Handicapping The NFL

Part 2: Assessing relative team strengths.

In Part one I talked about the basics of the scoring system and how it impacts on handicapping. In part two I'm going to stay with keeping it simple and talk about the basics of comparing one team with another. It's all very well me saying that if the two teams are about level then it's a 3 point home win, but how is a newcomer to the sport supposed to have a clue where to start when doing that?
I'll try to address some common errors to avoid and also point you in the right direction.

Previous scores.

It makes sense that any system you try to devise to predict future results would involve results from the past. Sadly the results of the weeks before are actually quite poor predictors of next week.
That's probably obvious to many of you but a bit worrying for anyone looking to learn this from scratch, so here's an example of why. I'm going to look at two ten point win scenarios. Both of which are common.
 
Scenario 1: The strong road team go 24-0 up by halftime. To protect their starters and also their lead, they revert to many second string players and a running game that involves picking up 3 yards a carry and taking up 40 seconds a play. They make a few first downs but end up punting several times too. On Defence they go to the prevent, designed to not give up a big play and more time coming off the clock. The home team score but slowly and the road team win 27-17 comfortably.
 
Scenario 2:  Home team leads 20-17 in a great match with 30 seconds to go. The opponents have 3rd and 15 with 40 yards to go for a TD and throw a Hail Mary. It's picked off, and even though he should kneel, the defender sees the open field and the home crowd and runs back the length of the field for a Touchdown against an offensive unit who already know they've now lost the game.
Final score 27-17.
 
Obviously there are more normal 27-17s in the middle of these two, but neither of the above are anything like rare occurrences and if you're basing your assessment on score you're not making the most of the data you have.

Other statistics.

Other stats are similarly influenced by game situations. A team a long way behind can rack up passing yards. A team a long way in front will get poor yards per carry rushing. An injury to a key offensive linesman can lead to more sacks than usual. The list is endless.
I don't want to dismiss the importance of stats because if you're armed with a computer, a huge play by play database and a very methodical approach, you can work out the correlation between various statistics and the outcome of the game. As this is something I haven't done, I wouldn't want to take guesses beyond what I know from personal experience. What I do know, is that without this kind of approach, it is easy to put too much store in statistics. Use them as a guide, but don't swear by them.
There are also sites out there that work out the value of a play and adjust for the situation in which the play occurred. Getting two yards on a 4th and 2 would be treated differently to getting the same on a first down. As with all statistics, this information is useful if you keep it in context.
One well known such site had the Falcons ranked highly in Pass offense last season. Although I like Matt Ryan, the stats weren't so much a reflection of his ability, as a reflection of how teams stacked against the strong Atlanta run which made things a little easier on Ryan. Again, its valuable information but also a reminder not to take anything too literally.

What is certain is that computer simulations have been written that are incredibly accurate. What is equally certain is that unless you write one yourself, you're not getting your hands on them any time soon. If you do embark on such a project, you will need to make sure that every assumption you make is accurate and that you don't neglect something important. You'll also need to do it better than anyone else trying to do the same thing. In short, if you have that kind of brain and computing power, there are probably less well researched sports to apply it to.

Thankfully there are other ways to work out who the best team may be. Sadly they take some time. As with last week, I'll go through how I do things and you should take out of it what you can and adapt it to suit your own needs.

Research and judgement.

After each round of games I try to read every game report written by the AP (Associated Press). They're the reports that most people will be reading and I feel it's important to see what the general perception will be. After a while, by comparing what you see to what is written, you'll get a feel for what actually happened.
Peyton Manning, for example, never seems to have an average game where he played okay without being spectacular. Some random defensive guy who scored a touchdown will be the new big thing and so on. When you manage to translate AP speak into real speak, you'll find it's actually not a total waste of time reading the reports.

However, the bulk of my notes come from fantasy football sites. These people have a vested interest in avoiding the hyperbole that comes with running a news site and run their business around providing accurate future predictions and past write-ups.
Fan sites for the teams involved can also be useful, although they tend to have polarised opinion and you'll have to factor that in for yourself.
Having read as much as you can stomach, you can now make notes on the players and teams for a given week based on what other people have written. You can throw in your own observations if you watched the games.
Below is a picture of how I organise my notes.



My folder is organised by division, and then four games per page within each team. Each week I note the score and should likely note the handicap underneath too, but I have that elsewhere as you'll see another time.
I start the note taking by quickly checking through the previous games to make sure there's no theme to keep an extra eye on. Last season for example would have been the improvement of Cassel and keeping an eye on Peyton's injury comeback.
With that in mind I make a note on the QB for how he performed this week, I note the stats above his name and try to note what kind of style he played (basically how many risks he took, was he given a free reign?) and how well he did it. How well he made decisions, particularly in younger players, is incredibly important to me. In this section I'll also note anything important about WRs, particularly how new players to the team forge a relationship with the QB.

Sometimes I'll put a note in brackets that there wasn't a consensus about how well someone played, or that I disagreed with the assessments I read. I never presume to know more than anyone else about the game though, so I still put both sides of the argument to look back on later in the season.

Underneath this 1-2 lines, I'll do the same for the primary RB and O-line. I’m looking for how hard he ran and how much help he got. I find this as helpful as how well he actually did; as each team will approach stopping him in a different way, but if he's always running his hardest then I can factor him in properly against any forthcoming opponents. If there's an important secondary back, I'll note him as well.

Next come stats for the Rush D and Pass D, and then any notes.

In the picture attached you may notice some scrawl under the Panthers games. I was following the false starts given up by the team as their rookie center had admitted to problems. I wanted to know if it was worse in noisier road games (it was). You should also try to note important injuries and significant game plan changes.


After all this reading and writing/typing (I assume most people will use a computer to store notes. I find that using pen and paper helps me to not get lazy and just copy paste chunks of other people's work, and that I remember better because of it... even if I do just copy someone else's work down from time to time) you should now have a great idea about how each game played out.
I don't find I have to do this, but you could consider putting your own score/rating in brackets under the real score. For instance, in Scenario 1 at the top of this article you might want to call the score "20 points better" and in Scenario 2 you might call it "Equal teams". If you do this, remember to subtract home advantage in a similar manner to how you added it on last week to generate lines. If it seems like the home team was 3 points better, call it even. You get the idea. I hope.

After a few games you should be able to put the teams in a relative order of strength. If you had Team A to be about equal with Team B, this will help you to judge where they stand against common opponents. The more time you spend on reading and watching and note taking, the better you'll be able to do this.

Combine this with the beginner's system for handicapping games and you'll hopefully find that already you're getting your predictions within a handful of points of the actual line.

I'll be back in two weeks time with some ways to further refine your lines and some other factors to consider.
Unless you all laugh at my handwriting that is.

Lori