Creating an Ideal Model For Evaluating MLB Prospects

Once upon a time, prospects were more valued as trade chips by fans of MLB teams than anything else, and young players, in general, weren’t seen as super valuable compared to established players. However, that time is past us- fans now care a great deal about their team’s top prospects, and front offices are more reluctant than ever to trade them. Therefore, public prospect rankings have grown tremendously in popularity, as fans want their teams to have as many top-100 prospects as possible. However, in my opinion, many of these prospect rankings are flawed. Two months ago,I came up with my initial hypothesis on how to evaluate prospects. Since then, I have done more research to refine my model of evaluation, which will be the topic of today’s piece.

A critical aspect of creating this model is to exclude all potential bias. Therefore, when assigning these prospects a grade for each of their skills, I use the average of multiple prospect reports. Therefore, this allows us to get a good overall assessment of each prospect. Then, to compile their overall grade, I take into account the following three skills:

Getting On Base

First off, it’s important for me to emphasize that getting on base is not represented by a prospect’s hit tool. By definition, a hit tool only measures pure hitting ability, which translates to batting average. Yet, it’s pretty obvious in the modern game that on-base percentage is a far superior metric to batting average, so a hit-tool grade is incomplete. Consequently, developing an on-base grade is critical for the evaluation process. After plotting each qualifying MLB hitter’s on-base percentage, batting average, and walk rate, I found that batting average and walk rates have roughly the same coefficient of determination in their relationship with on-base percentage. However, discipline numbers are far more stable than batted ball numbers, so after adjusting for stability, here is the final on-base grade formula:

65.5%: Discipline/Pitch Selection- Discipline numbers are very stable and are difficult to improve upon.

25%: Hit Tool- Provides a baseline for pure-hitting ability, which is important since players who have higher vertical bat angles tend to have higher batting averages on balls in play.

9.5%: Bat Control/Whiffs- Surprisingly, not striking out isn’t really correlated with getting on base, so having ideal contact skills is clearly overvalued.

Currently, evaluators often gush about a player’s pure-hitting skills. Yet, as the formula indicates, earning a 70 hit-tool grade isn’t very important if your pitch selection is lacking, so when projecting prospects to the big-league level, simply using a player’s hit-tool grade clearly would give you inaccurate results. For reference, Amed Rosario, Victor Robles, Byron Buxton, and Manuel Margot were all given 60-grade hit tools by multiple outlets, yet each of them have struggled to get on base. Meanwhile, players like Joey Gallo, Aaron Judge, and Cody Bellinger received below-average hit tool grades by specific outlets, yet their excellent plate discipline has allowed them to post high on-base percentages.

Power

Similarly to the on-base grade, you can’t simply slap someone with a “50” power grade based off of their previous production; believe it or not, but there are multiple types of power- raw power and game power. Raw power can be explained/measured by exit velocities and hard contact rates, which are more correlated to power production that ground ball/fly ball ratio, which is more associated with “game power”- maximizing on the raw power. In other words, raw power should be weighted slightly more than game power, though the basis for that goes beyond the correlation to power production. As teams such as Dodgers and Rays have shown with Chris Taylor, Max Muncy, and Yandy Diaz, it is much easier to improve a player’s power by changing the trajectory of their contact than to suddenly have them hit the ball harder. While that would appear to be obvious, the public seems to care too much about what their actual power numbers are, which makes no sense given its volatility. A prospect such as Tyler Stephenson of the Reds, for example, is seen as a below-average power hitter due to his power production in the minors, yet his exit velocities indicate he’s capable of much more with a slight swing/approach alteration. He’s definitely not the only one, and until we change our methods of projecting future power production, those types of players will continue to be overlooked.

Defense

Honestly, grading a player’s defense is very straightforward. Rather than having speed and arm strength be counted separately, they just are added to the equation of projecting their defensive ability. Still, defense is generally unstable and can be predicated on defensive positioning, so it’s important to not overemphasize that aspect of a player’s game.

Overall Grade

Based off of previous studies, we have found that offense is around five times more valuable than defense, and it’s also been revealed that getting on base is around 1.8 times more valuable than hitting for power; that’s without even mentioning how getting on base is far more stable than power and defense. With that in mind, this is my official formula for grading a position player:

53% On Base, 30% Power, 17% Defense

From this, it’s clear that on-base ability is definitely the most important trait for a player to have, while being elite defensively doesn’t really move the needle. However, this formula is only part of the equation. Since positional value is alive and well in the modern game, we need to also make necessary positional adjustments; a “50” catcher is far more valuable than a “50” first baseman or corner outfielder. To reflect this, I use the following positional adjustments based on the average statistics of each position from the past decade. These multipliers are applied to the prospects’ tool grades:

Catcher: On Base- 1.04, Power- 1.1, Defense- 1.083

First Base: On Base- 0.95, Power- 0.96, Defense- 0.9097

Second Base: On Base- 1.01, Power- 1.09, Defense- 1.049

Third Base: On Base- 1.01, Power- 1.03, Defense- 1.0081

Shortstop: On Base- 1.03, Power- 1.1, Defense- 1.0523

Corner Outfield: On Base- 0.98, Power- 1, Defense- 0.948

Center Field: On Base- 1, Power- 1.04, Defense- 1.0155

From this, it’s obvious which prospects tend to be more valuable than others; teams should be looking to build their farm systems, and rosters in general, with up-the-middle players. Not only do these players have a far greater margin for error in terms of bust-rate since they play valuable positions, but it’s far easier for them to shift to a corner position than vise versa. Elite prospects at corner positions like Spencer Torkelson, Nolan Jones, and Andrew Vaughn are still top-notch prospects, but the gap really appears in the next tier of players- you have to be flawless as a corner player to be as valuable as an above-average up-the-middle prospect. Plus, it doesn’t hurt that better athletes tend to age better than less-athletic corner players, as evidenced by the rough decline of corner infielders Miguel Cabrera and Albert Pujols.

Okay, but what about pitchers? Sure, a star pitcher isn’t as valuable as a star position player, yet with how volatile they are, teams should always be stocked with pitching depth, especially with how difficult it is to address on the open market. A previous study I conducted already demonstrated that “stuff” is about four times more important than command. With the 80% that “stuff” takes up in the formula, a repertoire should be weighted as such, based on the average pitch usage from last season:

32.5% Fastball, 29.5% Breaking Ball, 18% Changeup

The fact that a pitcher’s changeup is almost as important as command tells you all you need to know- command is all about being average, especially since it’s far easier to improve in that area than adding velocity.

However, this only takes into account a pitcher’s quality. In a correlation study I conducted, a pitcher’s efficiency (FIP) was around 2.5 times more predictive of WAR than innings pitched, but innings pitched is definitely a key factor. Those findings would have us weigh quality versus longevity in 71.5%-28.5% fashion, but since it is very difficult to accurately predict a player’s future innings, I recommend using a 85-15 formula. At the end of the day, what we can really hypothesize is the risk that they become a reliever, which I do using the following grading scale for their “stamina rating”:

60: 0% Reliever Risk

50: 25% Reliever Risk

40: 50% Reliever Risk

30: 75% Reliever Risk

20: 100% Reliever Risk

To be as precise as possible, I go by 2.5s, and overall, a player’s stamina is a lesser part of the equation for their overall grade. Yet, it would be foolish not to account for it all, since, at the end of the day, we’re trying to project which players will accumulate the most WAR, and the ability to pitch a lot of innings matters.

I’ve referenced certain data points, such as vertical bat angle and exit velocity, but which data matters the most when projecting untapped potential? For me, there are two that certainly stand out: spin rates versus spin efficiency and hard-hit rate versus launch angle. For the former relationship, raw spin rates are a great measure in talent, but in order to be effective, they need to have a high spin efficiency for it to have ideal movement. In other words, even if they have a high spin rate, it doesn’t necessarily equate to results, and that’s where teams can project improved performance- the spin rates measure their raw talent, and improving the spin efficiency can take their game to the next level. That similar line of thinking also works with hitters, as if they demonstrate an ability to hit the ball hard, an elevated swing path can allow them to convert that hard contact into more power. These two relationships stand out because they can actually be quantified to project advancement in a prospect’s abilities, which cannot be said for plate discipline. Refining a position player’s defensive skills, as well as adjusting a pitcher’s pitch usage and making their hip hinge more glute-domiannt rather than quad-dominant, also apply, but the two highlighted data associations are the ones that stand out for me when projecting a prospect’s development.

So, what should you take away from my proposed model of evaluating prospects?

  • On-base ability is the most important trait for a position player. However, it’s not reflected properly by a hit-tool grade. Therefore, we need to create an on-base grade, which heavily weights plate discipline, a far more stable metric than batting average.
  • Power is the next important trait. It’s far easier to improve in this area, and a player’s raw power should be taken into account more than what their current power production is.
  • Defense is very volatile and reliant on defensive positioning, so it’s by far the least important trait for a position player.
  • Up-the-middle players and pitchers tend to much more valuable than corner players, and that’s without taking into account aging curves. This can be reflected by using positional adjustments.
  • For pitchers, “stuff” is four times more important than command. Meanwhile, a stamina grade is necessary when projecting their future WAR production.
  • Spin efficiency versus spin rate and hard contact rate versus launch angle are the two relationships teams can look at when looking for untapped potential with prospects.

Whether it’s in the draft or just with current prospects, MLB front offices have specific models for evaluating players. While the public doesn’t have access to the data that they have, we can do our best to mimic said model, and perhaps even take it a step further. By using consensus tool grades, we can establish a baseline of how a player fares in each of the major categories. The extra twist with my model, though, is that it adjusts for positional value and weighs the skills that matter the most. Therefore, I believe it is an adequate model for not only evaluating prospects, but also projecting the players most likely to be under-the-radar stars, and which prospects are being overvalued. Sure, it is likely to lead to some unpopular results, but perhaps it shouldn’t be so controversial? If we eliminate biases, such as former draft position and reputation, and look through each prospect with an objective lens, we can better see where a prospect stands in their development. As they say, the numbers don’t lie, and until a prospect has quantifiable data to suggest they’re heading towards stardom, we cannot assume that they are destined to get there. Taking all of that into account, I’ve been able to come with a model that I believe is sufficiently backed up with different studies and research, and I recommend you create your own model as well!

2 thoughts on “Creating an Ideal Model For Evaluating MLB Prospects

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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