Strength of Schedule

**teecee** · 12-19-22, 10:41 AM

Here is an example.

I had Niagara yesterday at -3, or -3.5. My numbers showed that Niagara had a +3.5 off to defense efficiency rating, while NJIT stood at +10.5. Niagara was performing in this manner against a schedule that bested their opponent (NJIT) by over 12 points in efficiency. So, Niagara -3/-3.5 was a no-brainer, right? They never had a chance to cover that game. It's frustrating, bc I spend a fair amount of time on this stuff.

**teecee** · 12-19-22, 10:45 AM

Niagara had won 5 of 6 ats before that game, so maybe they had peaked for the time being and had become overvalued by method of handicapping. Their numbers had become inflated bc they have played so well recently. I'm just brainstorming here, bc yesterday was so bad.

**HeeluvaGuy** · 12-19-22, 11:55 AM

Couple of things that aren't a direct answer to your question, but nevertheless are relevant:

1) Teams don't perform consistently. If they did this would be pretty easy. Add to the fact that we're specifically talking about kids 18-22 years old and honestly it's amazing we ever get it right.
2) Any one game is pretty much meaningless in a season that has over 5,000 games.

You caught Niagara on their least efficient offensive performance of the season to date. A few days ago I lost an over because one of the teams was >20% below their worst efficiency of the season. Outliers happen.

That's why I find it very important to have a methodology that I can apply to every game and test for accuracy. The result is that I can look at the overall performance (in this case of my model that predicts every game) to understand that while it performs pretty well overall, any one game might miss wildly.

**teecee** · 12-19-22, 01:07 PM

Right. I appreciate all your responses. I get it is a one-off (or could be), but my line was so far off the line offered, I'm wondering (aloud, yes, and not expecting anyone to solve the issue for me) if it is a case where they (Niagara, for example) had reached sort of a peak and should be left alone. Unfortunately, I was so excited about yesterday, because I had a number of plays that easily qualified, but maybe a lot of the teams I played had "peaked" for my method. I went 3-14-1 and a number of the teams I played shot absolutely wretchedly. A good number of the so-so plays (ones I left alone) killed it.
Again, I was hoping for some of your valuable input. Thank you for chiming in.

**HeeluvaGuy** · 12-19-22, 01:48 PM

I think you use kenpom, right? If you're using his adjusted offense and defense, you probably don't need a separate SoS adjustment too since those numbers are based on performance vs an average team. In other words, you might be double counting the adjustment. Just a thought.

I took a quick look at the box score and Niagara was 1-16 from three. They shoot around 30% on the season. If they had hit their average they would have had at least 9 more points and would have had a chance to cover. Predicting shooting percentages is the most difficult part of handicapping these games IMO. I didn't see the game and am just looking at numbers, but NJIT's 3-point defense is actually pretty good and may have been a factor in the bad performance.