Calculating Ravens odds of winning next six games

Baltimore enters the bye at a crucial point in the season. Sitting at 6-4 lacking most meaningful tiebreakers (so far), the Ravens are set to play three games against quality opponents that stand to shape their season. Fail to win at least one of them and they will be sitting at home in January.

While the New Orleans game is probably the least important of the remaining six games (and to call any game unimportant is highly relative), the following contests against Miami and San Diego have significant implications.

Teams are often not what their record says they are

Bill Parcells popularized a euphemism that you are what your record says you are. While I understand the saying and I agree that's the right mindset for a coach to take, I would disagree with it as a general rule for the rest of us.

If I've learned anything from following the NFL it's that most games are decided in the fourth quarter and often by a single score margin.  Because of that, strange or even random events can often dictate the outcome. I'm not really into making game predictions for that reason. Even armed with a plethora of statistical and eye-test evidence, there's just no way to predict how a game will go and certainly not the score. The ones who are good at such things mostly work in Las Vegas for a reason where their job is to make their employer money not themselves. The NFL, like gambling, is an exercise in probability, not certainty.

However, it so happens we have a probability-based tool at our disposal to estimate what our chances are of winning each of our next six games.

Log5 Method

The method, known as the Log5 method, was developed by Bill James, the titan of sabermetrics made famous by Moneyball. One of my favorite professional sports writers, Bill Barnwell, employed the method recently to estimate the chances that Oakland would go 0-16. It was compelling.

While the method was developed originally for baseball, it has merits in football as well. Instead of using raw win percentage though, we use a stat I've written about regularly before called the Pythagorean Expectation, which is a foundational concept in sports analytics. This statistic uses a team's points scored and points allowed to estimate how many wins it should have.

Remember Bill Parcells? Yeah, he wouldn't like the Pythagorean Expectation, at least not publicly.  Separating a team's record from its actual performance may feel like an academic exercise but it is instructive to do so.  For instance, are all 10-6 teams created equal?  Absolutely not. Yet, we often think they are without context for how each team played and the quality of their opposition.

Calculating Ravens Odds of winning next six games

Following is a table employing the Log5 method. The calculations also incorporate home field advantage for each team in the third column, which tends to matter in the NFL.

Again, keep in mind this is using a team's expected win percentage — not its actual one. Baltimore has the highest win percentage (70%) because it was expected to have won 7.04 games using Pythagoras. We have a good example of why this matters in Miami and Cleveland. Cleveland is 6-3 with a higher winning percentage than 5-4 Miami. However, the Dolphins are operating with a top 5 defense while Cleveland has no unit higher than 14^th in the league so far. Miami has two blown leads in the final minutes of the fourth quarter against high quality teams while Cleveland has pulled those games out except for their loss to us.  The evidence suggests Miami is a slightly better team despite their record. Frankly, they did deserve to lose those two games to some degree with bad coaching by Joe Philbin and his use of timeouts, but they were still pretty unlucky to do so.

Thus, before accounting for HFA, we are given a 54% chance to beat Miami but 60% to beat Cleveland. Now incorporating HFA, Miami becomes slightly more favored and vice versa for Cleveland.

Taken together, this method gives us a 7.5% chance of winning out.  Not exactly a likely outcome but I'm sure few of us were expecting us to do that anyway after our most recent three games. Of course, we were given a 3% chance of winning it all in 2012 at the start of the playoffs. So, unlikely outcomes happen in the NFL. Don't rule it out but I wouldn't go to Vegas with that bet either.

If I was making the prediction though, I'd probably say the odds overstate our chances against New Orleans.  The Saints are one of the most successful home primetime teams in the league. It feels like the 2011 Chargers game where we could quite easily get beat up pretty bad early by a very good passing team at their house even if they may not have the performance to back it up to date. But we'll see.

Beyond that, the numbers say Miami is our toughest remaining game. This is a very important game for so many reasons, as is the Chargers game.  All three of us will be playing for their season and we need all the conference wins we can get in the AFC.

Of course, let's not forget about Cleveland in Week 17. It is entirely possible one of us has nothing to play for but it very much may decide the fate of both teams. Cleveland has been feisty this year and I'm glad it's at home.