Stiff Arm Trophy 2009
WE DID IT! It was the closest Trophy race in history, but we correctly projected the winner, Mark Ingram of Alabama. That gives us a record of correct projections eight out of eight years. Dig into the details here.
Our Projection: FINAL
We've now completed our final projection for 2009. Read the details here.
The tally is on the left is a vote count from actual voters. The projection takes the regional breakdown and extrapolates a final total. NOTE: Here's an explanation of why the numbers total more than 100%.
Last updated: 12/12/2009, 1:51 a.m. Pacific. 313 ballots, 869 votes.
Stiff Arm Trophy on Twitter
With over 160 official voters on Twitter, and thousands of fans, we're using Twitter to track the votes and take note of funny, insightful, and provocative commentary. Follow Stiff Arm Trophy on Twitter or track the chatter here.
Over at the Gainesville (Florida) Sun, Heisman voter Pat Dooley argues that it's about time for some transparency in the Heisman balloting.
You hear those of us in the media screaming from the top of the building that the coaches must have transparency when it comes to their poll. We won that battle finally, but shouldn’t the Heisman votes also be made available to anyone who wants to see them? I want to know the 43 people who voted Tim Tebow No. 1 on their ballots. I want to know which 538 people left Ndamukong Suh off their ballots and which two voted Golden Tate first.
I couldn't agree more. At a minimum, the Heisman Trust should reveal who the list of voters are.
Kari Chisholm | December 19, 2009 | Comment on This Post (6 so far)
According to the Associated Press, the closest Heisman Trophy vote of all time led to the highest ratings for the Heisman award show ever.
The closest vote in Heisman Trophy history drew a record television audience for the announcement ceremony.
ESPN's telecast Saturday averaged a 4.1 rating, up 32 percent from the previous high of 3.1 set the last two years and in 2002.
Alabama tailback Mark Ingram finished just 28 points ahead of Stanford's Toby Gerhart.
A ratings point represents 923,000 households.
Ponder that for a second.
How could the closest vote ever lead to high ratings? The vote's secret, after all.
Unless, that is, people knew it was going to be a close vote. Unless, somewhere out there, there was a website that was able to tell people that there was some suspense in the outcome.
Oh yeah. There is. Right here. 3.78 million households tuned into the Heisman ceremony. On the 12 days prior to the ceremony, Stiff Arm Trophy recorded 1.35 million pageviews from 1.267 million daily unique visitors. On top of that, I did nearly two dozen radio interviews - and our vote counts were featured in daily ESPN broadcasts and many local and national publications.
So, my dear friends at the Heisman Trophy Trust, you're welcome. No need to send a commission check - or even a fruit basket. We're all good.
I am, however, going to change one thing. We're going to start using the word "Heisman" around here again. There's no legal reason to avoid it, it was just my little way of tweaking the Trust's silly tantrum about our old name (HeismanProjection.com). We won't use the Heisman's trademarks in our branding, but we're not going to be afraid of using the word itself.
It's good to see some other folks take note of the problem we've been talking about for the last six years - that the distribution of H------ voters across the country is terribly skewed. Here's StatePaper.com's Samuel McKewon on the topic. (Ironically, writing from Nebraska - which is way overrepresented, with 16 voters, instead of the six that a population-apportioned method would offer them).
You know what else is a sham? Votes equally apportioned by region. Not only does that ignore the population quotient – the West Region has far more people in it than most of the geographical subsections – it hands 145 votes to the Northeast Region, which is, essentially, New England.
Want to guess how many Division I programs are in the entire region? Five. Army, Syracuse, Buffalo, Connecticut and Boston College. Even if one assumes that a decent percentage of the votes go to ESPN (located in Bristol, Conn.) you're still awarding votes to a region where baseball and college basketball reign supreme. How often – if ever again – will that region have a player in the Heisman running?
One-sixth of the Heisman race should not be decided by a region with no dog in the fight.
Before the announcement, we took a look at the various projections, polls, and predictions out there. So, how'd everybody do?
|Stiff Arm Trophy||Ingram||Gerhart||Suh||McCoy||Tebow|
|Scripps-Howard||Ingram||Suh||Gerhart||McCoy||Tebow & Spiller|
|Heisman Pundit Poll||Ingram||Gerhart||McCoy & Suh||Moore|
|ESPN Heisman Watch||Gerhart & Ingram||Suh||McCoy||Tebow & Spiller|
|CBS Heisman Watch||Gerhart||Ingram||Moore||McCoy||Suh|
|NBC Experts Poll||Ingram||Gerhart||Suh||McCoy||Spiller|
|USA Today||Ingram||Gerhart||McCoy & Suh||Moore|
|Heisman Pundit, himself||McCoy||Ingram||Gerhart||Tebow||Spiller|
Joining Stiff Arm Trophy in picking Mark Ingram as the winner were the Scripps-Howard poll, the Heisman Pundit poll, the NBC experts poll, USA Today's voters. We'll also give the ESPN Heisman Watch poll credit, as they called for a tie between Ingram and Toby Gerhart - and it was the closest Trophy outcome ever.
Predicting Gerhart as the runner-up were all of the above - except Scripps-Howard, who had Ndamukong Suh in the #2 spot.
Who got the top two wrong? Sports Illustrated's voters had Suh and Ingram in the top two. The stats-based ESPN Heisman Predictor had Colt McCoy and Ingram. The CBS Heisman Watch experts poll had Gerhart defeating Ingram. Chris Huston, the Heisman Pundit himself, personally predicted that it'd be McCoy and Ingram as well.
How about the surprise #3 finish for McCoy - well ahead of Suh? No one got that right, including us. And probably for the same reason.
Stiff Arm Trophy was the only projection to correctly peg Tim Tebow as the fifth-place finisher (though ESPN's poll and the SHNS poll both had a Tebow/Spiller tie for fifth.)
Kari Chisholm | December 13, 2009 | Comment on This Post (5 so far)
For the eighth year in a row, we've correctly projected the winner of the H----- Trophy. Congratulations to Mark Ingram and the Alabama Crimson Tide!
We projected that Ingram would get 1349 points, 48.6% of the maximum points possible. He earned 1304 points, 46.9% of max - just 1.2% off. We had also correctly pegged Toby Gerhart as the second-place finisher, with a projected 1225 points (44.1%) and an actual 1276 points (45.9%, also just 1.2% off.)
Ingram beat Gerhart by the closest margin ever, which validates our final projection: Mark Ingram, in a close one.
We were surprised by the third-place finish of Colt McCoy. He earned 198 points more than we expected. Tim Tebow, who placed fifth (as we projected), earned 100 points more than we expected. In retrospect, it's clear that voters who voted for McCoy and Tebow were underrepresented in our sample - likely because they didn't want public disclosure.
Ndamukong placed fourth; also a surprise. We projected 1191 points, and he earned only 815. (It's worth noting that that's the highest total for a fourth-place finisher in history, according to ESPN.) Why did Suh perform poorly compared to our projection? Most likely, the voters who voted for him - essentially arguing for an historic selection of a defensive lineman - were substantially more likely to publicly disclose their ballots.
It's almost certain that it was a combination of the "embarrassment" factor and the "campaigning" factor that led to McCoy and Suh being swapped in our projection. Over the next few days, we'll continue to dig into the numbers.
In 2010, we're hoping to start our analysis a little earlier in the season, and have some exciting things planned. Stay tuned!
Kari Chisholm | December 12, 2009 | Comment on This Post (128 so far)
In this clip, Kirk Herbstreit breaks down the H------ Trophy balloting by region, election-night style. Glad to see that our approach is getting out there.
And it's not just the approach - Chris Fowler uses some of our analysis and data in his discussion. We're flattered. (But how about a consulting deal next year? :)
And Kirk, since you're now looking at the regional voting patterns, here's something to ponder: Why does the Far West - with 21% of the population - have the same number of ballots as the Northeast, with only 12% of the population?
If the 870 media votes were distributed fairly, there would be exactly twice as many votes in each state as Members of Congress (435). But that's nowhere near the case.
Oregon has 5 members of Congress, 10 Trophy voters - excellent. Alabama has 7 members of Congress and 24 Trophy voters, ten more than they're due. Nebraska has 3 members of Congress and 16 Trophy voters - also ten more. South Carolina has 6 members of Congress and 24 Trophy voters - 12 more than their population would justify. How does any of this make any sense?
Region-wide, the Far West should have 180 votes, not 145. The Northeast should have 102 votes, not 145. Is it any wonder that the Pac-10 hasn't had a non-USC Heisman winner since 1970?
Kari Chisholm | December 12, 2009 | Comment on This Post (29 so far)
On a few of the radio interviews I've done this past week, a number of folks have noted that with the votes split among multiple top finalists, it's possible that Mark Ingram would have one of the lowest vote totals in H------ Trophy history.
Not going to happen. Our projection shows that Ingram will be somewhere in the ballpark of 48% of the max points possible. Sure, that's way below the two biggest landslide winners ever - Reggie Bush (91.8%) and Troy Smith (91.6%) - and even substantially behind the last two winners, Tim Tebow (70.5%) and Sam Bradford (62.1%).
But if Ingram gets 48% of the 2778 points possible, he'll land in the neighborhood of Carson Palmer (48.0%) and Steve Spurrier (48.2%). He'll certainly not be anywhere near 2001's Eric Crouch - who only earned 27.7% of the points possible.
Kari Chisholm | December 12, 2009 | Comment on This Post (68 so far)
It's such a close race, that we'd like to hang on for a few more days and convince another hundred voters to share their picks -- but alas, time has run out.
There's also something tempting about just declaring it "too close to call", but that just won't do.
We're now up to 313 ballots, 33.8% of all the ballots. And on those 313 ballots, we project that Mark Ingram will earn 1336 points - or 48.1% of the maximum 2778 possible.
We're now projecting that Toby Gerhart will earn 1241 points - or 44.7% of the max possible. Ndamukong Suh will earn 1188 points - or 42.8% of the max possible.
Colt McCoy comes in at 939 points, or 33.8%; and Tim Tebow comes in at 295 points, or 10.6%.
We are reasonably confident that Mark Ingram will win the Trophy. With only a 3.4% advantage over Gerhart, it's far from a sure thing - as our past projections have overestimated the top-two gap by 2.6 to 4.2%. That said, it would take a fairly dramatic difference between the known and unknown ballots for Ingram to lose.
We are not, however, confident that Gerhart will defeat Suh for second place. With less than a 2% gap, it's too close to call. Even late Friday, as batches of ballots were coming in, Suh and Gerhart traded places several times.
We're very confident that McCoy will place fourth - and that Tebow will place fifth.
In addition, we're projecting with confidence that Ndamukong Suh will earn the most first place votes. We also project that Toby Gerhart will earn the second-most first place votes (even though we've tracked more #1s for Ingram.)
Here's the final regional breakdown of the ballots we've tracked down:
Kari Chisholm | December 12, 2009 | Comment on This Post (339 so far)
This time of year, there's a lot of other projections and polls out there. Seven years in a row we've picked the winner correctly - and there's no other projection that can make the same claim. Nonetheless, it's always good to check in on what the others are saying.
|Scripps-Howard||Ingram||Suh||Gerhart||McCoy||Tebow & Spiller|
|Heisman Pundit Poll||Ingram||Gerhart||McCoy & Suh||Moore|
|ESPN Heisman Watch||Gerhart & Ingram||Suh||McCoy||Tebow & Spiller|
|CBS Heisman Watch||Gerhart||Ingram||Moore||McCoy||Suh|
|NBC Experts Poll||Ingram||Gerhart||Suh||McCoy||Spiller|
|USA Today||Ingram||Gerhart||McCoy & Suh||Moore|
|Heisman Pundit, himself||McCoy||Ingram||Gerhart||Tebow||Spiller|
The Scripps-Howard News Service is a weekly poll of 10 actual voters - and has correctly projected the winner 18 of the last 22 years (and 6 of the last 7; missing out in 2008.) This year, they've got a near tie at the top with Ingram leading Suh by a single point (their closest margin ever), followed by Gerhart (who got the most first-place votes), then McCoy - and then a fifth-place tie between Tebow and Spiller.
The Heisman Pundit poll of thirteen actual voters, pegged the final order of finish this way: Ingram, Gerhart, (tie) McCoy & Suh, Moore. Of course, we know that Tebow did better than Moore, so that's not quite right.
The ESPN Heisman Watch experts poll features 15 participants (of which most, and perhaps all, are actual voters.) In their final poll of 2009, they've got a dead-heat tie between Gerhart and Ingram, followed closely by Suh. Colt McCoy is a solid fourth, followed by CJ Spiller and Tim Tebow. Since Spiller's not a finalist, we know that's also off.
The CBS Sports Heisman Watch experts poll - which features five participants (but only one actual voter, as far as we know) - pegs the order of finish this way: Gerhart, Ingram, Moore, McCoy, Suh. Since Moore's not a finalist, we know that he won't get third place -- and Tebow's at least somewhere in the top five.
At NBC Sports, they've got an Experts Poll that includes a mix of 22 experts that includes a smattering of actual voters and quite a few that we know aren't. Their order of finish: Ingram, Gerhart, Suh, McCoy, Spiller. This one missed the Suh/McCoy tie that USAT and HP have by just a single point. Meanwhile, they've got Tebow in a three-way tie for 6th with Golden Tate and Dan LeFevour.
Over at USA Today, they've got seven official voters. And while they don't bill their combined votes as a poll, we thought we'd total up the winner there. The USA Today combined order of finish is: Ingram, Gerhart, (tie) McCoy & Suh, Moore... exactly the same as the HeismanPundit.com poll.
At Sports Illustrated, they've got ten official voters. It's not intended to be a predictive poll, but here's how SI's ten votes combine: Suh, Ingram, Gerhart - followed by a smattering of votes for McCoy, McClain, Spiller and Tebow.
Over at ESPN, Ryan McCrystal has put together a Heisman Predictor that does a prediction based on stats and claims to have been right the last six years (though I don't remember seeing it before.) It's tough this year, I suspect, because the system doesn't likely consider defensive stats. Nonetheless, the final ESPN Heisman Predictor predicts a close finish with McCoy just ahead of Ingram, followed by Gerhart. (In an online chat today, McCrystal discusses it.)
Finally, the Heisman Pundit himself, Chris Huston, told the New York Times that his personal expert prediction has the order of finish as: McCoy, Ingram, Gerhart, Tebow, Spiller. And again, we know he's off by a little bit - as Spiller didn't make the finalist cut, while Suh did.
The proof, as always, will be in the pudding on Saturday night.
Update: Over at Jacksonville.com, Mike DiRocco compiled most of these polls into a single compilation (including one of our recent projections) and came up with this order of finish: Ingram, Gerhart, Suh, McCoy, Spiller.
We've got 14 more ballots reporting in, but not much change in the vote totals. Mark Ingram's lead is now 4.3% - slightly more than than it was; and Toby Gerhart and Ndamukong Suh are still in a dead heat.
We did some analysis this morning, trying to sort out "how close is too close to call?" The answer seems to be right around 4%.
So, we're going to keep tracking ballots throughout the day - and see if we can't get more clarity.
Here's the regional breakdown:
Kari Chisholm | December 11, 2009 | Comment on This Post (153 so far)
Well, folks, as we realized from the beginning, this is a barnburner of a H------ Trophy race. In an effort to sort out just how close, we've gone back and looked at all the past years using our current methodology. (We have, of course, made some major and minor tweaks over the years.)
A number of folks have asked us: what's the margin of error? And that's hard to answer. After all, a typical "margin of error" you'd get on a public poll starts with the assumption that you've got a random sample of voters. That's not the case here. We've got a self-selected group of people who choose to publicly reveal their votes. There are all kinds of reasons that people would choose to conceal or reveal their votes that would affect the outcome of the projection. It hasn't been a major factor in the past, but it's something to watch for.
Using our current methodology, we have - on average - projected the finalists' vote total to within ±4.13%. Over the last five years, that's dropped to within ±3.73%.
When it comes to projecting the vote total of the winner, over the last seven years, we've been within ±4.99% - while over the last five years, that's dropped to within ±2.58%.
As for second place, over seven years, ±4.03% - and over five years, ±3.75%. For third, over seven years, ±4.35% - and over five years, ±5.59%.
We also analyzed the GAP between the winner and the runner-up. It appears that every year except 2004, we overestimated the gap. In other words, our numbers had a larger win than the actual totals. (That's not too surprising - after all, voters who vote for the favorites are more likely to disclose their vote than ones who vote for long-shots and underdogs.) In 2008, we projected a 5.7% win for Bradford, and it was 4.4%. In 2007, however, we projected a 19.4% win for Tebow, and it was only 9.2%.
In general, over the last seven years, we've overestimated the gap between first and second by 4.23%. In the last five years, 2.58%. This has huge implications in a year where we're looking at a projected win of somewhere between 2-4%.
Of course, not every year is the same. We've had varying numbers of ballots over the years - that's why we're looking at the five-year number. (In 2002, for example, we only had 127 ballots - and we overprojected Carson Palmer's vote total by a whopping 14.2% and his 1/2 gap by 16.3%, though we got lucky and he won anyway. In our first year, we probably had some selection bias - as it was a bunch of USC fans that were helping us.)
In addition, the vote distributions will be different in three-finalist years than in five-finalist years. So it's not always an apples-to-apples comparison.
Two technical notes to emphasize: Again, these numbers come from using our current methodology to analyze old voter pools. We're not interested in retroactively changing our old projections and taking credit - just analyzing the value of the projection system we're using now.
Also, these ± numbers are a better way to evaluate our accuracy over the years than the way we've been doing it. After all, if you're +5 one year and -5 the next, you shouldn't claim that average is perfect (instead, you're averaging a ±5%.) Thanks to a couple of our site commenters for bringing that to our attention.
Here's a PDF with all the numbers if you want to nerd out on the numbers with me.
Kari Chisholm | December 11, 2009 | Comment on This Post (65 so far)
We've now tracked down more ballots than ever before. Part of the reason? It's one of the closest races ever - and we may have a Trophy winner with one of the lowest vote totals ever. (More on that later.)
For now, the lead held by Mark Ingram is holding steady. But the race between Toby Gerhart and Ndamukong Suh for second-place is a dead heat. Colt McCoy is solidly entrenched in fourth place, and Tim Tebow remains pegged at fifth.
Ingram is at 44.2%, Gerhart at 40.6%, Suh at 40.3%, McCoy at 30.9%, and Tebow at 9.7%. We'll make our final projection Friday morning.
Here's the regional breakdown:
Kari Chisholm | December 10, 2009 | Comment on This Post (139 so far)
With 252 ballots, 27% of the total, Mark Ingram continues to extend his lead - now up to 3.8%. That's more ballots than we've ever had, though still one of the most narrow margins we've sent yet between the top three candidates.
We'll have another update tonight, based on some email correspondence we've had today. At some point tomorrow, we'll make our final projection.
If you're interested in helping us track down votes, check out the Big List of 600+ Voters. In an attempt to make this easier for you - and avoid the repeated (over and over and over) postings of ballots we already know about, we've bolded the voters we don't know yet. If it's not bolded, we got it.
- Post every vote you find. If you only post the ones you agree with, the projection will be wrong.
- Whenever voters identify their vote 1-2-3, post all three. If you only share the #1, the projection will be wrong.
- Be as specific as you can about the source. If there's a link, post it. If it was on-air, provide the station, the date, and the time. We need to be able to verify what they said (that it was a vote, not a prediction, or a hope), and that they claimed to be a voter. (There are bloggers and others out there who don't have votes, but are happy to confuse you.)
Thanks for all the help. We couldn't do this without you! But please, don't post votes we've already got. That just makes the updates slower. (No, seriously, stop posting about Monte Poole, Spencer Tillman, and Eddie George. We got 'em.)
Kari Chisholm | December 10, 2009 | Comment on This Post (247 so far)
Over at Rivals.com, they've got the story -- Toby Gerhart's been checking out Stiff Arm Trophy:
The stress of exam week hasn't stopped Gerhart from getting caught up in the Heisman hype. Gerhart admits it's tempting to see where he stands in various mock ballots or in the polling at www.stiffarmtrophy.com, which collects advance ballots from various Heisman voters.
"Some of my roommates are consistently updating me," Gerhart said. "They come running in and saying, 'Stiffarmtrophy has come in with new updates, and you're second now,' or 'You and [Alabama running back Mark] Ingram are tied.' There's always constant buzz from kids on campus."
It looks like
Mike Mark Ingram is starting to pull away from Toby Gerhart and Ndamukong Suh. At 242 ballots counted, he's now up to 43.8%, more than a full point higher than he was at 226 ballots. Gerhart and Suh are locked in a close race for second - Gerhart 40.6%, Suh 39.2%.
In addition, there's been a lot of attention paid to the hard count of votes that we've tracked down. Some folks have even reported in the last couple of days that Suh is "leading in the count", utterly ignoring the regional adjustments made in our projection. (You can't just look at the hard count, since our count has 54 votes in the Southwest and only 18 votes in the Northeast - and we know that actual voter turnout will be roughly the same.)
Nonetheless, for the first time, our hard count now shows Mark Ingram leading - with 325 points. He and Gerhart continue to trail Suh in the number of first-place votes, while Suh trails the two of them in the number of ballots on which he appears.
(Argh, sorry Mark! That's what you get when you're trying to write a post while talking on the phone to your pal Mike.)
Kari Chisholm | December 10, 2009 | Comment on This Post (112 so far)
So, it's come to my attention that folks in Nebraska -- apparently suffering from serious cabin-fever brought on by massive snowfalls -- are up in arms, with the mistaken impression that Trophy winner Johnny Rodgers put Ndamukong Suh third on his ballot.
In fact, Rodgers simply stated that he has Suh somewhere on his ballot. For the purposes of our projection math, we count a statement like that as a #3. But doesn't mean that he voted #3.
Rodgers even went so far as to call in to a local
Lincoln Omaha sports talk show to set the record straight. Listen here.
So, which voters are we still missing? Out of the 926 Trophy voters, we know the identities of just over 600 of them. Of those, we've now tracked down the ballots of 226.
Here's the list of the nearly 400 people who we know are H------ Trophy voters (or, at minimum, were voters very recently) but for whom we don't yet have votes.
Start by using the power of Google. Dig around, look for blogs, twitter feeds, facebook pages, and - last but not least - email addresses. If you do contact one of these folks directly, please be polite. There's no point in arguing with them; the ballot deadline has come and gone, and they couldn't change their mind if they wanted to.
Also, if you're going to be helpful do three things: #1, list every vote you find (not just the ones you agree with; the point isn't to "win" the StiffArmTrophy projection, it's to get it right.) #2, list all three spots on the ballot. #3, state your source - if there's a link, provide it. We need to be able to confirm that these folks are ACTUAL voters, and not just folks talking about "my vote" as a hypothetical thing.
And remember, there's another 300+ voters whose identities are still secret. We'd love to track down a few of them - but we'll be especially careful to verify that they're really a voter.
Without further ado...
Now at 226 ballots - 24.4% of all Trophy ballots - Mark Ingram's lead is holding at 42.4%. Toby Gerhart remains narrowly behind him, but his lead over Ndamukong Suh has shrunk from 3.1% to 1.6%.
Colt McCoy's total dropped ever-so-slightly. We're now confident that McCoy is going to remain in fourth place. Tim Tebow remains in fifth.
It is, of course, fairly stunning that 2008's second-place and third-place finishers will finish fourth and fifth in 2009. Most observers expected a huge Trophy battle royale between the two of them (along with 2008 winner Sam Bradford).
The regional breakdown continues to show substantial regional differences - though with Ingram leading the pack in three of the seven regions (plus leading our list of voters whose region we don't know, mostly national broadcasters.)
Kari Chisholm | December 9, 2009 | Comment on This Post (143 so far)
Since I'm stuck in jury duty today - excuse me, doing my long-delayed civic duty today - I thought I'd give you a little analysis to chew over.
A lot of folks have been asking me questions like "Are there people voting for Ingram and Gerhart because they're both RBs?" and "Are the Big-12 voters voting for McCoy and Suh together?" etc. (And until last weekend, lots of those questions were being asked about the SEC's Tebow and Ingram, or QBs Tebow and McCoy, too.)
Among the first 208 ballots we've tracked down, 169 have the two of the five finalists in the top two spots. (The rest have other #1s and #2s, or we don't know their #2 at all.)
The most common 1/2 ordered combos are Suh/Gerhart (26), Suh/Ingram (25), Ingram/McCoy (20), Gerhart/Ingram (18), Gerhart/Suh (13), and Gerhart/Ingram (10).
- The most common "alliance" was unexpected - 39 ballots had Gerhart and Suh together in the top two spots. Clearly, these are voters who like a non-traditional approach; they're not looking for "the best player on the best team."
- 28 ballots had Ingram and McCoy on them -- with both teams headed to the BCS title game, those are your "best player on the best teams" voters.
- The RB-alliance of Gerhart and Ingram does seem relevant, as another 28 ballots had them together.
- There doesn't appear to be much of a Big-12 alliance, with only 8 ballots with Suh and McCoy together (only 2 have McCoy above Suh.)
- The most common combo with Tim Tebow in it was McCoy/Tebow, with 9 ballots. His overall poor showing means he's not in a lot of top-two pairs, so there's not much of an Tebow/Ingram SEC alliance at all, just four ballots.
Kari Chisholm | December 9, 2009 | Comment on This Post (84 so far)
We've now tracked down 208 ballots, containing a total of 578 votes. That's a sample of 22.4% of the entire H------ Trophy electorate. That's more votes than we've had in four previous projections. Of course, this vote is still looking very close.
Mark Ingram now leads Toby Gerhart by just over a single percentage point. Ingram is at 42.2%. Gerhart is at 40.9%. We're not quite confident in making a projection between the two of them. Either could still win this thing.
However, Ndamukong Suh is now at 37.8% and Colt McCoy is now at 31.6%. We're confident in making the relatively limited (and obvious) statement that neither of those two gentlemen will be the winner of the Trophy. (And, of course, we noted earlier that Tim Tebow is a lock for fifth place - holding stead at 11.3%.) We'll reserve judgment on exactly where Suh and McCoy will land in the order of finish.
Here's a regional breakdown.
We seem to be getting deluged with comments from Cornhusker fans telling us that our projection is bunk because our math is off. After all, they say, Ndamukong Suh has the most points - how can he be in third place?!
So, rather than simply saying, "People, read the site!" we'll address it directly.
First, we see this every year. And every year, somebody cries foul. And every year, we've been right. (Last year was especially tough, since I was one of the few projecting a loss by Tim Tebow - even though he would have the most #1 votes.)
Second, if we simply counted up the number of ballots we track down - and then extrapolated a total - that would be a) too easy, and b) wrong. In fact, that's exactly what we did in our very first year. We got lucky, and correctly projected Carson Palmer as the winner, but we had Larry Johnson at #5 - not #3, as he actually ended up.
After that first year, we realized that the projection actually has to be six separate regional projections that are then added up. Why? Because you get lots and lots of public declarations from the regions that have candidates - and relatively few from the regions that don't, even though they all still vote.
This year, for example, so far we've tracked down 44 of the 145 actual ballots in the Southwest region - but only 13 of the 145 in the Northeast region. But since we know that both regions will be roughly equally represented in the final total, we have to adjust the math - so that the Southwest region doesn't appear to be three times the size of the Northeast.
A few of the other questions:
- "Do these projections assume that the people who have not responded are going to be similar to those that have?" Yes. That's a critical assumption. And so far, it's borne out to be true. We hear from very few former winners - so I've always worried about projecting that group of voters, but they're relatively few compared to the media voters.
- "What projection will you give voters who voted before Saturdays games?" About 10% of voters voted prior to the final Saturday games. I wish I could account for those voters, but the date of their ballot isn't something we know. Just have to assume that those votes will be equally represented in the known and unknown ballots like all the others. (Of course, if those people all voted Tebow/McCoy, and now they're too embarrassed to publicly declare, well, that could impact that critical assumption above - that known ballots accurately reflect unknown ballots.)
- "Dude! You're WRONG! Everybody knows MY GUY is the best player in the country! Moron!" I do love those comments. This is just a little reminder that I don't award the H------ Trophy, nor do I have a vote. I'm just a math and politics nerd in Portland, Oregon that has an obsession with college football. (Pro football nerds, check this out.) Rest assured, I think your player is a great guy and deserves the Trophy. (Seriously, spending all this time on it, you can't help but gain a healthy respect for all the leading candidates.)
Well, this is getting more and more interesting - and less and less clear - as time goes on. We're now showing Mark Ingram and Toby Gerhart in a virtual dead heat, separated by just couple tenths of a percentage point - 41.4% to 41.1%.
Ndamukong Suh is now solidly in third place, with 36.8% - and Colt McCoy clearly in fourth, with 31.6%. As we've been tracking all along, Tim Tebow is a lock for fifth.
Update: Here's the latest regional breakdown.
Well, we have a new leader in the Stiff Arm Trophy projection. Having added 42 more ballots to the mix, Mark Ingram is now narrowly leading - with 41.4%.
Toby Gerhart is close behind at 39.7%, followed by Ndamukong Suh at 37.2%. Colt McCoy seems solidly entrenched at fourth-place with 31% - and Tim Tebow is just happy to get a front-row seat in New York, with 12%.
So, what changed? A whole lot of ballots that included Ingram in the #2 spot - especially outside of the South. In terms of overall points, Ingram is now leading in the South, the Mid-Atlantic, the Northeast, and is in second place in the Midwest and the Far West. In the Southwest, he barely trails McCoy (though Suh is the huge leader there.)
Here's a regional breakdown of the points in each region. Keep in mind that the total number of ballots we've tracked varies a bit - and our projection methodology corrects for that (which is why you can't just add up the numbers.)
Kari Chisholm | December 8, 2009 | Comment on This Post (133 so far)
Critics of the H------ Trophy balloting process have long complained that too many voters vote prior to the final games of the season. This year, with four finalists playing in the Big-12 and SEC championship games, those critics are louder than ever.
Kirk Bohls, of the Austin American-Statesman, noted before those games that the Trust told him that 10% of all votes had been turned in early. He tweeted:
Just blogged at Statesman.com that about 10 percent of the 926 accredited Heisman voters have already turned in their ballots. Help Colt?
Given the disastrous game McCoy had (0 TDs, 3 INTs), it's likely that those 10% of ballots had him higher ranked than the subsequent 90%.
And sure enough, at least one voter regrets voting early.
Doug Krikorian of the Long Beach Press Telegram announced his vote on November 30:
Just cast my Heisman Trophy ballot, and voted for Texas' Colt McCoy, Florida's Tim Tebow and Stanford's Toby Gerhart, in that order.
But in retrospect:
After watching Colt McCoy's lame performance against the Cornhuskers, I now regret casting my Heisman Trophy vote for him.
Now that the Trust has gone to all-electronic balloting, is there any reason to open up the voting system prior to the finish of the games? You can't control when folks make up their minds - but you can ensure that they'll have all the information.
Well, the finalists have been announced by the H------ Trophy Trust. As we expected, they are Toby Gerhart, Mark Ingram, Colt McCoy, Ndamukong Suh, and Tim Tebow. The Trust announces them in alphabetical order, but we're glad to see that their top five matches our top five.
With 121 ballots, we're showing Gerhart maintaining his hold on the top spot with strong showings in a number of regions and a projection of 39%. Suh and Ingram continue their virtual tie at 38%, and McCoy right behind at 32%. Tim Tebow is solidly in fourth - but well ahead of the rest of the field at 13%.
Kari Chisholm | December 7, 2009 | Comment on This Post (160 so far)
Well, we've crossed the threshold of 100 ballots - and things are beginning to take shape. As close as this is, I won't be comfortable with a projection until we're past 200 ballots.
It looks like Toby Gerhart is starting to put some distance between himself and the field. He's now at 41% of the vote. Mark Ingram and Ndamukong Suh are in a virtual tie at 37%. Colt McCoy is in fourth at 30%.
Gerhart continues to show strongly across four regions - picking up roughly a fifth of his overall total in the Far West, Southwest, Midwest, and South.
Kari Chisholm | December 7, 2009 | Comment on This Post (64 so far)
A couple of folks on Twitter have noticed that the vote totals add up to more than 100%. Here's the explanation.
Each H------ voter identifies their #1, #2, and #3 pick. Top vote gets 3 pts, second place gets 2 pts, third gets 1 point. Therefore, each ballot has a total of six points - but the most any one candidate can get is three points.
A unanimous #1 selection would get three points on every single ballot. To make things as clear as possible - 100% is the maximum possible points each candidate could earn.
While each candidate can only get three points, there are six points per ballot. As a result, the total projection values will add up to 200%, not 100%.
Kari Chisholm | December 7, 2009 | Comment on This Post (42 so far)
For years now, I've been talking about the regional bias in the H------ Trophy balloting. If you haven't read it yet, dig into the item I wrote in 2003 about the structural regional bias that exists.
It's not as simple as East Coast voters not staying up late to watch evening games in the Pacific timezone. It's not as simple as Southern voters thinking the Pac-10 is soft. There's always been - and always will be - a certain amount of regional favoritism in the voting; whether deliberate or accidental.
Rather, the H----- Trust has simply shortchanged the West in terms of the number of votes. Here's an easy and crystal clear example. This year, we learned that the state of Alabama has 24 media voters. Last year, I learned that Oregon (where I live) has 10 media voters.
Here's the rub: According to the 2007 Census estimate, Alabama has 4.58 million residents. Oregon has 3.69 million residents.
So, Alabama has one H------ voter for every 191,000 people. Oregon has one voter for every 369,000 people. Alabama is 1.25 times the size of Oregon, but has 2.4 times as many votes.
That, in a nutshell, is structural regional bias. And it should be fixed.
Kari Chisholm | December 7, 2009 | Comment on This Post (39 so far)
A big update this morning - mostly based on a whole lot of email correspondence. We're now up to 79 ballots, and this thing isn't getting any clearer.
Ndamukong Suh leads our projection with 38.6%, in a virtual tie with Toby Gerhart at 38.4%. Mark Ingram is very close behind at 36.9%, and then Colt McCoy at 34.7%. Tim Tebow trails at 12.6%.
The H------ Trust is set to announce the finalists, via ESPN's SportsCenter, at 6 p.m. Eastern. In years past, they've always announced three to five finalists - with three being the norm in recent years. I'll be stunned if they limit to three this year, expect four, and won't be surprised if it's five.
I don't expect season-long crowd favorites CJ Spiller, Kellen Moore, and Case Keenum to make the cut.
Well, we're now up to 46 ballots that contain 103 votes. It's Gerhart 56.6%, Suh 39.0%, Ingram 27.9%, McCoy 22.2%, Tebow 14.4%.
Astute observers will note that Toby Gerhart is leading, even though we've actually tracked down 20 ballots that list Ndamukong Suh in first place - and only eight with Toby Gerhart at #1.
Why? Two reasons: First, Gerhart gets #2 and #3 votes on twice as many ballots (15) as Suh does (8).
Second, and this is key: A huge percentage of the Suh votes are coming from the Southwest region - specifically, Nebraska. As I mentioned earlier, a large number of the early ballots will come from folks campaigning - and today, that's been all about Suh.
In fact, 49% of all the votes that we've tracked for Suh are in the Southwest. Beyond that, he's also concentrated in the Midwest (23% of his votes) and the South (20% of his votes).
By contrast, Gerhart's picking up votes everywhere. 26% of his support is in the Southwest, 21% in the Midwest, 19% in the Northeast, 17% in the Far West, and 11% in the South.
If this trend holds up - with Gerhart picking up lots of votes geographically spread out, while other candidates are largely regional... well, that's very good news for Toby Gerhart.
One last note: We have, so far, tracked down ZERO votes from the Mid-Atlantic region. As I said, there's a long way to go.
Kari Chisholm | December 6, 2009 | Comment on This Post (12 so far)
I know that a lot of you are itching to help find more voters - and more votes. Since the voter list is secret, tracking down previously-unknown voters is tough (and mostly done via Google, or listening to local sports talk radio).
But over the years, I've built a list of some 570+ known voters. We may know who they are - but have no way of contacting them. I've been working on finding Twitter accounts for H------ voters. See the big list of 175+ voters on Twitter here.
If you want to help, here's one thing that needs doing: I'd like to track down Facebook pages and profiles for as many voters as we can. After all, some voters may post their votes there - and we'll be able to use Facebook to ask 'em directly.
To make this easy, I've created pre-built links that go directly to Facebook's search function. Just click, and then try and figure out if the results match anyone that looks to be a sportswriter (or a former Trophy winner.)
To minimize overlap, start wherever your name would be alphabetically - and work down from there as much as you can. Whenever you identify a Facebook "fan page" or personal profile for a voter, post it in the comments on this thread. I'll go through and remove the ones we get from this list.
Here's the list...
In our latest projection, Nebraska's DT Ndamukong Suh has surged to a lead - followed closely by Stanford RB Toby Gerhart. Clustered together behind them are season-long favorites Colt McCoy, Mark Ingram, and Tim Tebow.
But before Husker and Cardinal fans celebrate, let's remember how this thing goes in years past. The early votes (which are, admittedly, coming in later this year) tend to be the ones from H------ voters that are campaigning for a regional favorite or some other nontraditional pick.
Beyond that, consider the psychology: If you're a H------ voter that thinks that Tim Tebow or Colt McCoy earned the trophy after a season's worth of good games, are you really rushing to announce that right now? That's a recipe for all kinds of hate mail in the inbox. So those folks are going to be sitting tight today.
We're only 28 ballots in. Ballots are due at 5 p.m. Eastern on Monday, and that's when we'll see a real surge in public reports. Once we get around 100 ballots, we'll have a good sense of what's happening out there. Once we're up over 200 ballots, I'll start feeling comfortable making a projection.
Kari Chisholm | December 6, 2009 | Comment on This Post (8 so far)
Wow. What a final weekend of college football. I don't have to tell you - but those were some amazing games: Texas/Nebraska, Alabama/Florida, Cincinnati/Pittsburgh, East Carolina/Houston, Clemson/Georgia Tech, Fresno State/Illinois. Hey, even the two unranked-over-ranked upsets in the Pac-10 (Arizona over USC, Washington over Cal) that kept up the record of high-level Pac-10 parity in 2009.
This is a H------ Trophy site, not a playoff site, but if you'll allow me a moment of personal privilege... After watching all those amazing conference championship games - the main thing that comes to mind is this: Wouldn't it be incredible to have four more weeks of awesome matchups like that (with either a 16- or 12-team playoff) culminating in a true national championship?
OK, back to the Stiff Arm Trophy.
Obviously, even though ballots have been out for a couple of weeks, almost everyone was holding their ballots until yesterday's games were over. As you see and hear H------ voters making final vote declarations, be sure to post 'em on our voter tips thread.
Expect a fast rush of votes in the next 48 hours. I'll be updating the projection 2-3 times a day. By mid-week, we should be up over 100 ballots and things will have taken shape. By Thursday or Friday, we should be up over 200 ballots and I'll be comfortable making a final projection.
This year, we've added Twitter to our toolbox as we attempt to once again project the winner of the H------ Trophy. I'll be tweeting (and retweeting) lots of Trophy news - along with commentary from reporters and fans that's insightful, funny, bizarre, provocative, etc.
And while you're at it, check out the list we built of over 160+ actual H------ voters that are on Twitter themselves. (Admittedly, quite a few aren't tweeting much.)
I can hear you now... "Wait a minute! Those numbers are wrong! You suck!"
Just so everyone's clear here: These numbers are completely WACK right now. They're very early projections - based on just a handful of votes. It's kind of like a political election projection when the first three counties have come in - but the major metro areas are still out.
Hang in there, folks. Once we get upwards of 50 ballots, it'll start to take shape. Once we're over 100 ballots, it'll start to look real. We'll be happy once we're over 150 ballots - and the last two years, we've had a total of 249 ballots.
Some voters will vote this week. Most will vote after the final weekend games. The deadline, though, isn't until Monday, December 7.
OK, if you want to promote your favorite Trophy candidate - or trash talk someone else's -- do it right here. Anything violent, racist, or otherwise uncool will be deleted. Run-of-the-mill idiocy will be allowed.
If you have official voter tips, post 'em over here.
Use this space to post tips about official Stiff Arm Trophy voters that have declared their votes. If possible, give us a link to a web page where it's printed -- or at least tell us where on the air you heard it.
We can't do this without you. Thanks!
(Please keep this limited to voter tips. Do your trash-talking here.
This thread closed. Submit new voter tips here.
The H------ Memorial Trophy is a registered trademark of the H------ Memorial Trophy Trust. This site is not affiliated with the Trust, not even a little. We're not even using the H------ word, since they don't want us to. All other trademarks are the property of their respective owners.
Copyright to and responsibility for all posts and comments are owned by their respective authors.
Obviously, the posts and comments here are the views of their authors, and not of anyone else.
While we're strong believers in free speech, we reserve the right to delete comment spam or other offensive material. Our contributors, however, reserve the right to embarass themselves in public.