Underachieving, Recruiting, and Expected Wins

One of the criticisms of Les Miles is that while he is able to bring in a bunch of talent, that same talent underperforms once it gets here. It's the double-edged sword of recruiting: you need to bring talent to campus, but then if you don't win with that talent it's an indictment of your coaching abilities.

There's some truth to that criticism. LSU has recruited well under Miles, and it is frustrating to finish in the standings behind teams that supposedly don't have the same talent levels. Then again, Alabama has been snapping off No. 1 recruiting classes for so long, every single SEC team has a built=in excuse for losing.

But how fair is this line of thinking? How has LSU done relative to the expectations of the talent recruited? Well, first that means trying to measure how much talent is supposed to be on each team. This required looking at every recruiting class and the overall score assigned to the class by 247sports' composite ranking.

This is where I stop to tell you that we're about to do a bunch of math. If you don't want the nitty gritty, just skip ahead to the charts. You can meet us there. I promise to not get into anything except for quantifying the expected talent on hand from recruiting and then coming up with projected wins based upon that talent. We'll draw conclusions after the charts.

Still here? Cool.

We can't just look at the four most recent classes, add up the scores of each class and call it a day. That's because teams don't rely much on their freshmen class and quite heavily on their junior class in these days of early entry to the draft. I looked at the makeup of the All-SEC teams over the past few years and came up with a ratio of the representation of each class on those teams. This ratio is then used as a multiplier for the class scores from the consensus ranks. The multipliers are:

Freshman 0.5

Sophomores 2.0

Juniors 3.0

Seniors 2.5

Now, there are some huge caveats here. We're not accounting for redshirts. Nor are we tracking attrition per se, though the lower value for the senior class captures the effect of average attrition for a class. Juniors are more represented on All-SEC teams, which shows that teams rely most heavily on their junior talent.

This gives us a rough approximation of what a team's talent SHOULD be. Now, it could be lower like LSU last season, which had an extremely small senior class left on campus. Or it could be greater for a squad with a successful redshirt program. But those are results of coaching and managing talent. We're trying to first look at the talent that should be on hand, and then what the staff did with it. Let's take an example:

LSU, last season, had three consecutive top ten classes in the barn, and a weaker senior class by the recruiting gurus. The class scores by year were:

2012: 247.120

2013: 281.030

2014: 299.300

2015: 282.350

Now, we multiply each score by its class multiplier: 2012 by 2.5, 2013 by 3, 2014 by 2, and 2015 by 0.5. Then, we add up those scores for a total value which I'll call the Expected Talent Score. If you do the math, LSU's Expected Talent in 2015 was 2200.67, ranking third behind Florida. The SEC average was 1947.87 and Alabama led the conference with a 2527.36.

This makes sense that Alabama is lapping the field due to their run of top-ranked classes. There should be a massive talent gap, while most other teams are clustered closer to the league average. Which brings us to the more important step, trying to project a team's win total based upon expected talent on hand, and how the teams have performed against expectations.

I sense I'm losing most of you on the math, but I do like to show my work. After multiple false starts, I finally came up with a decent projection system based upon Expected Talent. I would warn you that it doesn't work as an actual projection system because the system is working off of Expected Talent on hand, not Actual Talent. For that, we'd have to track the redshirts and attrition of each individual team and make adjustments to the Talent Score. But the goal here is not to build the best projection system in the real world, but a projection system which isolates recruiting. Any decline in Expected Talent to Actual Talent is due to roster management and is a reflection of coaching ability. And we're trying to see if Les the Coach betrays Les the Recruiter.

Anyway, the Average SEC team keeps improving due to increased recruiting dominance in recent years. So the projection system accounts for the average SEC team and by design, makes the average team project at precisely 4.0 conference wins.

We then subtract a team's Expected Talent Score from a replacement level SEC team, which is a score of 1300. That is the lowest possible score imaginable for an SEC squad, and not even Vanderbilt is down at those depths. That would be the score if Vanderbilt churned out its worst class in the past decade in four consecutive seasons.

Then we divide a team's Expected Talent Score Over Replacement by the Average SEC Team's Score Over Replacement multiplied by two. Let's use LSU in 2015 again to illustrate.

The average SEC team had a score of 1947.866 which is 647.87 over replacement value. In order to give the average team a 500 winning percentage, we multiply 647.87 by 2 to get 1295.732. LSU's Expected Talent was 220.67 and 900.67 over replacement. We divide 900.67/1295.732 to get an expected winning percentage, then multiply by 8 (the length of the SEC schedule) to get LSU's projected number of wins: 5.6. LSU had 5 wins last season in SEC play, so it was pretty close to its talent projection. Let's see that for that entire SEC in 2015:

2015	Score	Wins	Over Rep	Proj %	Proj W
Alabama	2527.36	7	1227.36	0.947	7.6
Florida	2241.40	7	941.40	0.727	5.8
LSU	2200.67	5	900.67	0.695	5.6
Georgia	2145.07	5	845.07	0.652	5.2
Auburn	2138.85	2	838.85	0.647	5.2
Texas A&M	2108.26	4	808.26	0.624	5.0
Tennessee	1930.80	5	630.80	0.487	3.9
Mississippi	1903.14	6	603.14	0.465	3.7
South Carolina	1880.83	1	580.83	0.448	3.6
Arkansas	1733.96	5	433.96	0.335	2.7
Mississippi State	1716.78	4	416.78	0.322	2.6
Kentucky	1592.58	2	292.58	0.226	1.8
Missouri	1587.98	1	287.98	0.222	1.8
Vanderbilt	1562.47	2	262.47	0.203	1.6
Average	1947.87		647.87		4.0

Here's where we get into Auburn and the Sneaky Good Defense. Auburn has recruited well. Their recruiting ranks, by year, heading into 2015 were 11^th, 9^th, 6^th, and 8^th. And all of the expected talent on hand didn't mean squat, as Auburn won just 2 games, 3.2 less than their projection.

Another interesting bit of data is that this method shows Ole Miss as being slightly under the SEC average in its expected talent. How could this be, given their plethora of first round picks. Well, that's because in Ole Miss' case, the formula overrated their senior class, as the recruiting ranks by year were 47^th, 6^th, 15^th, and 17^th. That 47^th ranked class, which should have made up the bulk of their senior class, acted as an anchor on their overall projection. In real life, the Rebels dealt with this issue by moving on from their senior class almost entirely, and relying on the three other classes for the bulk of their roster, as well as the 20^th ranked class from 2011 for a few redshirt seniors.

But that shows exactly how good coaching and roster management can overcome a misstep on the recruiting trail. Hugh Freeze deserves all of the credit that came his way for getting the most out of his roster, and hiding any expected talent deficiencies. Which is exactly the sort of thing we're trying to find.

The SEC expanded to 14 teams in 2012, and we can look back over those past four seasons to see which teams got the most out of their talent:

TEAM	2015	Wins	Proj	2014	Wins	Proj	2013	Wins	Proj	2012	Wins	Proj	Wins+
Missouri	1587.98	1	1.8	1498.35	7	1.3	1535.13	7	1.6	1570.90	2	1.9	2.59
Vanderbilt	1562.47	2	1.6	1488.09	0	1.3	1403.09	4	0.7	1305.17	5	0.0	1.84
Mississippi State	1716.78	4	2.6	1678.36	6	2.5	1641.38	3	2.4	1674.82	4	2.6	1.73
Mississippi	1903.14	6	3.7	1804.21	5	3.4	1740.21	3	3.1	1771.36	3	3.3	0.89
Texas A&M	2108.26	4	5.0	1900.17	3	4.0	1780.07	4	3.3	1735.49	6	3.1	0.41
South Carolina	1880.83	1	3.6	1912.22	3	4.1	1867.30	6	3.9	1875.61	6	4.0	0.09
Georgia	2145.07	5	5.2	2168.14	6	5.8	2138.19	5	5.8	2146.69	7	5.9	0.06
Kentucky	1592.58	2	1.8	1515.04	2	1.4	1465.23	0	1.1	1468.18	0	1.2	-0.39
Alabama	2527.36	7	7.6	2475.27	7	7.8	2378.70	7	7.5	2310.75	7	7.1	-0.50
Arkansas	1733.96	5	2.7	1757.10	2	3.0	1701.03	0	2.8	1698.33	2	2.8	-0.58
LSU	2200.67	5	5.6	2138.17	4	5.6	2116.13	5	5.7	2191.07	6	6.3	-0.77
Florida	2241.40	7	5.8	2226.46	4	6.2	2307.24	3	7.0	2294.91	7	7.0	-1.24
Tennessee	1930.80	5	3.9	1916.27	3	4.1	2003.08	2	4.9	2069.24	1	5.4	-1.82
Auburn	2138.85	2	5.2	2148.50	4	5.6	2176.90	7	6.1	2058.55	0	5.3	-2.31
Average	1947.87	56	56	1901.88	56	56	1875.26	56	56	1869.36	56	56	0

Negative numbers are bad. Wins+ is how many wins over projection a program has per season. This shows the terrific job Missouri has done beating the projections, at a tune of two and half wins per year. It also shines a light on the team which has been the most disappointing in SEC play since the recent expansion: Auburn.

Now, we all wish we could squeeze in a national title game appearance in our four year run of disappointment, but it distracts us from how massively Auburn has underperformed these past four years.

LSU and Auburn have had remarkably similar expected talent levels. While LSU has churned out good year after good year, enraging the fanbase that wants great year after great year, Auburn's win total has gone all over the place. LSU's win totals have been 6 -€” 5 -€” 4 -€” 5. Auburn? 0 -€” 7 -€” 4 -€” 2.

This leads to LSU slightly underperforming its projection, as LSU should be winning at a 6-win pace instead of its 5-win pace per season. Auburn should also be winning near 6 games a season. Instead, it won 7 games in 2013, and 6 games in the other three season combined.

There's also the matter of Alabama showing up in the negative column despite winning 7 games in each of the last four seasons. You could wind your watch by the Bama win total, if people wound their watches anymore. However, they have amassed such a talent advantage over the rest of the SEC and college football, that 7 wins doesn't meet their expected win total in any of those seasons. In 2014, Bama was so loaded that they projected at 7.8 wins, while only one other team cleared a projected six wins (Florida at 6.2).

This chart also shows that both Mississippi schools are playing far above what we would expect based upon recruiting which highlights not only how good they've been on the field, but the fact a class ranked 15^th nationally doesn't even get you into the top half of the SEC anymore.

Then, we should acknowledge Vanderbilt, who despite winning just 2 games in the past 2 seasons combines, is still well above projection. That's because the 2012 was as close to zero as an SEC tam could get in expected talent, but still won 5 games.

The concern for every fan should be how quickly the "Average" column is climbing. In 2012, the average SEC team had a score of 1869.36, which is great, but it was up to 1947.87 last season and 1975.74 this year. An Expected Talent Score of 1975.74 would have projected at 4.7 wins in 2012. Now, it's the average.

Let's compare talent from the SEC to some of the top recruiting teams in the nation.

Talent	2016	2015	2014	2013	2012
Alabama	2526.29	2527.36	2475.27	2378.70	2310.75
Arkansas	1739.58	1733.96	1757.10	1701.03	1698.33
Auburn	2189.86	2138.85	2148.50	2176.90	2058.55
Florida	2118.25	2241.40	2226.46	2307.24	2294.91
Georgia	2174.44	2145.07	2168.14	2138.19	2146.69
Kentucky	1651.92	1592.58	1515.04	1465.23	1468.18
LSU	2310.48	2200.67	2138.17	2116.13	2191.07
Mississippi	2042.06	1903.14	1804.21	1740.21	1771.36
Mississippi State	1711.51	1716.78	1678.36	1641.38	1674.82
Missouri	1585.18	1587.98	1498.35	1535.13	1570.90
South Carolina	1861.11	1880.83	1912.22	1867.30	1875.61
Tennessee	2055.48	1930.80	1916.27	2003.08	2069.24
Texas A&M	2164.54	2108.26	1900.17	1780.07	1735.49
Vanderbilt	1529.75	1562.47	1488.09	1403.09	1305.17
Florida St	2235.51	2223.03	2272.48	2259.39	2161.56
Ohio St	2346.48	2345.02	2294.81	2133.26	2098.31
Michigan	1954.56	2121.66	2049.71	1909.02	1895.35
USC	2186.55	2127.38	2173.08	2262.67	2302.98
Clemson	2035.61	1983.22	2029.95	1947.86	1797.11
Texas	1983.78	2060.42	2185.86	2338.45	2324.45
Notre Dame	2150.86	2119.12	2105.89	2019.94	1959.15
Miami	1943.08	2032.26	1919.00	1855.29	1846.32
Oklahoma	1966.59	1974.45	2001.22	2096.36	2129.90
UCLA	2077.52	2050.80	1829.81	1792.84	1888.09
Oregon	1862.52	1888.79	1956.25	1988.15	1886.74
SEC Average	1975.74	1947.87	1901.88	1875.26	1869.36

Yeah, that's the SEC average ahead of Michigan, Miami, Oregon, and even Oklahoma. And nipping on Texas' heels in Expected Talent if not already past them in Actual Talent given their defections under Strong.

An absolutely loaded squad like Clemson is barely keeping up at recruiting against the middle tier of the SEC. About the only teams truly holding their own are Florida St, Ohio St, USC, and Notre Dame. And going into this season, the only team with more Expected Talent than LSU is Alabama and Ohio St.

Les Miles has slightly underachieved the past few seasons in regards to the talent on hand, but not as dramatically as his detractors insist. However, this is the season in which LSU is back on relatively even footing with Alabama. You're not going to catch them in talent, but the talent gap between the two is more narrow this season than it has been in the 14-team SEC era.