When reliance on metrics go too far ( Clemson 44 Alabama 16 )

You have to give it to Dabo Swinney and his 2018 Clemson team credit. Their dominant 44 – 16 national championship win over Alabama, 6 months later, still seems incomprehensible to some. The Alabama players didn’t seem to understand Clemson was just better. For the SEC media, Clemson has become their worst nightmare.    Now we have advanced metric analysts trying to essence explain away a 28 point victory while trying to convince us that Alabama really was better. It’s just as mind-boggling as it sounds.

I’m a numbers guy. I’m engineer by day, with 2 degrees in the discipline. I’m all about metrics and data. That said there are times when my eyes tell me the answer regardless of what the data says.

The culprit in this case is @SBN_BillC (Bill Connelly). Bill seems like a bright guy. I’m sure he is. He has some great metrics. I’ve read them and have probably referenced them in a article or tweet somewhere, but he’s wrong on this one.

Add all this up, and Clemson’s adjusted scoring margin in this game was 4.9, which is to say that S&P+ views this game, with these primary stats, as a five-point Clemson win. Not 28.

— Bill Connelly (@SBN_BillC) June 4, 2019

Let’s revisit the advanced box score:

Succ rate: Bama 56%, Clemson 41%
SD succ rate: Bama 58%, Clemson 42%
FP: Bama 29.3, Clemson 22.9
Scoring opps: Clemson 7, Bama 6

Those are major predictors of victory, and it suggests something between “Bama by a little” and “Bama by a lot” pic.twitter.com/7LFrTh3QyH

— Bill Connelly (@SBN_BillC) June 4, 2019

To be sure, the ACC’s hotgarbageness meant Clemson was unchallenged after Syracuse and got to keep plenty of clubs in the bag, so to speak. I assume they were saving stuff for Bama starting in, like, October.

— Bill Connelly (@SBN_BillC) June 4, 2019

Because of the season-long results, and because this was more like a 5-point game on paper than 28, I’d have maybe still picked Bama to win in 6 or 7 in a best-of-7 series.

ugh, why do I do this to myself delete tweet delete tweet

— Bill Connelly (@SBN_BillC) June 4, 2019

I don’t even know where to begin with this. I really don’t. It’s fine that you’ve used metrics as a score predictor, but then after the fact you double down and still predict Alabama to win 6 or 7 games in a 7 game series to defend your metrics.

Some of Clemson’s yearlong metrics are based on a different quarterback playing, and the development of QB Trevor Lawrence.  If you couldn’t see Trevor Lawrence was a better by the end of the year than when he took over, then what can anyone tell you? You’re going to say it’s clear and sunny when it’s raining outside.  You just aren’t going to believe what your eyes tell you.

Here’s what I saw – when Alabama began facing competent passing attacks their secondary was exposed. When Tua Tagovailoa faced top 30 defenses, his statistics were significantly worse.  

Even I’ll admit Clemson wouldn’t beat Alabama by 28 every time they played, because in some of those games it would have been more than 28. Clemson backed off with subs in the 4th quarter and went to primarily running the ball. It could have far worse. Now I do think if they played 7 times, Alabama wouldn’t go winless. Alabama was very good, but they weren’t better than Clemson.

I know this because the only statistic that matters was 44-16.

Make sure you follow the All Sports Discussion Twitter account at @AllSportsDACC and please like our Facebook Page