Use Statistics Responsibly When Discussing Alabama and Oklahoma State
by Year2 on Jan 3, 2012 4:41 PM EST
in SEC Football
Email
Print"He uses statistics as a drunken man uses lamp-posts... for support rather than illumination." -Andrew Lang
I finally have begun getting into the massive college football data dump from last summer both to try to discover new things and judge whether the methods I currently tend to use are indeed the right ones. I'll give you an example.
I tend to use averages a lot, as does most everyone who works with sports statistics. Averages work best if what you're taking averages of has a normal distribution. If the distribution is skewed to any great degree, then using an average is as helpful as you might think.
So for starters, I plotted out the frequency for ranges of points scored in I-A vs. I-A games from 1998-2010. This is what you get for that:
It kind of looks like a bell curve, but it's awfully jagged and kind of leans to one side. Those variations from what you see in a classic normal distribution could have important implications for the study of football.
They also could be the side effect of a poorly constructed graph. If you look at the X-axis, the point ranges are in groups of five. While that might be good for, say, basketball, it's not good for football. Points are most often scored in groups of three or seven in football, and two field goals are together almost seven. Here is a better graph where point frequencies are grouped by seven instead of five:
That's much better. The scale of the graph better matches the underlying activity that produced the numbers, and we see a distribution that looks a heck of a lot more like a normal distribution. Taking about average points per game scored or allowed does indeed make sense given that the distribution of points scored is roughly bell-shaped.
That conclusion is not the only important one here, though. You can see that the left side of the mean has a much steeper slope than the right and also that the boundary of zero is quite important. I think it's fair to say looking at this distribution that, if it was possible to finish a game with negative points, a noticeable amount of games would end up with at least one team below zero on the scoreboard.
That point is important because of the fact that we often try to compare the best offenses to the best defenses. There is no upper boundary to the number of points an offense can score. In theory, two teams can go on scoring forever as long as they match each other point-for-point in overtime periods. On the flip side, a defense cannot by rule allow fewer than zero points no matter how well it plays. The zero bound is a very real factor there.
I am not a good enough mathematician to come up with a nice, elegant solution for comparing offenses and defenses on equal grounds (though some people have made good tries to that end). However, it's an important distinction to bear in mind when going through this season's most contentious debate regarding the relative merits of Alabama and its fierce defense versus Oklahoma State and its explosive offense.
For that matter, points scored and points allowed measure two completely different things. There are obviously some mitigating and confounding factors, but generally speaking, points scored measures how often and to what degree an offense succeeds whereas points allowed measures how often a defense fails. Success rates and failure rates are two different things.
In the coming days, the Bama vs. Oklahoma State fight is going to be fought here and there around the world of sports media. It will only get more intense if Alabama defeats LSU in the BCS National Championship Game.
If you plan to take and argue a side, I plead with you to use statistics responsibly.
30 comments
|
0 recs |
Do you like this story?
Comments
Enjoyed the article
2nd graph isn’t any more normal, just a bit smoother. The curve would be better approximated as Poisson rather than Gaussian (“normal”). My guess is that individual (team) points allowed follows a similar distribution. What you would want to do then is compare rates of team points scored/allowed.
Then again, nobody really likes statisticians anyway.
What are you glorifying with your life?
OSU has only 1 argument to make
that is they haven’t yet lost to LSU. When you consider pretty much every other argument, with or without statistics, you come up with Bama having the stronger argument.
"Difficulties strengthen the mind, as labor does the body."
― Seneca
Actually they have at least 2
They haven’t lost to LSU. And they won a stronger conference outright playing every team whereas Alabama did not win their division and escaped the two best teams from the opposite division.
Others may focus on additional points (for instance, strength of schedule) but I think the above two cover it pretty thoroughly with the first being much less important than the second. For instance, had Oregon beaten USC, and otherwise played to the exact same result, I would have considered their argument better than Alabama’s because of the second factor. You and I argued this point ad nauseum in a different post, however, so there’s no sense in rehashing it here.
In the case of a Bama victory, it will be interesting to note how much, if at all, LSU outperforms them in non-scoreboard related categories as Stanford did against OkSt last night.
Let's take this argument you make
The won a stronger conference outright: Well, considering they didn’t play LSU and since Alabama lost to LSU and that is the sole reason Alabama didn’t win their division/conference, you are only restating the fact that Alabama did, in fact, lose to LSU. It’s not really Alabama’s fault they happened to play in the same division as the #1 team in the nation, was it?
"Difficulties strengthen the mind, as labor does the body."
― Seneca
Not only that but our scoring average isn't that far apart.
And the Big 12 teams have no defense so scoring points is that much easier in their conference.
by burmbuster on Jan 3, 2012 8:13 PM EST via iPhone app up reply actions
And offenses are much, much worse in the SEC
Really, the two best teams in the conference play and they can’t even manage a touchdown? Oregon scored three or four against LSU. WVU scored three. I’m sure if Alabama had played anyone of consequence out of conference similar totals would have been observed. (Penn St. is not a team of consequence.)
Alabama (pre-bowl game) has faced on average of 20 fewer plays on defense than has Oklahoma St (post bowl game). That’s certainly an indicator of its strength. It’s also an indicator that SEC teams really struggle to generate much offense. Which means if you have a pulse on offense, you’re at a real competitive advantage. And that’s borne out by Alabama’s average number of plays on offense, which was 10 more than what they faced on defense. OSU represents the converse, averaging 8 fewer offensive plays than defensive plays per game.
Had honestly forgotten about my comment up until now...sorry for the delay
It’s not Alabama’s fault that they played against the #1 team in the country, it’s Alabama’s fault that they weren’t (aren’t?) good enough to beat them (at home). As I said earlier, had they not lost to USC, I would have voted for Oregon to play LSU again above Alabama because of the conference championship. Moreover, we don’t know whether or not Alabama would have beaten UGA in the SECCG had they been their instead, though I agree they would have been favored.
Sometimes it’s just tough s*** that you can’t be the best in your conference. But we know as surely as we can that Alabama was not up to beating LSU. We have no idea whether or not Wisconsin, Stanford, or Oklahoma St is capable or incapable of such an achievement. Which is why the national championship game should not be a rematch of a conference game from the first weekend in November.
No prob on the delay
Again, you are restating the one argument which OSU has: Alabama lost to LSU. All other arguments are either happanstance (ie that Bama happens to share a conference with the best team in the nation which they effectively tied as closely as is possible to do), or strength of schedule. I do not deny that this is a powerful argument, but I do deny that any other reasonable argument can possibly be made, as every other fact points to Bama being a better team than OSU and by that logic, more deserving of a NC shot.
It all comes down to 2 views: Should a team which has lost to LSU get another shot at them or not? OSU’s best claim is the answer should be no, since they cannot reasonably say they played better over the whole season than Alabama. Bama’s best argument is that the best team should have a shot at #1, regardless of schedule. By this logic, Bama is pretty clearly the deserving claimant.
I get tired of all of the distractions between these two very simple questions. I posted this comment in “use stats responsibly” but I maybe should have made my own “use logic/debate responsibly” post, as to me all of the muddle is due to poor logic and debate skills on the part of most involved.
Of course ESPN and other want to intentially muddle the argument for their own purposes, but I am supposing that most true football fans want the most deserving teams to both play each other for the NC – by one logic, they are, by another, they are not.
"Difficulties strengthen the mind, as labor does the body."
― Seneca
OSU has (at least) two arguments
1: It won the strongest conference in the country outright
2: It compiled the same record as Alabama against a significantly more difficult schedule
3: Its loss did not come to the same team it is set to replay.
It all comes down to 2 views: Should a team which has lost to LSU get another shot at them or not? OSU’s best claim is the answer should be no, since they cannot reasonably say they played better over the whole season than Alabama. Bama’s best argument is that the best team should have a shot at #1, regardless of schedule
You want to focus on the final claim and say it’s okay because LSU is the #1 team and use that to simply wish away the first two. But who Alabama lost to is so much less important than the fact that it revealed they were not even the best in their conference. I don’t know what your stance was in 2007, but recall that many in the SEC vehemently argued against UGA being selected to face the Buckeyes specifically because they did not win their conference (or division) even though they were 3 (or maybe 4) spots higher in the rankings than LSU. If every other conference champion had 2+ losses then you might have something, but even then, why not Boise St? They smoked UGA about as convincingly as did LSU.
I do not deny that this is a powerful argument, but I do deny that any other reasonable argument can possibly be made, as every other fact points to Bama being a better team than OSU and by that logic, more deserving of a NC shot.
How on Earth do you know that? Alabama would probably be favored against OkSt, I admit that, but guess what, Clemson was favored over WVU last night and how did that work out? Maybe the statistical discrepancies between the two teams are simply the result of OSU playing so much more difficult a schedule?
For the last time, Alabama could not beat LSU at home. Why would you possibly think that some other location would give them a better chance to prove otherwise?
Of course ESPN and other want to intentially muddle the argument for their own purposes, but I am supposing that most true football fans want the most deserving teams to both play each other for the NC – by one logic, they are, by another, they are not.
Leaving aside the problem of the ‘true football fans’ statement (google the ‘no true Scotsman fallacy’), you are equating most deserving with ‘the best.’ Most deserving is reflected by what’s accomplished on the football field, in short, its resume. ‘The best’ is an airy notion that can hardly be proven but is frequently disproven, one way being by already having lost to the team you are scheduled to replay.
*Sigh*
Point 1) OSU Won the strongest conference in the country: This is first of all a strength of schedule argument, saying essentially that OSU played a harder schedule than Alabama. Maybe you can argue they did, maybe not, but most observers would say Alabama played a harder schedule.
2) They had the same record as Bama: ok, this is not an argument saying OSU is better, but it is worth noting, I suppose
3) Alabama lost to LSU – as I’ve noted, this is a good argument.
Of course, the only reason OSU was able to win their conference is because the one team the DID lose to sucked bad enough they were not a threat to win their conference even beating OSU. I suppose with similar logic, if Bama had beated LSU but lost to Vanderbilt, Alabama would be more deserving to play for the NC than they currently are, having lost to the #1 team in the country. This logic is, to my mind, poor. But then again I agree with most of the voters in the nation which ranked Bama ahead of OSU due to all the varabiles.
I suppose you have me in one thing, however. I should amend my statement that OSU has 2 arguments: Bama lost to LSU, and Bama didn’t lose to a team that couldn’t win their conference even beating the best team in it, as did OSU. I suppose if the roles were reversed, I might agree with your point.
As far as your last point, Alabama’s resume is better based on who they played and how well they played them than OSU. What they accomplished on the football field is, at the end of the day, more impressive than what OSU did. That’s the only reason Alabama is getting a shot against LSU next week, nothing else.
"Difficulties strengthen the mind, as labor does the body."
― Seneca
This is first of all a strength of schedule argument, saying essentially that OSU played a harder schedule than Alabama.
No. It’s a conference championship argument. You use the word logic a lot so please explain by any means you choose how Alabama could logically be the National Champion but not the Division or Conference champion.
Maybe you can argue they did, maybe not, but most observers would say Alabama played a harder schedule
Cite one. I’ll wait.
They had the same record as Bama: ok, this is not an argument saying OSU is better, but it is worth noting, I suppose
This is a strength of schedule argument. Oklahoma St managed an identical record playing a more challenging schedule. Therefore, all else being equal, they deserve the nod due to equal achievement in the face of greater degree of difficulty.
If you still don’t see what I mean, think of it like a treadmill at the gym. You run a mile in the same time as your neighbor, but you do it with the incline set to 5 while his is set to 0. Who did more work in the end?
Alabama lost to LSU – as I’ve noted, this is a good argument.
Though sufficient, I would not call this argument the best or even particularly good. Its validity does not impact #1 or #2 at all. Switch the Tide’s loss from LSU to literally any SEC team and if it kept them from winning the CCG then nothing changes.
Of course, the only reason OSU was able to win their conference is because the one team the DID lose to sucked bad enough they were not a threat to win their conference even beating OSU.
I would say that the only reason that OSU was able to win their conference was because they went 8-1 while no other team did better than 7-2. There was not even a tiebreak, as was the case, for instance, with Oregon and Stanford.
But then again I agree with most of the voters in the nation which ranked Bama ahead of OSU due to all the varabiles.
I guess I just agree with most of the computers then, which ranked OSU ahead of Bama while explicitly taking into account most of the variables.
Bama lost to LSU, and Bama didn’t lose to a team that couldn’t win their conference even beating the best team in it, as did OSU. I suppose if the roles were reversed, I might agree with your point.
Since you are focused on losses rather than wins, which is understandable since that’s where Alabama’s case lies, I’ll just point out that OkSt lost in double OT on the road in a short week.
Alabama’s resume is better based on who they played and how well they played them than OSU.
Alabama’s resume might be better based on how well they played their opponents. It is not better based on whom they played.
I take issue that the Big XII (-4) is a stonger conference than the SEC.
Look at their final standings:
1. Oklahoma State
2. Kansas State
3. Baylor
4. Oklahoma
5. Missouri
6. Texas
7. Texas A&M
8. Iowa State
9. Texas Tech
10. Kansas
In contrast to the SEC:
1. LSU
2. Alabama
3. Arkansas
4. Georgia
5. South Carolina
6. Auburn
7. Florida
8. Mississippi State
9. Vanderbilt
10. Kentucky
11. Tennessee
12. Ole Miss
The SEC is considered weak due to the awful Ole Miss team. Tennessee and Florida were decimated by injuries. Kentucky was just not good. Even to even I think that the SEC would beat the Big XII except for Ole Miss (who can’t beat anyone)
"I solemnly swear to tell the truth as I know it, the whole truth as I believe it to be, and nothing but what I think you need to know."
If we played 1 v 1; 2 v 2; 3 v 3 and so on and so on.
Baylor vs Arkansas….. OH THE OFFENSE!! But I think that Bama is hindered by the fact it didn’t play USC or UGA. It probably helped UGA and USC have inflated rankings, too.
- FOW
It's certainly possible you're right
But Jeff Sagarin would disagree with you as would the majority of the other computers.
Let me ask you this. Given that LSU and UGA both had to play 9 conference games because of the SECCG (but only 8 during the regular season). The remaining 10 teams only played 8 conference games however. That accounts for 96 total regular season conference games and 98 total conference games.
The Big XII played 9 conference games apiece, meaning 90 conference games total. How significant do you think those extra 12 regular season games against OOC opponents were in shaping our perception of the SEC as a whole, especially relative to the Big XII which saw half the teams record an additional loss? Pretty significant I’d say.
Not to parse words
But the SEC is not weak and I hope that’s not the sense I left you with.
It is just not as good as the Big XII this year top-to-bottom. Since there was no hiding in the Big XII this year with a full round robin format, OSU’s achievement is all the more impressive. The SEC is much more generous in that regard with LSU and Alabama both missing South Carolina and UGA during the regular season. Of course, it’s fair to point out that that may have had something to do with why those two teams were at the top of the East.
South Carolina and Georgia went a combined 9-1 against the East.
So I think they still would have been at the top of the division. Though they would have had fewer wins and would have been lower in the rankings accordingly. Might I suggest a minor change to your argument? I would state that Oklahoma State played a significantly more difficult schedule that featured more ranked opponents and in doing so won its conference outright. This is rather than saying the tougher conference because while there are Saggarin rankings and computers, it is still somewhat subjective. However, there is no denying that Oklahoma State has more wins against ranked opponents than Alabama.
- FOW
I'd argue Bama only has one argument to make
and it’s the most important. . . they are the better team.
Truly believe that Bama and LSU are a TD better than everyone else in the country. That was pretty much confirmed by the Fiesta Bowl when a Stanford defense, of which not one player would start on either Bama or LSU, held their own against Oklahoma State.
Were they?
How do you know? In 2008 when USC lost to Oregon St on the road and Florida lost to Ole Miss at home there was a very strong argument to be made that USC was the better team. Why didn’t they get to play Oklahoma instead? The answer, in part, was because Florida was the more deserving team having played a more difficult strength of schedule. (This despite USC having played 12 BCS conference opponents to Florida’s 9.)
You’re certainly entitled to your opinion in your second paragraph but your conclusion is simply not right. Stanford’s Defense did not hold their own against OSU. They let up 41 points on 55 plays in only 18 minutes TOP, which is only 7 points below OSU’s season average on 20(!) fewer plays. OSU was “stopped” by Stanford’s Offense, which ran 81 plays and held the ball for nearly 42 minutes.
Year2 (or others)
In comparing the Alabama Defense to the OSU Offense, wouldn’t this be a case where standard deviations above/below the mean result in the most meaningful comparison? It seems like that would account for the ‘zero boundary’ you note by shrinking the size of the standard deviation for the respective categories (e.g. scoring offense/defense).
There is also a limiting factor in discussing the offensive statistics in that most coaches will take their foot off the gas through play call selection and personnel after a certain margin/quarter is achieved. Oklahoma St (or Oregon or Wisconsin or many other teams) could probably have scored 100 points multiple times this season if they’d wanted, but that would be regarded as unsportsmanlike. In contrast, no team, even after pulling the starters, is going to aim to give up more yards per play, which is the approximate defensive equivalent of switching to a run-heavy offensive play call selection. Any changes are more likely to be the result of less skill among the back-ups.
A couple of years ago the guy over at National Championship Issue attempted something similar by comparing average over/under performance in scoring offense/defense against the opponent’s mean to determine each team’s likelihood of over/under performing. In plain speak, was the team’s offense that good or defense that bad and vice versa? Whether or not that’s better than comparing standard deviations I don’t know, but I do think it’s an interesting approach to take.
College football stats are just not advanced enough. Too many variables to account for.
by SenorChuy on Jan 3, 2012 7:09 PM EST via mobile up reply actions
At least for this post we’re only talking about points per game scored or allowed. That seems pretty straightforward to me with relatively few variables to track. Do you say that because of the effect defensive or special teams points scored have on the average figures?
Why not use Standard Deviation on YPP on an offensive and defensive basis then? That won’t be skewed by defensive or special teams points scored/allowed. It might not take into account field position due to the same, but it seems like at least a step in the right direction.
There are too many intangibles. OSU can't run the ball
so no matter how good their passing game is, a one dimensional team would get destroyed by Bama’s D.
by burmbuster on Jan 3, 2012 8:11 PM EST via iPhone app up reply actions
Who do you think would
win an Oklahoma State-Alabama game? Ignore all the stats and don’t answer who you think DESERVES to play LSU. Who do you think would win an OSU-Bama game?
Well
Since we needn’t concern ourselves with who actually deserves to be in the game, or maybe even whether or not they’re eligible, I think Oregon and USC could both beat Alabama and would therefore be a more appropriate choice to play in the National Championship game.
Please clarify your response.
DoubleB asked who you thought would win an OSU-Bama game. And you responded that USC and Oregon could both beat Bama. Are we to infer that OSU is better than Oregon and USC; and, that OSU could thusly also beat Bama? And obviously any team can beat anyone, Iowa State can beat OSU, Oregon State can beat Southern Cal in 2008, UK beat LSU in 2007, Ole Miss beat Florida.
Personally, I am in the camp that would like to have seen the Okie State vs LSU matchup. I understand the flaws on both scenarios and the Okie State vs LSU matchup is what I prefer. This is despite the fact that I personally believe Alabama is a better team than Oklahoma State. As terrible as the SEC offenses rank, we have seen high powered offenses whither and die against SEC defenses in several high profile games recently. And some may not have died but have certainly not enjoyed the success that they enjoyed against other opponents.
This is leading into a statistics debate, in which I found myself not adequately armed, so I’ll simply restate my position. I personally believe Okie State would lose to Alabama but would like to see them play LSU because it pits a conference champion against another.
- FOW
by skandrewj62j on Jan 6, 2012 11:07 AM EST up reply actions
True
I did avoid the question somewhat, but only to strike at the assumption underneath. He wants me to admit that Bama is better than OSU which would strengthen his claim that Bama is the #2 team and the consideration that as the second best team they are the second most deserving. I reject his premise and believe that we can do nothing but offer conjecture about the results of an OSU/Bama match-up
To the question then: Oklahoma State’s offense is for real and would do significantly better than did Arkansas’. They also have a considerable gift for generating turnovers, as does LSU, which could easily leave Bama unable to catch-up. Still, I think Bama would be favored, though by anything more than one possession would be too much.
After my question about using Standard Deviation to measure scoring offense/defense, I went ahead and calculated it myself based on CFBstats.com. OkSt’s was based on 13 games, Bama’s was based on 12. Even with their underperformance against Stanford, OSU’s scoring offense was a greater absolute distance from the mean than was Bama’s scoring defense. However, the combined Std Dev of scoring offense + scoring defense favored Bama.
To your other point, I don’t think there’s any doubt here that I would prefer to see OSU vs LSU. This is true for tangible as well as philosophical reasons. Tangibly, OSU has accomplished significantly more and Alabama has already swung at LSU and missed. Philosophically, letting this game occur only increases the mythology that ESPN, et al., have developed around the SEC that allows them to get away with playing weak OOC teams and paper over occasions when the conference is very top-heavy.
That is completely irrelevant
Alabama may be the second best team in the country, but they have already been eliminated by the best team in the country. Sometimes the two best teams aren’t the two last teams standing. Sometimes the two best teams play in the conference championship, or the final four. The only way Alabama (or Oregon had they made it with one loss) should get another shot is if ever single other one loss team had lost again.
by mypisceannature on Jan 6, 2012 9:27 AM EST up reply actions
But the BCS is designed to put the two HIGHEST RANKED teams in the MNCG, not the two MOST DESERVING.
And since cfb rankings are generally understood to be an ordering of teams from Best to Worst we can say that the teams ranked #1 and #2 are the two BEST teams.
Make no mistake about my words, I hate the BCS system, but in this instance the BCS has done what is was designed to do.
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
