Setting the Table
With the string of commits from prospects rated as mid-four stars (5.9 rating) over the past month and recent thread where @Ketchum basically poo-pooed on this tier of prospects, I thought now would be a good time to have a board-wide discussion on the value of a mid-four stars and where they fit into the puzzle of building a competitive program. I must admit, reading Ketch argue back and forth with @Horny Gambler and @gamer_989 over statistics related to this tier of prospects gave me flashbacks to when the Inside the NBA crew got into a debate about filling up their gas tanks, and I’m not sure who is Shaq in this scenario. All I can do is be Ernie and shake my head.This is going to be a lengthy post (TL;DR at the end) that I hope everyone finds informative and adds to the recruiting perspective/narrative of this board, which I personally feel has gotten a bit hyperbolic. Before I dive in, I want to first state a few things every poster on this board intuitively understands. It is better to obtain commitments from higher ranked players. Higher ranked players “hit” at higher rates than lower ranked players. Evaluation matters. Development matters. Coaching matters. Winning matters. But above all, when it comes to analyzing data, context matters. Now let’s wade through the minutia.
Are 5.9s closer to 6.0s or 5.6s?
Let’s first address the claim by Ketch that began this whole debate:
I must admit, I chuckled seeing Ketch condescendingly reply to posters about simple mathematics only for him to later post in the thread a graphic from his 2023 NFL draft analysis undermining his own claim (be better Ketch
). However, when using the 2014 draft rate data posted originally by Horny Gambler the truth is that both Ketch and gamer_989 are technically correct about whether 5.9s are closer to 6.0s or 5.6s. The difference is that Ketch uses absolute difference in draft rates for comparison to justify the latter whereas gamer_989 uses relative difference in draft rates to justify the former. This is a great example of how two people can look at the same data and arrive at different conclusions.But which method of comparison is better? They both have drawbacks, namely that absolute differences on large numbers can look significant even if the relative change is small while relative differences on small numbers can look overly significance in the setting of small absolute change. Since the rate at which 5.6s are drafted is small, I tend to lean towards using the absolute difference in percentage points of draft rates between rating levels for comparison. Below are the draft rates (%) for each tier of 3, 4, and 5 stars from the 2014-2023 NFL drafts (as provided by Ketchum here). Note that a given draft rate for a single draft is calculated based on a three-year window, e.g., for the 2014 NFL draft hit rates Ketch used the 2009-2011 Rivals high school rankings.
NFL Draft "Hit" Rates by Rivals Rating Tier (2014-2023):
Rivals Rating Tier || Year | 2014 | 2015 | 2016 | 2017 | 2018 | 2019-2022 | 2023 |
6.1 | 47.62 | 69.93 | 61.35 | 83.16 | 58.22 | ? | 68.8 |
6.0 | 43.67 | 42.37 | 49.5 | 30.47 | 29.1 | ? | 40.7 |
5.9 | 23.92 | 24.88 | 30.77 | 15.79 | 21.18 | ? | 27.6 |
5.8 | 16.72 | 15.15 | 18.55 | 16.03 | 19.28 | ? | 13.1 |
5.7 | 12.3 | 11.99 | 8.43 | 8.59 | 13.97 | ? | 7.6 |
5.6 | 5.72 | 7.27 | 6.37 | 6.57 | 7.53 | ? | 8.7 |
5.5 | 3.49 | 4.71 | 5.03 | 4.37 | 4.34 | ? | 5.0 |
If we look at the absolute differences in draft rate percentage points year-by-year we get:
| 2014 | 2015 | 2016 | 2017 | 2018 | 2019-2022 | 2023 | |
| Δ6.0-5.9: | 19.75 | 17.49 | 18.73 | 14.68 | 7.92 | ? | 13.10 |
| Δ5.9-5.6: | 18.20 | 17.61 | 24.40 | 9.22 | 13.65 | ? | 18.90 |
Impressions:
- I am not a huge fan of a year-by-year comparison analysis due to each NFL draft utilizing a three-year window. This means a single high school class is used to calculate hit rates for three different NFL drafts (e.g., The 2011 Rivals high school class is used in the 2014-2016 NFL draft calculations) so their data is overrepresented in a year-by-year comparison analysis. Rather than relying on three-year moving windows, a better analysis would be to take every player from every recruiting tier from 2004 (the first year I can tell Rivals implemented a rating tier system to complement their star system) through 2020 and calculate the rate at which these tiers “hit” in the NFL draft. This would eliminate using arbitrary three-year windows and provide a much larger sample size to better understand the true differences between the rating tiers.
- However, because Ketch’s aggregate data is the best data set I am aware exists, we shall rely on the year-by-year comparison to determine a victor of this semantically insignificant debate. Although close, I must declare gamer_989 the winner as a greater absolute difference in NFL draft rates existed between 5.9 and 5.6 stars than between 5.9 and 6.0 stars in 4 out of the 6 years of the NFL draft with aggregate analysis available (not sure what happened between 2019-2022 but I could not find a tiered analysis breakdown for these years in your One-stop Shop page Ketch).
I note above the debate between Ketch and gamer_989 is semantically insignificant because anyone can see there is an obvious gap between 6.0s and 5.9s as well as between 5.9s and 6.0s. The true heart of the debate rests on their views of what the value a 5.9 four-star recruit represents. To even begin answering this question one must first understand how Rivals implements their rating tier system. A 6.1 represents a 5* with roughly 32 players each cycle earning this grade (i.e., the same number of 1st round picks in the NFL draft). Over the past decade an additional 32-38 players receive a grade of 6.0 and high 4* status - think players ranked 33-70ish. Our hotly debated mid-four star prospect is where things start getting tricky, because the number of players who receive a 5.9 rating ranges anywhere from 70 (~#71-#140) to 120 (~#70-#190) over the past decade. This massive variation in prospect allotment makes it difficult to pin down the true value of this tier. My suspicion is that the 5.9 tier is actually composed of two sub-tiers: a group of prospects with NFL draft rates closer to 28-33% and a group of prospects with NFL draft rates closer to 20-25%. I would argue the former group represents very high value, albeit not elite value like your 6.0 and 6.1 players. Take Derek Williams for example – the #70 ranked player in 2023 with a measly 5.9 rating. If I had to guess, I would predict the prevailing sentiment on this board is that Williams represents a high value recruit with a good shot at making the NFL (likely closer to 40% than 25%).
Number of Prospects per Rivals Rating Tier per Year:
| Rivals Rating Tier | | | |
Year | 6.1 | 6 | 5.9 | Cumulative Total |
2006 | 28 | 40 | 113 | 181 |
2007 | 29 | 47 | 99 | 175 |
2008 | 30 | 52 | 94 | 176 |
2009 | 33 | 55 | 82 | 170 |
2010 | 26 | 26 | 84 | 136 |
2011 | 26 | 16 | 47 | 89 |
2012 | 32 | 29 | 49 | 110 |
2013 | 33 | 54 | 73 | 160 |
2014 | 33 | 37 | 71 | 141 |
2015 | 34 | 36 | 97 | 167 |
2016 | 31 | 41 | 94 | 166 |
2017 | 34 | 38 | 108 | 180 |
2018 | 33 | 37 | 105 | 175 |
2019 | 30 | 41 | 97 | 168 |
2020 | 35 | 35 | 71 | 141 |
2021 | 31 | 32 | 101 | 164 |
2022 | 33 | 36 | 104 | 173 |
2023 | 32 | 37 | 120 | 189 |
If I could encourage Ketch to make a single graph, it would be this one: player rank (e.g., 1, 2, 3, etc.) versus percentage of players that rank drafted in the NFL for the top 250-300 prospects from the first year of available data to 2020. This graph would no doubt be heavily right-skewed (i.e., a long tail to the right), but it would give great insight into where significant step-offs in draft rate occur as one moves down the national recruiting ranks. This graph would allow us to better see the cutoff prospects go from 1 in 2 odds of being drafted to 1 in 3 odds, to 1 in 4 odds, etc. I would argue a prospect with 1 in 3 odds still represents excellent value and this cutoff probably lies around #110-#125 nationally. With this data set you could also perform a supervised binning task (e.g., using a decision tree) to develop more optimal rating classifications based on maximizing the differences in NFL draft rates. I suspect the 6.1 cutoff would remain similar, but the 6.0 and 5.9 cutoffs would be shifted.
Recruiting Philosophy
So, what should the coaches recruiting philosophy be with respect to 5.9 mid-four star prospects? As shown above, this tier of prospect is almost equidistant to 6.0s as they are to 5.6s in terms of NFL draft rate absolute percentage points, with significant drop-offs of around . However, remember that when it comes to data analysis CONTEXT is king. The context of this discussion is not only within a 25 player recruiting class and an 85 man college football roster, but also the competitive landscape of college football recruiting.Ketch and other posters like the consistently gloomy @CS (wait that is just Ketch twice), consistently advocate for and push the notion that the ONLY caliber of player which define the quality of a recruiting class are 6.1/6.0s, or as ketch calls them “super blue chips.” History shows that landing these players in significant enough quantities is, to borrow Tom Herman’s favorite expression, REALLY HARD. Let’s take a look at how many of these players Texas has landed since 2002:
- Texas 2002-2023: 7, 4, 3, 3, 6, 5, 4, 5, 6, 2, 4, 2, 0, 1, 3, 1, 5, 4, 1, 1, 3, 7
- Alabama 2010-2023: 1, 6, 7, 12, 10, 9, 7, 12, 4, 6, 7, 13, 8, 10
- Clemson 2010-2023: 1, 4, 1, 2, 3, 5, 4, 3, 3, 7, 3, 9, 5, 2, 2
- Georgia 2010-2023: 3, 3, 5, 1, 3, 5, 3, 7, 11, 9, 7, 4, 9, 9
- Ohio State 2010-2023: 0, 3, 2, 6, 7, 4, 4, 10, 10, 3, 8, 7, 4, 5
Impressions:
- Ketch is right when he states the top programs have the highest number of super blue-chip prospects, often by A LOT. When it comes to winning conference titles and championships, these prospects matter more than other rating tiers.
- There is a positive feedback loop in recruiting whereby teams with high levels of success are then able to recruit more high-level players which in turn brings high levels of on-field success. Again, nothing earth shattering, but this consolidation of talent appears to be strengthening within the past decade. This may be because of the CFP, social media, increased number of camp circuits, or likely a mixture of all the above.
- Consistently landing 5-7 super blue chips will likely lead to a very good football team that contends for a championship. Consistently landing 8+ super blue chips leads to elite, dynastic levels of success. This level of recruiting has only been achieved by Alabama, Georgia, and Ohio State under CUM over the past decade plus. If Texas wants to reach their level then they will need to replicate their success when it comes to landing super blue chip prospects.
Although the goal for Texas should be to match the pace set by Alabama, Georgia, and [now to a slightly lesser degree] Ohio State in recruiting, the reality is that the Texas program has not achieved enough on-field success to harness the positive feedback nature of recruiting. Sark and Co. landing 4-6 super blue chips per class would represent very high level recruiting given the state of our program. The good news is that it would appear the staff are positioned to accomplish this once again in the 2024 recruiting class. Furthermore, they have the roster entering the 2023 season to be positioned as favorites to win the the 2023 Big 12 title, which might just represent the thing needed to shift Texas' recruiting prowess up a gear.
Even with elite level recruiting, teams cannot fill out a recruiting class or roster solely with super blue chip prospects. There just aren't enough of them. This begs the question then, how should Texas fill out the rest of a recruiting class. The obvious answer is from the lower rated tiers of high school recruits or the portal. Here I favor the line of thinking that consistently strong programs are built from the high school ranks rather than the portal. While I acknowledge the portal can be used to effectively plug holes in a program’s roster, I would push back that it can be leveraged as a main source of long-term success. As @Suchomel says, there are a reason these players are entering the transfer portal and more times than not they were not BAMFs at their original school. Given this entire discussion has revolved around recruit rating tiers, let’s see how many players from each tier were available in 2023 transfer portal:
- 6.1: 1
- 6.0: 4
- 5.9: 24
- 5.8: 129
This brings us back to the high school ranks and our 5.9 mid-four star prospects. Although it is true they are not in the same ballpark as your super blue chip 6.1/6.0 prospects, they still hit at high enough rates to field a competitive team. For example, if Texas were to land 10 mid-four stars and 5 6.1/6.0 per class, the data suggests they would produce 5-6 future NFL players. Stack that across four years and we are talking 20 to 24 NFL players in your program from those 3 rating tiers alone. While that might not be enough to unseat Alabama or Georgia, it certainly is enough to shift the trajectory of the Texas program upwards and have them consistently in the 12-team CFP and threatening to get over the hump. Put simply, Texas needs to land not just as many 6.1/6.0 players as they can, but they also need to hoard as many 5.9 prospects as possible (the higher ranked the better obviously) because they represent significantly greater value (both statistically and materially) than the tiers of recruits below them. Landing 10 5.9s vs 10 5.8s per class for 4 classes equates to an extra 4 future NFL players in your program. Landing 10 5.9s vs. 10 5.7s equates to an extra 6 future NFL players in your program. Again, this won't be good enough to unseat Alabama or Georgia, but in a game where fine margin often decide who wins and who loses, I would argue these margins represent significant value to a program. While super blue chip prospects may represent the bar for competing for championships, the notion that recruiting success is 6.1/6.0 or bust is simply an over exaggeration and acting like there isn’t really good value in the 5.9 tier is plain folly.
TL;DR:
1. Although the absolute difference in percentage points is close, 5.9 mid-four star prospects are drafter at a rate that is closer to a 6.0 prospect than mid-three star 5.6 prospect. Furthermore, with a rate of 1 out of every 3.5 to 4 prospects developing into NFL drafted players, this tier of recruits represents very good value.
2. The 5.9 mid-four star tier should probably be broken into two sub-tiers. Further analysis is needed, but it is likely the top 125 prospects or so have at least a 33% chance of being drafted into the NFL.
3. Outside of Alabama and Georgia (who are on another level), the goal of any program should be to consistently land 5-7 super blue chip 6.1/6.0 prospects PLUS as many highly ranked 5.9 mid-four stars as they can if they want to compete at a high level (top 5 program).
4. While the portal can be great for shoring up a roster and adding depth, there are simply not enough “super blue chip transfers” to allocate 8-10 spots for each recruiting cycle. A more realistic number would be 5ish. Regardless, it is still to be determined whether sustained success can be built by relying on the portal for several high level impact transfers each year.
5. While I appreciate Ketch’s goal to make recruiting a science, I personally feel there is more gray area to recruiting than his narrative of building a program solely through super blue chips and the transfer portal allows, even if I agree with the essence of his super blue chip sentiment.
