Plays/minute vs Time of Possession...

Easton Pa ND Fan · Sep 23, 2010

You are twice as likely to score points based on plays/minute
compared to time of possession (ToP). ToP is a better indi-
cator of margins and win percentages.

A 2008 database (120 teams/1521 games) provided the correla-
tions. Thanks to "Gen" on UHND for the idea. Other relation-
ships to time statistics are also listed...Tom

<pre>ToP Statistics OffPts/g M o V /g DefPts/g %gameWon Off plays/ min 0.365327 0.072558 0.276451 0.034805 Time of P/game 0.175417 0.272929 -0.25609 0.294807 </pre>

</PRE>

IrishinSyria · Sep 23, 2010

Ok, this has the makings of a really interesting discussion, but the data is a little raw as presented. Do you have a link or some more in-depth analysis?

Old Man Mike · Sep 23, 2010

hmmm...the data is still a bit confusing to me. But, for talking purposes, let's say that the numbers show that if you're a hurry-up team, then you score a lot, but you don't possess the ball much, and the other team scores a lot too. Is that what we're indicating here? Then, because the other team scores a lot, the numbers of wins to losses approaches more closely to 50% than you'd like it to. Is this still what we're saying here? -----------------------------------------------------------------------------------------------------------------------------------------------------On the other hand, if you're a competent slower paced team, you maintain possession of the ball alot, but you score less than the fast-paced team. Still making sense? However, your defense doesn't stay on the field that long, and consequently gives up fewer points. What you've lost by not scoring as much is more than made up for by the defense's ability to limit the opponent [apparently because the opponent has less opportunities against a fresher defense] and you win more games percentage-wise. Is this the analytical hypothesis? --------------------------------------------------------------------------------------------------------------------------------------------------------If the lay-out of the idea above is correct, and the statistics actually bear this out, then it would seem that coaches employing hurry-up offenses have some explaining to do---a VERY interesting concept. I would like to toss out two points to add to the stew. A). do coaches who like high-octane offenses prejudice the player distribution in favor of the offense, and thereby create this leaky defense statistic as a internally determined result? [i.e. if the quality of personnel were equally shared to the defense, would the defense not create a significant differential of points even given the Time-Of-Possession "problem"? B). Kelly's system seems to have worked terrifically despite the general statistics. Was that a several year fluke, or does he do something different total-system-wise? Texas Tech seemed to "work" as well. So did WVA in Pat White's tenure. ...confusing, but interesting.

phork · Sep 23, 2010

What it shows me is that while your points double, you are simply trying to outscore your opponent. Defense be damned. While Kelly insists that TOP is not a big deal, I still think running the ball and defense are the keys to winning any game.

No sense scoring 50 if they score 51.

Easton Pa ND Fan · Sep 23, 2010

The data links are at...

The data links are at...

IrishinSyria said:
Ok, this has the makings of a really interesting discussion, but the data is a little raw as presented. Do you have a link or some more in-depth analysis?

The link is below. The data was copied and pasted from
this site to an Excel spreadsheet. After getting garbage
results, we realized that all columns should have been
in alphabetical order.:embar:

Statistics, calculated from this data, are listed.

cfbstats.com - 2009 National

<pre>Teams Play/m Min/g Opt/g Dpt/g Win %. MoV/g A F . 2.2572 31.18 26.77 22.23 0.6154 4.54 Akron 2.3268 30.05 30.00 31.25 0.4167 -1.25 Al . 2.0103 32.05 30.14 14.29 0.8571 15.85 Ariz. 2.1896 32.18 36.62 21.31 0.6154 15.31 ArizS 2.1963 28.99 22.83 22.67 0.4167 0.16 Ark . 2.2484 30.43 21.92 31.17 0.4167 -9.25 Ark S 2.0198 32.84 27.00 23.75 0.5000 3.25 Army. 2.0025 31.75 14.75 23.67 0.2500 -8.92 Aubrn 2.2351 30.20 17.33 18.00 0.4167 -0.67 BallS 2.1698 30.45 34.93 20.50 0.8571 14.43 Baylr 2.3012 27.92 28.00 29.33 0.3333 -1.33 Boise 2.3663 28.80 37.62 12.62 0.9231 25.00 BostC 2.2271 30.95 24.71 18.36 0.6429 6.35 BowlG 2.3030 29.60 27.67 23.17 0.5000 4.50 Buflo 2.2003 31.10 30.29 28.21 0.5714 2.08 BYU . 2.3273 30.31 34.23 21.92 0.7692 12.31 Cal . 2.1986 29.11 32.46 19.76 0.6923 12.70 CenMi 2.4866 29.82 29.54 30.23 0.6154 -0.69 Cinci 2.2129 29.92 25.86 20.14 0.7857 5.72 Clmsn 2.1801 29.32 25.15 17.31 0.5385 7.84 Col . 2.4266 28.95 20.17 29.25 0.4167 -9.08 Col S 2.1278 31.92 25.15 30.23 0.5385 -5.08 Conn. 2.1470 31.71 24.92 19.77 0.6154 5.15 DukeU 2.2583 30.59 20.08 23.42 0.3333 -3.34 E Car 2.1851 30.63 23.43 21.07 0.6429 2.36 E Mic 2.3299 32.80 25.75 35.58 0.2500 -9.83 FL . 2.0734 30.11 43.64 12.93 0.9286 30.71 FL A 2.4277 27.66 25.08 28.69 0.5385 -3.61 FL I 2.3313 27.56 24.67 28.33 0.4167 -3.66 FL S 2.1209 31.59 33.38 20.15 0.6923 13.23 Frsno 2.1900 29.89 29.62 31.08 0.5385 -1.46 Ga . 2.0530 30.95 31.46 24.54 0.7692 6.92 Ga T 2.0272 30.51 24.38 20.31 0.6923 4.07 Hi . 2.1076 30.57 24.64 28.86 0.5000 -4.22 Houst 2.8622 27.28 40.62 30.92 0.6154 9.70 Id . 2.2056 30.30 19.58 42.75 0.1667 -23.17 IL . 2.3828 29.83 28.67 26.58 0.4167 2.09 In . 2.4942 26.83 20.50 35.25 0.2500 -14.75 Iowa. 2.0384 31.51 30.31 13.00 0.6923 17.31 IowaS 2.2775 30.77 25.33 35.83 0.1667 -10.50 Kan . 2.3895 30.42 33.38 28.85 0.6154 4.53 Kan S 2.5098 27.66 34.92 35.83 0.4167 -0.91 KentS 2.3659 29.13 25.67 31.67 0.3333 -6.00 Ky . 2.2430 29.46 22.62 21.46 0.5385 1.16 La T 2.3164 28.76 24.62 23.69 0.6154 0.93 LaLaf 2.1943 31.14 33.08 33.67 0.5000 -0.59 LaMon 2.0573 31.23 23.67 31.17 0.3333 -7.50 Louvl 2.2292 31.59 24.67 29.83 0.4167 -5.16 LSU . 2.2559 30.21 30.92 24.15 0.6154 6.77 Marsh 2.3168 27.84 20.50 27.67 0.3333 -7.17 Md . 2.3084 27.79 21.77 22.46 0.6154 -0.69 Memph 2.2905 32.81 27.15 27.15 0.4615 0.00 MiaFL 2.1846 29.47 27.08 24.15 0.5385 2.93 MiaOh 2.2742 30.67 18.42 32.67 0.1667 -14.25 Mich. 2.3841 27.65 20.25 28.92 0.2500 -8.67 MichS 2.2566 31.02 25.08 22.08 0.6923 3.00 MidTn 2.3743 28.99 22.83 24.92 0.4167 -2.09 Minn. 2.1813 30.01 23.23 24.77 0.5385 -1.54 Mis . 2.1909 30.23 32.08 19.00 0.6923 13.08 Mis S 2.1544 30.25 15.25 24.67 0.3333 -9.42 Mo . 2.7165 25.82 42.21 27.21 0.7143 15.00 Navy. 1.9787 31.41 27.15 22.00 0.6154 5.15 Neb . 2.0785 34.01 35.38 28.54 0.6923 6.84 Nev . 2.3778 32.74 37.62 32.31 0.5385 5.31 New M 2.2712 30.38 21.08 22.75 0.3333 -1.67 NewMS 2.3116 29.20 22.17 34.08 0.2500 -11.91 No C. 2.1853 27.14 27.69 21.15 0.6154 6.54 No CS 2.1687 29.58 23.46 26.31 0.4615 -2.85 No Tx 2.4012 30.68 20.00 47.58 0.0833 -27.58 NorIL 1.9562 31.97 24.15 18.00 0.4615 6.15 Nor'W 2.4480 29.82 24.38 20.15 0.6923 4.23 Notre 2.1714 31.28 24.69 22.15 0.5385 2.54 Oh . 2.1704 29.87 24.08 27.25 0.3333 -3.17 Oh S 1.9590 31.69 27.62 13.92 0.7692 13.70 OK . 2.6843 29.43 51.14 24.50 0.8571 26.64 Ok S 2.2424 31.15 40.77 28.08 0.6923 12.69 Or . 2.9297 25.18 41.92 28.23 0.7692 13.69 Or S 2.2574 31.86 30.54 23.08 0.6923 7.46 Pa S 2.1673 31.73 38.92 14.38 0.8462 24.54 PittS 2.1413 31.36 27.08 21.46 0.6923 5.62 Purde 2.4858 29.97 24.67 25.08 0.3333 -0.41 RiceU 2.4062 30.85 41.31 33.31 0.7692 8.00 Rutgr 2.1169 30.45 29.00 18.85 0.6154 10.15 S D S 2.4578 26.21 19.25 37.17 0.1667 -17.92 S J S 2.1833 29.35 18.67 21.58 0.5000 -2.91 SMU . 2.1834 26.45 21.33 38.17 0.0833 -16.84 So C 2.2508 29.70 20.77 21.08 0.5385 -0.31 So FL 2.3176 30.70 27.62 20.00 0.6154 7.62 So Ms 2.4772 30.68 30.62 24.08 0.5385 6.54 Stnfd 2.1476 30.15 26.25 27.42 0.4167 -1.17 Syrse 2.1130 27.25 18.08 32.67 0.2500 -14.59 TCU . 2.2307 34.93 33.62 11.31 0.8462 22.31 Templ 2.2856 27.42 23.42 23.08 0.4167 0.34 Tenn. 2.0956 28.55 17.33 16.75 0.4167 0.58 Tex . 2.2603 32.50 42.38 18.77 0.9231 23.61 TexAM 2.3175 27.94 25.00 37.42 0.3333 -12.42 TexT. 2.4871 30.28 43.77 27.85 0.8462 15.92 Toldo 2.2333 28.21 22.42 31.42 0.2500 -9.00 TroyS 2.5783 29.00 32.77 21.31 0.6154 11.46 Tulne 2.1525 31.28 16.67 34.50 0.1667 -17.83 Tulsa 2.6572 29.49 47.21 27.93 0.7857 19.28 UAB . 2.0994 31.28 22.75 31.25 0.3333 -8.50 UCF . 2.2821 27.79 16.58 24.08 0.3333 -7.50 UCLA. 2.2594 29.95 17.67 29.00 0.3333 -11.33 UNLV. 2.1417 30.70 25.58 32.58 0.4167 -7.00 USC . 2.1940 31.24 37.54 09.00 0.9231 28.54 Ut . 2.1663 32.35 36.92 17.23 1.0000 19.69 Ut S 2.2644 28.74 24.00 34.67 0.2500 -10.67 UTEP. 2.1091 32.36 32.92 37.00 0.4167 -4.08 Vanbl 2.0365 29.84 19.15 19.62 0.5385 -0.47 Vir . 2.3287 27.84 16.08 21.67 0.4167 -5.59 Vir T 2.0261 33.00 22.07 16.71 0.7143 5.36 WakeF 2.2454 30.32 21.00 18.31 0.6154 2.69 Wash. 2.2427 29.17 13.25 38.58 0.0000 -25.33 WashS 2.1290 29.30 12.69 43.85 0.1538 -31.16 WestK 2.1670 28.69 17.50 27.17 0.1667 -9.67 WestM 2.4341 28.22 28.62 24.92 0.6923 3.70 WestV 2.1781 29.03 24.54 17.00 0.6923 7.54 Wisc. 2.2595 30.64 27.46 26.54 0.5385 0.92 Wyo . 2.2383 29.04 12.67 27.75 0.3333 -15.08 </pre>

</PRE>

Easton Pa ND Fan · Sep 23, 2010

Mike, are you a retired statistician?

Mike, are you a retired statistician?

Old Man Mike said:
hmmm...the data is still a bit confusing to me. But, for talking purposes, let's say that the numbers show that if you're a hurry-up team, then you score a lot, but you don't possess the ball much, and the other team scores a lot too. Is that what we're indicating here? Then, because the other team scores a lot, the numbers of wins to losses approaches more closely to 50% than you'd like it to. Is this still what we're saying here? -----------------------------------------------------------------------------------------------------------------------------------------------------On the other hand, if you're a competent slower paced team, you maintain possession of the ball alot, but you score less than the fast-paced team. Still making sense? However, your defense doesn't stay on the field that long, and consequently gives up fewer points. What you've lost by not scoring as much is more than made up for by the defense's ability to limit the opponent [apparently because the opponent has less opportunities against a fresher defense] and you win more games percentage-wise. Is this the analytical hypothesis? .

Mike,

You must be a retired statistician to have interpreted
those correlations. I only pointed out the individual
"vertical" contrast differences in my post.

I didn't quite have a handle on how to interpret all
the numbers, but you sure make a lot of sense out of
it. I looked at the "up-down" and you looked at the
horizontal numbers. Reps to you...Tom :king:

Old Man Mike · Sep 23, 2010

Actually I have no idea if i know what i'm talking about. All I've done is to lay out the logic of the hypothesis. When you look closely at your list, there are many questions. The main thing to me is that it seems that you can be a success either way, even though the statistics seem to indicate that coaching staffs which coach "slow to moderate" tend to have a little more success. Example: most of the big powers are "slow" but so was Tennessee, Vanderbilt, Army, and Syracuse. On the other hand Oregon was the fastest team in the NCAA. Oklahoma and Houston, hardly slouches, were also fast. One might argue that the reason that Tulsa competed last year was that they were fast. Kelly's Cincinnati oddly was almost "normal" on this plays/minute thing. --------------------------------------------------------------------------------------------------------------------------------------------------------------As i look through the whole list, it seems to me that you can win either way just fine---but you have to have players---and you have to have coaching. I will bet that the reason for the statistical edge that has been noticed here has to do with the power programs being mainly "slow" "control" teams, and the reason that so many of them are that way is that this is the style that the coaches who build long-term reputations so as to get hired at such institutions have always used and are comfortable with. The fact that they get elite talent allows almost any well coached scheme they throw out there to win. What Kelly has probably done is to come up with a creative way to compete with the big boys without having big boy talent. Now that he is getting it, he should be able to employ his "schematic advantage" to the full. That's my hope and my faith. -----and thanks very much, Tom, for another very stimulating post.

phgreek · Sep 24, 2010

Easton Pa ND Fan said:
The link is below. The data was copied and pasted from
this site to an Excel spreadsheet. After getting garbage
results, we realized that all columns should have been
in alphabetical order.:embar:

Statistics, calculated from this data, are listed.

from what I can see...few have kelly's W%...
on the filed too much? not on the field enough? there will always be concern.

Also saw some pretty cool analysis on who has done the most with their players based on their average original star rating as recruits, and their W/L %...can't remember where, but it was an ND related fan site...anyway

Guess who was among the top in that list too?

He seems to have a grasp on the elusive W most of the time...he appears to develop the hell out of young men...I think the numbers help me see the light at the end of the tunnel...and make believe its not a train...

Love the data and analysis...cool stuff

IrishinSyria · Sep 24, 2010

Old Man Mike said:
hmmm...the data is still a bit confusing to me. But, for talking purposes, let's say that the numbers show that if you're a hurry-up team, then you score a lot, but you don't possess the ball much, and the other team scores a lot too. Is that what we're indicating here? Then, because the other team scores a lot, the numbers of wins to losses approaches more closely to 50% than you'd like it to. Is this still what we're saying here? -----------------------------------------------------------------------------------------------------------------------------------------------------On the other hand, if you're a competent slower paced team, you maintain possession of the ball alot, but you score less than the fast-paced team. Still making sense? However, your defense doesn't stay on the field that long, and consequently gives up fewer points. What you've lost by not scoring as much is more than made up for by the defense's ability to limit the opponent [apparently because the opponent has less opportunities against a fresher defense] and you win more games percentage-wise. Is this the analytical hypothesis? --------------------------------------------------------------------------------------------------------------------------------------------------------If the lay-out of the idea above is correct, and the statistics actually bear this out, then it would seem that coaches employing hurry-up offenses have some explaining to do---a VERY interesting concept. I would like to toss out two points to add to the stew. A). do coaches who like high-octane offenses prejudice the player distribution in favor of the offense, and thereby create this leaky defense statistic as a internally determined result? [i.e. if the quality of personnel were equally shared to the defense, would the defense not create a significant differential of points even given the Time-Of-Possession "problem"? B). Kelly's system seems to have worked terrifically despite the general statistics. Was that a several year fluke, or does he do something different total-system-wise? Texas Tech seemed to "work" as well. So did WVA in Pat White's tenure. ...confusing, but interesting.

Very good start... in addition to the framework you give here, I'd add another wrinkle: do teams change their pace of play at the half or in the fourth quarter based on score. If they do, you'd expect teams that tend to be losing to run a faster paced offense in the later parts of the game and teams that tend to be winning to run a slower paced offense in the later parts. This could distort the data, essentially speeding up the pace of play of losers and slowing it down for winners.

edit: Ok, so I thought and thought about this, and I finally remembered basic stats. The first thing is that it would be wrong to view them as opposing categories. Thus, you're not twice as likely to score if you run more plays, it's just that the number of plays you run per minute is a better indicator of scoring than time of possession is. Maybe everyone else is reading this and saying "obviously" but...I had to think about it.

Having re-examined everything, OMM's observstions are on the money. A few more things that jumped out to me:

1) Both off plays/min and time of P/game have a possitive correlation to Margin of victory/ game. Granted, it's much higher with P/game but still...more plays and more ball time both seem to be at least somewhat tied to victory.
2) Time of possession is HUGE for helping out your own D. If we were to graph out off plays/min and P/game for def points/game it would look like an X...

To sum it up, this data is interesting but not very surprising. The more offensive plays you run per minute, the more likely you are to score. The longer you hold the ball, the less likely your D is to be scored on. As Mike pointed out, the real eye opener is the much stronger correlation to Margin of victory and winning percentage that time of possession has versus O plays per minute. We have to remember that correlation is not causation, and several factors could influence this outcome (for example, teams running the play clock down while holding a lead). Nonetheless, it does appear that Coach Kelly's success over the past few years is a bit of an outlier. I guess only time will tell if this is due to a run of strange luck, distorted results, or Kelly being a truly mold-breaking coach.

edit #2: Thank you for posting this Tom, makes me wish I'd paid more attention in Stats.

Plays/minute vs Time of Possession...

Easton Pa ND Fan

New member

IrishinSyria

In truth lies victory

Old Man Mike

Fast as Lightning!

phork

Raining On Your Parade

Easton Pa ND Fan

New member

Easton Pa ND Fan

New member

Old Man Mike

Fast as Lightning!

phgreek

New member

IrishinSyria

In truth lies victory