Recently we debuted a new tool on RotoWire we are calling Prospect Trends. The page is built on a model that tries to quantify MiLB performance and develops a ranking for players based on their current season performance. It is important to note that this page relies entirely on stats from a player's most recent season (current season in-year and the last year in the offseason). There is no prospect grading that goes into this; it is entirely stats based.
How Should I Use This?
Like all other prospect pages on RotoWire, if you leverage the My Leagues feature you can connect your favorite league and find the guys who aren't rostered in your dynasty leagues. The idea behind the tool is to try to help you to get ahead of the pop-up prospects or risers in between updates to the Top 400 rankings. These rankings and the model output is meant to be used for Fantasy purposes only, there is no defense or any other positional considerations. So if you have waivers in your Dynasty leagues and are trying to find out what to do with your last few roster spots, this tool will be your best friend. Every Friday over the course of the season we will update these rankings frequently trying to highlight some of the possible biggest risers to help you get in on them before the rest of the industry catches up. So, the first update for the 2026 season will come in a few weeks once Minor League games get going. For this reason, the model prefers high upside fantasy impact bets over safety and floor.
How Does The Model Work?
I will try my best to explain this concept without too much of the gory math details and concepts and try to speak to the idea conceptually. In order for a player to be considered in the model they need to have at least 50 PAs or 10 IP at a given level (A and above). So if a player has 100 PAs at AA and 10 at AAA, we will only be looking at the 100 PAs at AA. Each level's results are taken independently and then for the final output are weighed according to the PA/IP obviously giving more credence to the larger sample.
The entire crux of the model is based on a concept called Mahalanobis Distance. This is a measure of distance or similarity that allows us to provide a series of statistics and weights for those different statistics to find the most similar seasons. So, for every player season, we are using metrics like Age, wRC+, K%, and more to find the most similar seasons among previous MiLB performances. AA seasons will only be compared to AA seasons. Each stat has a weight in determining the similarity that is determined by how strongly it correlates with future MLB performance. Future MLB performance is defined in this context by SGP Values for a standard Rotowire Online Championship style 12-Team league per 600 PAs or 150 IPs. The idea is to isolate Fantasy Performance over the course of a player's career defined within the context of a season.
Beyond the final ranking, the model provides three different metrics: Value, Adjusted Value, and Elite Rate. Let's take a look at the top-10 non-debuting players from 2025 and discuss the different values and what they are conveying.
Name | Value | Adjusted Value | Elite Rate | Rank |
7.072 | 1.889 | 13.8 | 1 | |
6.848 | 1.665 | 13.8 | 2 | |
6.917 | 1.734 | 13.7 | 3 | |
6.095 | 1.544 | 14.6 | 4 | |
6.878 | 1.695 | 13.7 | 5 | |
6.781 | 1.598 | 13.7 | 6 | |
6.763 | 1.58 | 13.7 | 7 | |
6.761 | 1.578 | 13.7 | 8 | |
6.725 | 1.542 | 13.7 | 9 | |
6.585 | 1.402 | 13.7 | 10 |
Value is pretty simple, it is the weighted average (closer comps get more weight) of the SGP per season values for all of the player's comps. As you'll see, Dax Kilby is the top player in the rankings and with a Value of a little over 7. However, this does not tell us much about the distribution of that output and does not allow us to differentiate between players too much.
Adjusted Value is similar to value except it compares the player to the average player at their level (Hitters compared to Hitters and Pitchers to Pitchers). As players get closer to the majors, the average player they are compared to tends to be better. More players who have played at AAA make it to the majors when compared to guys in A-ball. So this just helps to contextualize players at different levels.
Elite Rate in my opinion is the most important portion of the output. This is simply the average of the two best performing comps among the player's closest 100 comparable seasons. The idea here is we are trying to isolate the idea of if this players hits a ceiling, what does that fantasy performance look like. The goal of prospecting is to find the hits, we do not really care too much if our player carves out a decent role and settles in as a 400+ ADP pick, we want to try to get in on the ground floor of players who will be perennial Top-100 redraft picks.
So the simple way to view these metrics is Value determines a player's floor while Elite Rate is their ceiling. Brennen Davis put up a AAA season that gave him some Elite comps, the highest in the sample in fact, but his skillset gives him a much lower floor compared to these other top ranked options. The Rank blends Value and Elite Rate with a penalty for advanced age prospects. As alluded to before, we are giving more weight to ceiling and upside in this ranking as that is our real goal as dynasty managers.
Check out James Anderson's Top-400 MLB prospects here.
What Are the Flaws?
I think as a data person who has built a lot of different models over the years, one of the most powerful things to highlight when evaluating a model is what it struggles with. First of all, as stated before, there is absolutely no Prospect Grading or context of that form in the model. James Anderson does incredible work for us in that area and as we all know with prospects statline scouting can lead us astray at times. This is not meant to replace James' work at all - it is a supplement and another tool to help evaluate prospects. Additionally, this is only looking at one season at a time and does not have context on a player dealing with an injury or any other contextual information that may help to explain a performance. This is another reason why we should not be dumping our highly ranked stud prospects who do not grade out well in the model.
From an actual output standpoint, the model tends to love players who show any sort of speed pairing with above average offensive performance. Kilby for example, stole 19 bases in 81 PAs while posting a 159 wRC+ in A. The offensive performance and speed, if continued at the MLB level would make Kilby a superstar. This is a feature and not a bug within the model. I built this for my own use and as you can likely tell, I am taking big swings on prospects in my Dynasty leagues. I want truly elite performers and taking chances on the players with speed alongside offensive performance is one of the easiest ways to find that elite fantasy performance.
On the flip side, it tends to be lower on some of the more "boring" all around profiles. Someone like Kevin McGonigle, who has what is a more rounded offensive profile, leads to a higher Value but a lower Elite Rate and causes him to be rated lower in this ranking set than his elite prospect ranking. Also, sometimes the super elite seasons that take place across multiple levels may get caught up in small sample weirdness and lead to a lower than expected ranking. Konnor Griffin falls into this boat. All of this is to say, this is meant to be used as a supplement and not a source of truth. If you remember that we are trying to find players who will be the biggest risers, that context will make this tool extremely useful in your deeper dynasty leagues.
What's Next
In the coming weeks, I will be working on a few additional pieces going a bit more in depth into the hitter and pitcher sides of the model and looking at some players that the model likes that did not make James's Top 400 Prospects list and understanding why it may be higher on them than he is. We can look a bit more at some player archetypes and maybe highlight some of the actual player comps that have been spit out for some of the more interesting prospects.
