How Do Prospect Grades Translate to Future Outcomes?

Reggie Hildred-USA TODAY Sports

Hello, and welcome to Prospect Week! (Well, closer to Prospect Fortnight — as you can probably tell from the navigation widget above, the fun will continue well into next week, including the launch of our Top 100.) I’m not your regular host – that’d be Eric Longenhagen – but not to worry, you’ll get all the Eric you can handle as he and the team break down all things minor leagues, college baseball, and MLB draft. I’m just here to set the stage, and in support of that goal, I have some research to present on prospect grades and eventual major league equivalency.

When reading coverage of the minor leagues, I often find myself wondering what it all means. The Future Value scale does a great job of capturing the essence of a prospect in a single number, but it doesn’t translate neatly to what you see when you watch a big league game. Craig Edwards previously investigated how prospect grades have translated into surplus value, but I wanted to update things from an on-field value perspective. Rather than look at what it would cost to replace prospect production in free agency, I decided to measure the distribution in potential outcomes at each Future Value tier.

To do that, I first gathered my data. I took our prospect lists from four seasons, 2019-22, and looked at all of the prospects with a grade of 45 FV or higher. I separated them into two groups — hitters and pitchers — then took projections for every player in baseball three years down the line. For example, I paired the 2019 prospect list with 2022 projections and the 2022 prospect list with 2025 projections. In this way, I came up with a future expectation for each player.

I chose to use projections for one key reason: They let us get to an answer more quickly. In Craig’s previous study, he looked at results over the next nine years of major league play. I don’t have that kind of time – I’m trying to use recent prospect grades to get at the way our team analyzes the game today. If I used that methodology, the last year of prospect lists I could use would be 2015, in Kiley McDaniel’s first term as FanGraphs’ prospect analyst.

Another benefit of using projections is that they’re naturally resistant to the sample-size-related issues that always crop up in exercises like this. A few injuries, one weird season, a relatively small prospect cohort, and you could be looking at some strange results. Should we knock a prospect if his playing time got blocked, or if his team gamed his service time? I don’t think so, and projections let us ignore all that. I normalized all batters to a 600 plate appearance projection and all pitchers to a 200 innings pitched projection.

I decided to break future outcomes down into tiers. More specifically, I grouped WAR outcomes as follows. I counted everything below 0.5 WAR per season as a “washout,” including those players who didn’t have major league projections three years later. Given that we project pretty much everyone, that’s mostly players who had either officially retired or never appeared in full-season ball. I graded results between 0.5 and 1.5 WAR as “backup.” I classified seasons between 1.5 and 2.5 WAR as “regular,” as in a major league regular. Finally, 2.5-4 WAR merited an “above average” mark, while 4-plus WAR got a grade of “star.” You could set these breakpoints differently without too much argument from me; they’re just a convenient way of showing the distribution. There’s nothing particularly magical about the cutoff lines, but you have to pick something to display the data, and a simple average of WAR projections probably isn’t right.

With that said, let’s get to the results. My sample included 685 hitters from 45-80 FV. Allowing for some noise at the top end due to small sample size, the distribution looks exactly like you’d hope:

Hitter Outcome Likelihood by FV
FV Washed Out Backup Regular Above Average Star Count
45 51% 25% 17% 6% 1% 295
45+ 52% 18% 19% 11% 1% 91
50 23% 24% 30% 21% 2% 197
55 17% 17% 30% 31% 6% 54
60 14% 12% 19% 38% 17% 42
65 0% 33% 33% 0% 33% 3
70 0% 0% 0% 0% 100% 2
80 0% 0% 0% 0% 100% 1
Note: Projections from three years after the player appeared on a prospect list

Consider the 55 FV line for an explanation. Of the players we graded as 55 FV prospects, 17% look washed three years later – Jeter Downs, a 2020 55 FV, for example. Another 17% have proven to be backup-caliber, like 2022 55 FV Curtis Mead, or 2019 55 FV Taylor Trammell if you don’t think Mead’s trajectory is set just yet. Continuing down the line, 30% look like big league regulars – 2021 55 FV Alek Thomas, perhaps. A full 31% appear to be above-average major league contributors three years later, like 2019 55 FV Sean Murphy or 2021 55 FV Royce Lewis. Finally, 6% project as stars three years later – Jackson Merrill, a 55 FV in 2022, feels appropriate as an example.

Two things immediately jump out to me when looking at this data. First, the “above average” and “star” columns increase at every tier break, and the “washout” column decreases at every tier break. In other words, the better a player’s grade, the more likely they are to be excellent, while the worse their grade, the more likely they are to bust. That’s a great sign for the reliability of our grades; they’re doing what they purport to do, essentially.

Second, each row feels logically consistent. The 45 FV prospects are most likely to bust, next-most-likely to end up as backups, and so on. The 45+ FVs look like the 45 FVs, only with a better top end; their chances of ending up above average are meaningfully better. The 50 FVs are a grab bag; their outcomes vary widely, and plenty of those outcomes involve being a viable major leaguer. By the time you hit the 55 and 60 FV prospects, you’re looking at players who end up as above-average contributors a lot of the time. The gap between 55 and 60 seems clear, too; the 60 FVs are far more likely to turn into stars, more or less. Finally, there are only six data points above 60 FV, so that’s mostly a stab in the dark.

This outcome pleases me greatly. Looking at that chart correlates strongly with how I already perceived the grades. For a refresher, roughly 30 prospects in a given year grade out as a 55 FV or above, give or take a few. Something like three quarters of those tend to be hitters. That means that in a given year, 20-ish prospects look like good bets to deliver average-regular-or-better performance. The rest of the Top 100? They’re riskier, with a greater chance of ending up in a part-time role and a meaningfully lower chance of becoming a star. But don’t mistake likelihood for certainty – plenty of 55 and 60 FVs still end up at or below replacement level, and 45 FVs turn into stars sometimes. Projecting prospect performance is hard!

How should you use this table? I like to think of Future Value in terms of outcome distributions, and I think that this does a good job of it. Should a team prefer to receive two 50 FV prospects in a trade, or a 55 FV and a 45 FV? You can add up the outcome distributions and get an idea of what each combination of prospects looks like. Here are the summed probabilities of those two groups:

Two Similar Sets of Prospects, Grouped
Group Washed Out Backup Regular Above Average Star
Two 50 FVs 46% 49% 60% 42% 4%
One 55, One 45 68% 42% 47% 37% 6%

Another way of saying that: If you go with the two-player package that has the 55 and 45 FV prospects, you’re looking at a higher chance of developing a star. You’re also looking at a greater chance of ending up with at least one complete miss, and therefore lower odds of ending up with two contributors. Adding isn’t exactly the right way to handle this, but it’s a good shorthand for quick comparisons. If you want to get more in depth, I built this little calculator, which lets you answer a simple question: For a given set of prospects, what are the odds of ending up with at least X major leaguers of Y quality or better? You can make a copy of this sheet, define X and Y for yourself, and get an answer. In our case, the odds of ending up with at least one above-average player (or better) are 40.7% for the two 50s and 41.4% for the 45/55 split. The odds of ending up with two players who are at least big league regulars? That’d be 28.1% for the two 50 FVs, and 16.1% for the 45/55 pairing. Odds of at least one star? That’s 4% for the two 50 FVs and 6% for the 45/55 group. In other words, the total value is similar, but the shape is meaningfully different.

For example, you’d have to add together a ton of 50 FV prospects to get as high of a chance of finding a star as you would from one 60 FV. On the other hand, if you have three 50 FVs, the odds of ending up with at least a solid contributor are quite high. Meanwhile, even 60 FV prospects end up as backups or worse around a quarter of the time. That description of the relative risks and rewards makes more sense to me than converting players into some nebulous surplus value. Prospects are all about possibility, so representing them that way tracks analytically for me.

Take another look at the beautiful cascade of probabilities in that table of outcomes for hitting prospects, because we’re about to get meaningfully less pretty. Let’s talk about pitching prospects. Here, the outcomes are less predictable:

Pitcher Outcome Likelihood by FV
FV Washed Out Backup Regular Above Average Star Count
45 53% 26% 16% 5% 0% 230
45+ 38% 24% 25% 13% 0% 68
50 27% 27% 24% 20% 2% 96
55 17% 20% 37% 27% 0% 30
60 17% 33% 25% 25% 0% 12
65 0% 0% 0% 100% 0% 1
70 0% 0% 100% 0% 0% 1
Note: Projections from three years after the player appeared on a prospect list

I have tons of takeaways here. First, there are substantially fewer pitching prospects ranked, particularly as 50 FVs and above. Clearly, that’s a good decision by the prospect team, because even the highest-ranked pitchers turn into backups at a reasonable clip. Pitching prospects just turn into major league pitchers in a less predictable way, or so it would appear from the data.

Second, there are fewer stars among the pitchers than the hitters. That’s true if you look at 2025 projections, too. There are only six pitchers projected for 4 WAR or higher, while 42 hitters meet that cutoff. It’s also true if you look at the results on the field in 2024; 36 hitters and 12 pitchers (22 by RA9-WAR) eclipsed the four-win mark. You should feel free to apply some modifiers to your view of pitcher value if you think that WAR treats them differently than hitters, but within the framework, the relative paucity of truly outstanding outcomes is noticeable.

Another thing worth mentioning here is that pitchers don’t develop the same way that hitters do. Sometimes one new pitch or an offseason of velocity training leads to a sudden change in talent level in a way that just doesn’t happen as frequently with hitters. Tarik Skubal was unmemorable in his major league debut (29 starts, with a 4.34 ERA and 5.09 FIP). Then he made just 36 (very good) starts over the next two years due to injuries. Then he was the best pitcher in baseball in 2024. Good luck projecting that trajectory. Perhaps three-year-out windows of pitcher performance just aren’t enough thanks to the way they continue to develop even after reaching the majors.

There’s one other limitation of measuring pitchers this way: I don’t have a good method for dealing with the differential between reliever and starter valuation. Normalizing relievers to 200 innings pitched doesn’t make a ton of sense, but handling them on their own also feels strange, and I don’t have a good way of converting reliever WAR to the backup/regular/star scale that I’m using here. A 3-WAR reliever wouldn’t be an above-average player, they’d be the best reliever in baseball. I settled for putting them up to 200 innings and letting that over-allocaiton of playing time handle the different measures of success. For example, a reliever projected for 3.6 WAR in 200 innings would check in around 1.2 for a full season of bullpen work. That’s a very good relief pitcher projection; only 20 players meet that bar in our 2025 Depth Charts projections.

In other words, the tier names still mostly work for relievers, but you should apply your own relative positional value adjustments just like normal. A star reliever is less valuable than a star outfielder. A star starting pitcher might be more valuable than a star outfielder, depending on the degree of luminosity, but that one’s much closer. This outcome table can guide you in terms of what a player might turn into. It can’t tell you how to value each of those outcomes, because that’s context-specific and open to interpretation.

This study isn’t meant to be the definitive word on what prospects are “worth.” Grades aren’t innate things, they’re just our team’s best attempt at capturing the relative upside and risk of yet-to-debut players. Being a 60 FV prospect doesn’t make you 17% likely to turn into a star; rather, our team is trying to identify players with s relatively good chance of stardom by throwing a big FV on them. And teams aren’t beholden to our grades, either. They might have better (or worse!) internal prospect evaluation systems.

With those caveats in mind, I still find this extremely useful in my own consumption of minor league content. The usual language you hear when people discuss prospect trades – are they on a Top 100, where do they rank on a team list, what grade are they – can feel arcane, impenetrable even. Breaking it down in terms of likelihood of outcome just works better for me, and I hope that it also provides valuable information to you when you’re reading the team’s excellent breakdown of all things prospect-related this week.





Ben is a writer at FanGraphs. He can be found on Twitter @_Ben_Clemens.

112 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Cool Lester SmoothMember since 2020
2 months ago

Absolutely incredible piece, Ben!

soddingjunkmailMember since 2016
2 months ago

Yeah, this is the stuff I come to Fangraphs for. Thanks Ben.

darren
2 months ago

Agreed!