Week 15 Recap - An Autopsy

There are no excuses.

What will go down as perhaps the worst total week in Owl Eats Football, Week 15 was an unmitigated disaster.

But it wasn’t so bad that we can’t learn from it. This isn’t a point where we should lose hope, but rather a point where we should get to work. Failure is just data, and by processing and analyzing that data we can more accurately predict where it might occur in the future.

So, the point of making this an autopsy is not to make excuses, but rather to identify reasons. Consider this an active learning exercise as we grow and shape our plan for WORMS heading into the off season.

With that being said, here is the breakdown of place we can look to find out more about a terrible, horrible, rotten week.

Finding #1 - Variance

The dirty little secret of gambling is that it is a completely insane proposition.

You are betting on the outcomes of games played by people with entirely different objectives than you. While you may be confident that a team will win by six points, that team doesn’t really give a flip (pardon my language) how many they win by.

Coaches don’t coach to spreads, players don’t play to spreads, standings don’t care about spreads.

And so, on a Monday like today, it can feel futile to place money on something that the participants don’t even consider. Imagine betting on whether your Starbucks barista spells your name right.

This week we had two examples of this costing us in Best Bets and WORMS alike.

Anyone (like me) who had money on Washington anywhere from +4.5 to +6 felt this in real time. Up 3 with 36 seconds left, it seemed like a lock for this bet to cash. The Eagles scored to go up 4, but again that was within our realm of outcomes and we still felt good.

Then, on the last play of the game, Dwayne Haskins fumbled and the Eagles returned it for a touchdown with 0 seconds on the clock. It wasn’t a good football play (he should have gone down to avoid re-fumbling) and it was a terrible gambling loss, but that’s the point. The players are not playing against the spread, they are playing to outscore their opponent by as much as possible. This bounce of the ball cost us a win in Best Bets, a win in WORMS, and plenty of money.

The other example was the Seattle - Carolina over. We here at OEF were invested in this game being under 50 for the reason that Seattle likes to get leads and take the air out of the ball and Carolina has a non-functioning offense outside of Christian McCaffrey. Both of those things came to fruition, but we still lost.

Heading into the fourth quarter, we were looking and feeling good as the Seahawks had a 23-10 lead and the game had only had six points since halftime. Then, half of the fourth quarter went by and no more points had been scored, putting us in position to feel very confident. A Chris Carson touchdown made it 30-10 with seven minutes left, but Carolina had done nothing all day and Seattle wouldn’t be pushing the envelope.

Well, the Panthers proceeded to drive 75 yards and score a touchdown, Seattle went three and out in 28 seconds, and the Panthers drove 84 yards for their second touchdown in less than two minutes. Seattle recovers the onside kick, kneels, and wins 30-24 while we lose our under.

The end sequence is no sweat off Seattle’s back as they get the win they needed, but it costs us.

And the point of all of this is not to complain or ask for sympathy, but rather to identify something in the autopsy of this week that we should be cognizant of moving forward. Variance can swing picks made on arbitrary lines that are of no consequence to the participants. Our job is to minimize this and find the areas where it would take the most variance to stop our bet.

We will never be in a place where we beat variance, we just want as few weeks as possible where it costs us big.

Finding #2 - The Elusiveness of Consistency

One of the biggest problems with having a statistically based algorithm is that it is impossible to account for things outside the numbers.

In the case of this week, there was plenty of evidence piled up that some of our picks were correct, but real life has a way of intervening and rendering statistics just a part of the puzzle.

Perhaps the perfect example of this was the San Francisco-Atlanta game. Coming off of a huge win against the Saints and heading into huge games against the Rams and then the Seahawks, San Francisco hosted Atlanta, a team that was 4-9 and next to teams like Washington and Carolina in the standings.

But, as has been proved for decades, consistency is the hardest thing to accomplish in football.

San Francisco lost this game at home to a bad team, which the numbers do not support. In real life, though, they are coming off one of the hardest stretches in NFL history, where they have played the Packers, Ravens, and Saints all in a row (a combined record of 33-8) and came away with two wins.

The long and short of it is that this was a letdown game, and there are n numbers to capture that.

And so, another finding of this autopsy is that while numbers seek to find patterns and consistency, the NFL is a breeding ground for inconsistency.

Last year the Rams were in the NFC championship, now they will miss the playoffs. Two years ago the Jags were one play away from making the Super Bowl, now they will pick in the top ten.

Putting this inconsistency into numbers (ironically) let’s look at longest consecutive playoff appearance streaks (including already clinched spots this year). Number one is New England at 11. Number two is Kansas City at five. Number three is New Orleans at three. No one else is higher than two.

If we raise the bar and look at consecutive 11-win seasons here is the list: The Patriots at 11 consecutive seasons, the Rams and Saints at two consecutive seasons (with the Saints able to extend that to three with another win), and no one else more than one.

Given that it is so hard for teams to be consistent year in and year out or even game in and game out, finding a way to capture these improbabilities can be difficult. We are getting closer, but this is the biggest challenge in creating a prediction system.

We might currently be the Saints, but we’re trying to be the Patriots.

Finding #3 - Growing Pains

This may be the biggest and perhaps the most difficult for readers to swallow.

Despite the fact that WORMS was backtested off line for a while and was fairly successful, the product itself is less than one season old. Because of this we don’t have a fully developed adult on our hands, we have a toddler who can be irrational and petulant.

Because of this, WORMS is not as good as it will be. And the hope is that weeks like this can be attributed to an underdeveloped vision rather than a false prophet.

So, moving forward, WORMS is still our prize possession and our most exciting element, but it is not what it will be quite yet. The Night Owl and I have big plans in the off season and the future for what this can be, but just because a toddler likes playing with trucks doesn’t mean you should buy them a Maserati.

The future is bright, but don’t let it blind you to the present. We are in a place where you can trust WORMS to deliver actionable information, but perhaps not one where you are going to double your life savings.


13 views0 comments

Recent Posts

See All