “insider” when they can get better analysis and information for free here?
I’ve been chewing on this overnight and came up with the following thought experiment: let’s say you had a team with 10 3-win players, all of whom are going into their third year in the majors. How many of those players would you be willing to trade for an 6-win player going into his third year?
I think my answer is certainly 3 and probably 4 – I figure these players can be replaced pretty easily and cheaply with 1-win players. If I trade 3 players, then, I lose 9 wins, add 6 wins with the player I’m trading for and 2 more wins with the 1-win players I add to replace the other two positions. I end up at -1 wins. If I go to 4 players I end up at -3 wins.
So why would I be willing to do this? A team with 10 3-win players has almost no easy upgrades – by replacing two or three of them with 1-win players, I suddenly have clear places where I can upgrade talent. Furthermore, it is very hard to bring up young talent on a team with so many solid players – a solid rookie might put up a 1-win first season but end up a reliable 2- or 3-win player for many years. You’re not going to bring that rookie up if you already have a 3-win player at his position.
So bringing this back to the question of this article: if I have a prospect who I think has a 1/4 to 1/3 chance of becoming a 6-win player, I value him about the same as a player who I think is clearly a 3-win player. Of course, prospect valuations are never so precise…
It’s interesting that despite Law’s focus on upside players, he placed the Padres atop his organizational rankings. San Diego has very little in the way of premium/five-star prospects, but the system is packed (like 15 deep) in guys who will probably have meaningful big-league careers. Apparently this netted out (in Law’s eyes, anyway) as a more desirable system than any other team’s, irrespective of impact potential.
I’m not sure if many scouts have adjusted to the post steroid era yet. The offensive environment is so different now, and I get the feeling that many fans and scouts are having a hard time wrapping their heads around what a premium offensive prospect really is anymore.
Great article, and, of course, I totally agree with it. It just seems like too much weight is given to the potential and less to the lower risk guys. It just seems like with the high risk/high potential guys, they forget that they most likely won’t turn out. In business, their expected value (value X probability) is probably much lower than a player like Alonso, who is so close to the majors.
The Minor League Baseball Analyst Annual book tackled this conundrum many years back by having a two character rating system. The first was a number showing what that player’s potential appears to be at that point in time (HOF, AS, regular, utility), the second was a letter showing what that player’s closeness to the majors is. I like that system.
I have seen other places go with the multi-tiered rating: describe his best or perfect world projection, plus what he’ll likely become.
Now, this is probably not the intention, but I have seen some places provide a Top 100 Prospect list along with a Top 100 Fantasy list, and that works to your point in the article, the former is more about potential and the latter more about proximity, giving a taste for each.