Saturday, June 09, 2007

"The Chris Allen Football League: Playoff Preview"

Skip to the random thought

I don't even remember how long ago I started the Chris Allen Football League season, but finally, I've finished the regular season, and it's playoff time!

First, before I talk about the playoff matchups, some words about the progress of my "random number football" algorithm. I think in my previous post about the CAFL, I said that the ratings system I implemented was "unstable". The better you are, the better ratings you get, and then it accumulates, as long as random luck doesn't push you back in the other direction (which, it could, but has less than a 50% chance of doing so). Well, I wasn't exactly correct about the rating system being unstable. Near the mean, the ratings are unstable. But towards the edges of my predefined range of acceptable ratings (necessary so that teams don't keep progressing towards infinity), the system is actually a little more stable. That's because I didn't do a good job of implementing the impact of extreme ratings. Because of that, it's hard to maintain extreme ratings. I suppose somewhere in the middle, there's a point of neutral stability. I'm not sure where the "tropopause" is in my range of ratings (-10 to +10), but instead of waiting around trying to find out, I'm implementing a minor tweak in the system to help the ratings hold true towards the edges. For the pass ratings, I had been using a fixed sack rate (7 sacks every 100 pass attempts) that wasn't affected by the ratings, but now I will change the sack rate as ratings progress towards the limits. (For example, teams with the minimum rating advantage of -10 will have 14 sacks every 100 pass attempts, while teams with the maximum advantage of +10 will have no sacks whatsoever. For the record, a +10 advantage is not realized very often, but it has happened a couple of times in the league this season. And, I don't think I've seen a -10 advantage yet.) I don't know of a convenient way to tweak the run ratings, but I've learned that in my algorithm, the running game doesn't really matter that much anyway.

Now, onto the league results and the playoff "preview". Let's start in the Eastern Conference, due to my East Coast bias. The playoffs have an identical format to the NFL - 12 teams, 6 from each conference, 2 teams in each conference (the top two division winners) get a first-round bye. Here are the seedings, along with their ratings: (seed, team, record, pass offense, run offense, pass defense, run defense)
#1 - Greensboro, 14-2 (+3.6, +6.2, +1.3, +2.2)
#2 - Harrisburg, 14-2 (+2.7, +4.6, -3.0, +4.5)
#3 - Birmingham, 12-4 (-2.2, -1.0, +1.4, -1.9)
#4 - Scranton/Wilkes-Barre, 12-4 (+3.6, +2.0, +2.1, +0.9)
#5 - Raleigh, 11-5 (+4.6, +2.4, -0.4, +0.7)
#6 - Dayton, 8-8 (+0.2, +6.7, +2.0, +1.1)

Greensboro is certainly the best team, and they earned their #1 seed. I'd consider them the favorite, although Scranton appears to be their biggest threat. Birmingham didn't have a very good finish to the season, and they lost all of their momentum. One thing about the ratings is that they only consider each team's last five games. So while Birmingham had an excellent middle of the season, they haven't been so good over their last five games. So, they'll find themselves to be the underdog in their first round matchup against Dayton. Also, it should be noted that the ratings will change during the playoffs, so it's not only important that teams win, they need to win impressively so that they can maintain their ratings.

Now, the Western Conference seedings:
#1 - Little Rock, 16-0 (+2.9, +0.6, +4.6, -3.6)
#2 - Las Vegas, 12-4 (+3.8, -0.4, +2.5, +1.3)
#3 - Fresno, 11-5 (-3.8, +1.9, +3.2, +2.5)
#4 - Milwaukee, 11-5 (-5.0, +2.6, +3.7, +4.7)
#5 - Salt Lake, 12-4 (-0.2, -0.7, +1.8, +0.6)
#6 - Los Angeles, 11-5 (+7.7, -4.4, +2.6, -0.3)

Little Rock made it the entire season undefeated. Amazing! However, their ratings aren't that good. And, they had a weak schedule, only playing three other playoff teams all season. So, it's anything but a guarantee that they'll make the championship game (which I'm calling the "Big Bowl"), even though every team in the West has a weakness. However, given the importance of the passing game, I have to give a nod to Las Vegas and Los Angeles. Los Angeles currently has the best passing rating in the league. However, their best games were their 4th- and 5th-most recent games, so they will lose their high rating if they don't produce in their first two games. And it will be difficult to do so, because they'll face a good pass defense in their first game (v. Fresno), and if they get past them, they'll face Little Rock in the next round, who currently has the best passing defense in the league. So, Little Rock's ratings don't look that great on the surface, but a good pass defense is hard to beat. Maybe they will get it done. Or, maybe not. It all depends on the random numbers. That 16-0 record doesn't lie. The fact that a team was able to go 16-0 definitely lends credit to my algorithm and my ratings system, particularly if they don't win the "Big Bowl", which would demonstrate that there are obvious tendencies, but no obvious guarantees, which to me is the perfect system. (Also, in case you're wondering, nobody went 0-16. There was one 1-15, though.)

I'll post again once I finish the playoffs, which will probably take no longer than a week, because there are fewer playoff games total (11) than games in a single week of the regular season (16). Oh, the excitement!

Today's random thought:

- Amber and I played mini-golf at the local Putt-Putt this past week, and I had what was perhaps my best round of mini-golf ever - a round of 34. And, I got a hole-in-one on four of the first six holes - I know I've never done that before. Granted, the Putt-Putt courses are quite easy, and much easier than other mini-golf courses, but still. That's part of what makes them fun, actually - they're simple, and a hole-in-one is within reach on every hole. I don't care what kind of mini-golf course it is; when you average less than 2 shots per hole, you're doing something right.

No comments: