My goal is to get more and more people interested in these weird stats that I've been using for the 'Beyond the Box Score' bits, and while I've explained these stats in different places at different times, before I post this week's BTBS Preview of Mizzou-Tech, I figured I could walk through the stats and explain them all in one place...and then link to this post in each future post.
Click 'Full Story' for more.
Update [2007-10-18 21:57:48 by The Boy]:See bottom of post.
Most of these stats are culled directly from Football Outsiders. Some I've modified, some I've created out of thin air.
Success Rate is the standard for a lot of the stats I use. I've modified it slightly so that there's roughly a 44% chance of 'success' on any given play. Here are the rules according to down:
1st Down: 50% of necessary yardage. If it's 1st-and-10, you need 5 yards for 'success'. Football Outsiders use 40% for 1st down, but with the games I've entered, that led to a 1st down success rate of about 51%. Bumping the requirements to 50% led to the 44% rate for which I was aiming.
2nd Down: 65% of necessary yardage (rounded up to the nearest yard, of course). If it's 2nd-and-10, you need 7 yards for 'success'. 2nd-and-15? 10 yards. This makes sense, really, because to succeed regularly on 3rd downs, you need to stay at 3rd-and-5 or less. Getting most of the way there on 2nd downs sets you up infinitely better for 3rd down.
Football Outsiders uses 70%, but the success rate for that was around 42%. Weakening the requirements slightly got me into the range I was looking for.
3rd and 4th Downs: 100% of necessary yardage. I figure this requires no explanation.
The ways I've been using Success Rate to date are pretty widespread. I can look at a team's Success Rate by quarter or down or game scenario (ahead, behind, tied, close, etc.), and it tells a pretty interesting story. I'm probably not dealing with enough of a sample size to draw any serious conclusions, but a team's trends for Success Rates don't tend to stray drastically from game to game. To understand what I mean, look at the OU-Mizzou game. Coming into the game, OU's strongest quarter was Q1. In Q1 of OU-MU, OU held a 62.5%-35.7% advantage. Meanwhile, Q3 has been Mizzou's strength all season, and sure enough, they held a 64.0%-46.7% advantage in Q3.
Success Rate doesn't lead to points--points are just as much determined by lucky bounces, turnovers, clutch individual plays, etc. However, Success Rate can tell you how a game will flow, and who will probably be creating better opportunities in specific situations, and over the long haul that can probably tell you who's winning and who's losing. I like this stat a lot, which is pretty obvious considering how much I use it.
Other uses of Success Rate: QB Success Rate (how a QB directs the offense as a whole, counting both runs and passes), Run Success Rate, Receiver Success Rate, etc.
Defensive Success Rate
This one's pretty easy: if it's a 'successful' play for the offense, it's an 'unsuccessful' play for the defense. Defensive Success Rates are simply the inverse of Success Rates. I break this out by position/unit a lot--looking at DL, LB, and DB success rates.
% of Success
I haven't used this in a while, but I occasionally break out the "% of Success" figure. I should use it more, actually. This is a pretty good gauge of explosiveness. On 3rd-and-7, gaining 7 yards = success. So does gaining 26. However, 7 yards is 100% of success, while 26 is 371%.
On the flipside, 6 yards = non-success, as does 1 yard. However, 6 yards = 86% of success, while 1 is just 14%. Get it? If a WR averages 110% of success, that's pretty damn good. However, if he averages 220%, that's way better. This gets into more detail than simple success rate.
The one trick here is, if you gain 50 yards on 3rd-and-1, that's 5000% of success. One big gain on a short-yardage situation, and you've skewed your numbers for the entire season. So I capped the success % at 1000% for any given play. 1000% is still pretty damn good, but it doesn't affect an overall average nearly as much.
There are plenty of stats to measure the effectiveness of a QB, RB, or WR/TE. Even defensive players have quite a few to go by--tackles, sacks, defensive success rate, etc. However, the play of the O-line or D-line as a whole goes a long way toward determining the outcome of a game, and there really aren't many measures for this. One of the better ones, particularly for rushing plays, is Line Yards.
Here are the rules:
• For a play that resulted in negative yards, the O-line is given 120% of the effort (i.e. a 3-yard loss would be a 3.6-yard loss for the O-line).
• For a play that resulted in a 0-4 yard gain, the O-line is granted 100%.
• For a play that resulted in a 5-10 yard gain, the O-line is granted 50% of the yards over 4 (i.e. an 8-yard gain would be a 6-yard gain for the O-line).
• For a play that resulted in a 10+ yard gain, the O-line get no extra credit—by that point, the runner is into the secondary, and the line won’t get much chance to block. Therefore (if the math in my head is correct), the most credit an O-line can get is 7.5 yards.
You get the idea here. If a runner gets 10 yards or 50 yards, the O-line did its job equally well because at some point it's all on the runner. Linemen are a bit too big and slow to follow all the way down the field.
My own theory is, this works a lot better for runs than passes since, obviously, the O-line isn't allowed to run down the field and block for receivers until the ball is actually thrown, and in most cases (some screen passes being the obvious exception) the ball is thrown pretty far away from the line. So I use Line Yards to judge run blocking and...
...to judge pass blocking. Both of these measures are flawed--if you have a quick, elusive QB who's good at escaping a collapsing pocket, you're probably going to have a lower sack rate than if you have a sedentary guy back there. However, this measure is as good as anything else.
This one's pretty easy:
Sack Rate = (sacks) / (sacks + passes).
There you go. One thing to keep in mind here is that, as a whole, there's a higher sack rate on 3rd down (more likely to be obvious passing situations, and therefore more likely to see a blitz of some sort) than on 1st or 2nd. That isn't the case for Missouri in 2007, or other pass-happy teams like Texas Tech, but they're the exception. So I measure 1st-2nd Down Sack Rate separately from 3rd-4th Down Sack Rate. In 2006, the 1st-2nd Down Sack Rate was around 5.5% in the Big 12, whereas the 3rd-4th Down Sack Rate was about 8.1%.
Each turnover is assigned two values: 1) the point value (see below) of the offense's field position at the time of the turnover, and 2) the point value of the resulting starting field position for the opposition.
Turnover Costliness = (0.75*the higher of the two values)+(0.25*the lower of the two).
I previously had a factor in here regarding closeness of the game, and I'm sure I will again, but for now this is what I'm working with.
So there you have it. As I come up with new stats and measures, I'll probably just add them to this post. This way I have a specific reference for all future BTBS posts. If this interests you at all, or if you come up with other things I can look at, feel free to comment below. I'm all ears.
First Update, 10/18: So this is pretty cool.
This is the average points per possession for a play taking place at a given yard line. Possessions with a play at your own 1-yard line average about 0.9 points per possession. Possessions with a play at your opponents' 1-yard line average about 6.0 points per possession. Your likelihood of scoring doesn't go up much anywhere between your goal line and 20, but it goes up quite a bit between your 30 and 40...and then again between your opponents' 30 and 40. The slope increases the closer you get to the goalline, which makes sense really. Either way, this is a new toy for me to play with. You've been warned.
Second Update, 10/29: I have tinkered with Turnover Costliness and have made some changes to the original. See above. Also, I've added the following category: Points Per Play.
Points Per Play comes from assigning a point value to each run/pass/penalty according to a scale made from the chart above. You can see from the chart which gains would be worth more than others.