Wednesdays are a bit of a reset day on the NFL beat, so let me show you something we did over at the old blog and update those numbers.
There are some standards that NFL broadcasters are always saying that I always like to challenge just to see if they are telling the truth. One in particular started back in the 1990s when the talking heads used to tell us that if Emmitt Smith had 20 carries then the Cowboys always won. I could never appreciate that stat (even if it was true) because I thought there was no correlation between that and winning. If you want to say when Emmitt gets 100 yards, I would believe it. Getting 100 yards has all sorts of winning attributes, but merely calling running plays would not lead to winning unless they were effective. The number of carries is obviously not as important as what he does with them. Otherwise, couldn’t they hand him the ball on the first 20 plays, he takes a knee, loses 40 yards in the process, and the Cowboys win? It just didn’t compute to me.
Am I crazy?
Let’s See:
Emmitt Smith Career
| Year | Rec W/ 20+ Carries | Rec W/ 100 yards+ |
| 1990 | 5-0 | 3-0 |
| 1991 | 9-0 | 7-1 |
| 1992 | 9-1 | 6-0 |
| 1993 | 9-0 | 8-0 |
| 1994 | 9-3 | 4-2 |
| 1995 | 11-3 | 9-2 |
| 1996 | 9-2 | 2-2 |
| 1997 | 4-1 | 0-2 |
| 1998 | 7-2 | 7-0 |
| 1999 | 6-5 | 4-5 |
| 2000 | 4-4 | 3-3 |
| 2001 | 5-1 | 3-1 |
| 2002 | 2-2 | 1-1 |
| 2003 | 1-0 | 0-0 |
| 2004 | 3-2 | 2-0 |
| Totals | 93-26, .781 | 59-19, .756 |
Wow. The talking heads were right. According to my math, the Cowboys actually had a better record with Emmitt getting 20+ carries than when Emmitt ran for 100+ yards. I guess that means controlling the clock is more important than rolling up yards? By the way, Emmitt had 78 (Edit: Oops) 100 yard games according to my research – so that should blow your mind. This is why we need to put theories to the test.
So, that caused me to find stats that have meaning. 100 yards for a Running Back is actually a meaningful number across the NFL. In 2008, 131 teams had games in which they had a runner go for 100 yards. The record was 102-29 (.779 win percentage). We will track these numbers all season long right here:
100+ Yard Rushers in Week 1:
| Name | Team | Opp | Yards | W/L |
| A Peterson | Minn | Clev | 180 | W |
| Mike Bell | NO | Det | 143 | W |
| J Jones | Sea | STL | 117 | W |
| Ray Rice | Balt | KC | 108 | W |
| Th Jones | NYJ | Hou | 107 | W |
| Totals | 5-0 |
300 yards QBs? Based on the idea that teams that are playing from behind generally throw more, you would expect that this number would be much lower. And it is. 74 games had 300 yard passers in 2008, and QBs who threw for 300 yards brought home just 46-27-1 record (.630 win percentage).
300 Yard Passers in Week 1:
| Name | Team | Opponent | Yards | W/L |
| Tom Brady | NE | Buf | 378 | W |
| Roethlisberger | Pit | Ten | 363 | W |
| Drew Brees | NO | Det | 358 | W |
| Tony Romo | Dal | TB | 353 | W |
| Joe Flacco | Balt | KC | 307 | W |
| P Manning | IND | Jac | 301 | W |
| Totals | 6-0 |
I have been obsessed with numbers for years. And I have been looking for that magical statistic that always correlates to winning. There is one in the NFL. And I don’t think that it has a rival. Turnovers. Turnovers are seemingly the single most telling statistic in the NFL to find a winner. Point Spread favorites don’t matter the same. Home Stadium doesn’t matter the same. Yardage advantage doesn’t matter as much. Turnovers matter most.
I have been tracking turnovers for every game in the NFL for 4 years. 2008 was the least convincing year in the turnover ratio. Usually, the turnover winners in each game has won before 82-85% of the time. In 2008, it dropped to just 157-41 for 79.2%.
Here is the sliding scale of winning for turnover winners from 2008:
| Total | Record | Win % |
| +6 | 1-0 | 100% |
| +5 | 6-0 | 100% |
| +4 | 13-0 | 100% |
| +3 | 24-6 | 80% |
| +2 | 44-11 | 80% |
| +1 | 69-24 | 74% |
| Totals | 157-41 | 79.2% |
And then, Week 1 – 2009 results:
| Winner | +/- | Loser |
| Phil | +5 | Car |
| GB | +4 | Chi |
| Atl | +4 | Mia |
| Den | +2 | Cin |
| Min | +2 | Cle |
| NYJ | +1 | Hou |
| SF | +1 | Arz |
| Pit | +1 | Tenn |
| SD | +1 | Oak |
| NO | E | Det |
| Dal | E | TB |
| NYG | E | Wash |
| NE | E | Buf |
| Balt | -1 | KC |
| Ind | -2 | Jac |
| Sea | -2 | StL |
| Totals | 9-3 |
Turnover Winners through week 1 are just 75%. That will improve, but already, 2 teams won at -2. Not easy to do.
These numbers in all categories will be far more interesting as the season goes on. After week 1, too early to read any meaning into it. But, since I keep these numbers (With intern TC’s help) I want to share them with you as we go…
RANDOM FOOTBALL EMAIL OF THE WEEK:
Hi Bob (been a few years since we spoke last),
You are right about Football being the ONE true judge of a college to a degree, in Texas anyway. So, “Here Come the Bears!” Great clip below on his interview after Wake Forest win and human highlight film, fastest QB in the nation and sophomore player that wants to graduate like Mike S. did….being the largest Baptist school in the world gets us one All Pro every two decades looks likeJ
Enjoy, com’on give Griffin some air time (or maybe you did and I missed it)……..
Chris
I’ll be honest, I have not given Griffin enough time or attention yet. A quick look at the two videos below demonstrate that Baylor appears to have a rare talent for sure. The dude won the 400m hurdles? As a freshman? And he plays QB? Geez.
A quick scan of the Baylor Schedule suggests that UConn this weekend should be good, but I don’t see it on TV. Therefore, my first good look at the young man this season will be his trip to Norman on October 12. I can’t wait.
Bob,
To really prove the correlation, wouldn’t it be interesting to know the opposite. eg. What was the Cowboys record in games where Emmitt played and did not run for 100 yds or 20 carries. Maybe the Cowboys just won a lot. Maybe just having Emmitt on the field cause defenses to focus on him opening up the offense. Just think it would make your argument better.
Hey Bob!
I think your initial thoughts were on the right track despite the numbers. Number of carries to me is an indication of the degree to which a team is controlling the game. In other words, the Cowboys’ wins in the 90′s didn’t correlate with ES22′s carries because his 20+ carries caused the win. It was that during the possessions in which he had his first 12 or so carries, the team got a 10+ point lead, which then led to his getting those last 8-12 carries. Kind of like how really bad teams that get blown out often are near the top in fewest pass yards allowed — when a team “only” throws for 175 yards when running up a 24-3 lead midway in the 3rd Q, they can just run the rest of the way. We wouldn’t say that a team getting blown out in this fashion has a stellar pass defense.
I’m interested in the stats too, but (as you know), correlation ≠ causation.
One note on Griffin…he graduated early from high school, so he essentially won the Big 12 400m hurdles as a senior in high school, barely 18 years of age. He’s a pretty special talent.
Griffin looks great for Baylor, exciting things going on in Waco this year.
Bob,
I always assumed the 20 carries stat had to do with sticking with your game plan, which usually means you control the game. The Cowboys ALWAYS wanted to run Emmitt a lot, so when they didn’t, it meant the other team was knocking them off their game. The fact that the stat held true even when the Cowboys were sliding strengthens it.
I wonder if the Giants have a similar stat for rushing attempts per game? And what would the Brady Patriots or Manning Colts winning stats be? I wonder if there is a completion percentage number for those guys that always spells victory? Or yards per catch? Percentage of completions to wide receivers? I bet there is some type of stat that demonstrates that those teams had success within their game plan and almost always won.
Sounds like a good project for an intern!!!
I never listen…
Baylor/UConn is not televeised. There will be a tape-delayed airing of the game later on Saturday night at BaylorBears.com.
@redstripe – I agree with the causation/correlation idea on rushing, but the turnovers are very much related to making the outcome – not because of the outcome. Agree?
Bob
What about missed/blocked field goals? Personally, I think they count as half of a turnover, since you’re giving the ball back to the other team, there’s a field position swing involved, momentum shifts, and you’re leaving points off the board. Do you know if anyone has ever done the math on missed FGs affecting the outcomes of games from a statistical perspective?
Bob,
Run a regression and determine the P-value!
You get a lot of carries when you have a lead.
So even if he didn’t gain a lot of yards, he’s getting the ball because the are winnning. Hence OBVIOUSLY the better winning %.
How is 59-19 = 119 games?
Emmitt had 78 100 yard games, 1 more than Payton’s 77 for the NFL record. He did however, have 119 games with 20+ carries.
Bob,
Did you notice that in 1999-2000 when the cowboys were waning that carries and yards were much weaker indicators? Turnover ratios seem to hold up much better for bad teams.
@BarrySandersWasBetter nice observation, hadn’t noticed that. Maybe attribute it to lack of ability to run the ball consistently. 100 yard games probably imply 1 or 2 freak plays that are big gainers, but do little to affect the outcome of the game. Also, numbers could have been accumulated in garbage time as well. I think an even better indicator of the overall offensive skill of a team would be avg. ypc for an offensive line and…similarly ypc allowed by a defense. Any team averaging 4+ ypc and allowing less than 4 ypc…guarantee they have winning records.
Interesting stuff, Bob.
I’m not surprised that turnovers appear to contribute so much to winning.
I came to the same conclusion looking at college football stats and posted it a couple of weeks ago on my blog.
I found that adding scoring defense to analysis produced even stronger correlatiion in the college game, I wonder if the same would hold up in the pros.
Defense and turnovers, wins games. No doubt.