Building a Farm of the Best Unranked Prospects – Part 1: Pitchers

A lot is made about the strength of some farm systems, and likewise the weakness of some others. There’s plenty of reasons why stockpiling noteworthy, ranked, and on-the-radar prospects can be advantageous:

  • Longevity and cost control at the positions held by your prospects
  • You can easily fill the voids of outgoing free agents
  • Prospects are the preferred currency of low-budget teams who can land big names on the trade market but not the free agent market
  • A healthy farm is often seen as a proxy for the future health of the organization

With all that said, how good of a farm could you possibly have if you don’t have any ranked prospects? If your farm is comprised of nothing but under-the-radar guys, the benefits bulleted above pretty much all go away. The best you could really hope for is that you saw something that everyone else either missed or undervalued. But I’m guessing there’s still something in the public data that could help us find a few hidden gems. So I’m going to build a farm of unranked prospects and see it becomes anything some time down the line.

I’m going start with pitchers, TINSTAAPP be damned. Looking at THE BOARD over at Fangraphs, I noticed that there’s only 31 pitchers in the top 100 (compared to 49 at as of 11/25/18), so it’s not exactly uncommon for a team to find themselves without a ranked pitching prospect. Since my goal is to build the best possible farm system that appears to be worse than basically every other MLB team – but (hopefully) only on the surface – I have to set some baseline criteria to establish who’s eligible for my Island of Misfit Prospects, and do so in a way that ensures my farm looks really bad.

Eligibility Requirements & Composition of Farm

If this is done right, my fantasy farm will be at or near the bottom in each of the charts and tables above, and the eligibility requirements can assist with that. So here are the rules for a pitcher to be eligible for my farm:

  1. The pitcher cannot be ranked any higher than 220 overall on THE BOARD at Fangraphs.
    1. This is due in part to how common it is for a team to lack a pitcher in the top-100, which is how we generally understand “unranked” prospects.
    1. This increases the likelihood that no pitcher in our farm will be included on the most recent iteration of the major top-100 prospect lists (Baseball America, Keith Law,, etc.)
    1. Finally, the Brewers are the only team without a pitcher in the top 220. I’d have made the cutoff lower than the Brewers’ highest ranking pitcher, but unfortunately Caden Lemons is ranked 807th on THE BOARD, and that would have been extremely limiting.
  2. The pitcher cannot have any MLB experience.
  3. The pitcher must have thrown at least 10 IP in 2018.
  4. The pitcher must be a part of a MLB organization.
  5. The pitcher must have already reached the MiLB assignment level
    1. This means that a pitcher on my triple-A team must have reached triple-A at some point in his career.
    1. So I’m not necessarily advancing anyone to a higher level, but we’re going to say my fantasy farm is being created retroactive to the start of the 2018 season.

The farm should reflect the experience at each level realistically, just so I’m not throwing a bunch of unranked pitchers who threw well in low-A ball onto my triple-A team. I’m also not really interested in inexperienced pitchers who didn’t get any further than short season A-ball either, so I’m only going as deep as high-A in my fantasy farm. That gives me three teams to make up my farm; A+, AA, and AAA. Let’s also say I’m building the 2018 version of my farm, just so I don’t have to worry about 2019 assignments or promotions; the goal is to see how everyone pans out moving forward. So here’s how each team’s pitching will be comprised:

  • 8 relief pitchers; over half of minor league appearances must have been made in relief
  • 5 starting pitchers; must have started at least 50% of minor league appearances

So without further delay, let’s move on to the players I’ve selected…




Zac Lowther


In retrospect, it seems that I maybe should’ve been more restrictive on the eligibility of Fangraphs BOARD players by either setting a rank cutoff at something worse than 220, or by capping the amount of players who appear on the BOARD at a maximum that better reflects a bottom-of-the-barrel collection of minor league pitching. I say this not because my Fantasy Farm is “good”, but because it didn’t turn out remarkably “bad” in terms of perception. It is completely devoid of stud prospects, and also appears to be pretty below average when the analysis gets deeper, it isn’t exactly a bottom feeder. Check out the following charts:

Notice how the red column (representing my Fantasy Farm) moves a little more to the left as the charts descend – that’s by design. It illustrates how my collection of pitchers may appear a little less hopeless depending on how we want to evaluate farms.

Nevertheless, by common standards, I’ve put together an unremarkable, below-average farm system with no big-time prospects as of the end of the 2018 season. Other than the eligibility rules I set up, the biggest factor in my selection was a data model of MLB performance based on minor league performance. The end-result is the collection of no-name prospects you saw in the tables above, who are generally young for their league, and possess either good K%-BB% or GB% rates, if not both.

Evaluating Outcome & Final Thoughts

In terms of how I’ll measure my success, I probably put my selections at a slight disadvantage by prohibiting any pitchers with MLB experience and assigning them only to levels they’ve already pitched at. Those two rules make it so I have a collection of guys who are, in all likelihood, set to repeat triple-A in 2019. I don’t have data to cite, but I’d imagine non/unranked prospects who repeat triple-A are considerably less likely to make it to the Big Leagues than a guy who ended the prior year in double-A and is assigned to triple-A at the start of the following year. Additionally, it’s hard to find a diamond in the rough who’s managed to reach the highest levels of the minors without gaining any traction as a prospect, so I’d imagine the true talent level of my Fantasy Farm is much lighter at the top than at the bottom.

That being said, I still think my farm will outperform any expectations set forth if my farm actually existed. I don’t have anything set in stone as far as how I’ll determine the success of my farm, but I have a few ideas, and I’ll probably use at least a couple of them (save for those that I ultimately conclude are relatively useless). Keeping in mind that I’m referring to future performance/debuts/numbers, here’s what I’m thinking…

  • Measure success by MLB debuts and/or performance
    • 160 pitchers made their MLB debut in 2018 (though not all were full-time pitchers, i.e. Willians Astudillo)
    • The average number of debuts by a team was 5.3
    • This number is skewed toward bad teams and doesn’t necessarily represent talent on the farm
    • Only about 20% of debuts came from playoff teams, which make up a third of MLB (33.3%)
    • These pitchers went 265-276 with a 4.55 ERA and 1.39 WHIP, striking out 4248 and walking 1978 in 4766.1 IP, so the bar probably won’t be set too high
  • Measure success by comparing performance across all levels to that of top-ranked farms (compile aggregated figures by org and compare)
  • Measure success by aggregating numbers of all pitching prospects ranked better than 239 on the 2018 update of the Fangraphs BOARD
    • According to the version of the Fangraphs BOARD in reference, the highest ranking member of my Fantasy Farm is Jose Suarez of the Angels at 239
    • By compiling figures of all pitchers ranked above 239, I’m looking exclusively at pitchers regarded more highly than my top-ranked pitcher
    • 98 pitchers are ranked in front of Jose Suarez, so there’d likely be a large body of work to compare
    • Since both my fantasy farm and the top 98 pitchers are likely to put up numbers at all levels, the data can be collected and measured both as a whole, and/or by level

Consider this particular exercise just a part an ongoing analysis. Regardless of how the results look at the end of 2019, it still won’t be finished. The conclusion of all this probably won’t come for a few seasons, but I’ll keep checking in on the results periodically and reporting them. Sure, in the end I’d like to be able to say my farm did better with a bunch of no-names (at the time) than the top-ranked farms, but I might crash and burn too. All we can do is watch it unfold.