Kirk Ferentz and the Myth of the Percentages

[NOTE: Before you dive into this, read HEC's stellar analysis of situational percentages below.]

In his Saturday post trying to explain the inexplicable late-game decisionmaking of Kirk Ferentz, particularly at the end of Iowa's loss to Iowa State, Marc Morehouse said that Ferentz was playing the percentages, and that we all know it works more often than it doesn't.  Two-minute drills are difficult and risky, after all, and Ferentz's late game coaching experience, bend-don't-break defense, and pounding running game work better when they are not rushed.  When Ferentz shies away from a possible game-winning drive despite having 1:17 on the clock, two timeouts in his pocket, a tie on the scoreboard, mismatches on both sidelines, a kicker who had already connected from 50 yards and was 5-for-5 on the season, and an opposing offense that hadn't been stopped in three quarters, he's just following the numbers.

Not risking a late drive despite having virtually every circumstance in your favor might be MANBALL dogma, passed down from Schembechler to Carr to Tressel to Ferentz, and it might indeed be smart in aggregate to go hyperconservative in close games.  The problem is that, while "the percentages" worked for Carr and Tressel, they quite clearly don't apply for Ferentz and his "unique" brand of endgame decisionmaking.  On the contrary, Kirk Ferentz is an especially poor coach in close games, and his philosophy is counterproductive on both sides of the ball in late-game situations.

CLOSE GAMES

Earlier this offseason, as part of its study of Clemson, Football Study Hall looked at the record of highly-ranked teams late in games.  FSH ranked the 20 best teams in overall performance from 2005-10 based on winning percentage in games decided by less than one possession.  These are the games where Ferentz's vaunted clock management and playcalling system should shine through, where we should see the maestro's master plan pay off, as it did during the 2009 campaign.  We were shocked to find that Iowa finished nineteenth out of twenty, with a 12-19 (.387) record over that period.  In fact, Iowa was one of only two teams that was more than one game under .500 in one possession games over that period of time (the other being the aforementioned Clemson).  If you take 2009 out of that total, Iowa is a putrid 8-17 (.320), one of the worst records in all of football.

This isn't just isolated to the FSH top 20 or the last six seasons.  Following is a list of the records of Big Ten teams in one-possession games since Kirk Ferentz took over at Iowa in 1999:

Compare that to a chart of overall wins and losses in that same period of time:

It turns out that, except for the living embodiment of statistical anomaly that is Northwestern, winning percentage in close games is a fairly accurate barometer of overall conference success.  We should expect that winning teams would have a winning percentage at or slightly above .500 in close games; in a close game where the teams are basically equal, the result is basically a coin flip, and winning teams probably get those flips at a slightly higher rate than others.  In other words, while we would expect a decrease in winning percentage in close games when compared to overall winning percentage, a winning team should not have a large spread between the two or fall drastically below .500.

This is why it is staggering that, not only does Iowa's winning percentage drop by 15% in close games (among Big Ten teams, only Penn State is worse over that time, and their coach doesn't even know when the game is over), but it falls from nearly 10 points above .500 to 5 points below.  While good teams play close games at a coin flip or better, Iowa is measurably worse.  The conclusion is that, contrary to popular opinion as spun by Kirk Ferentz, Iowa is not especially good in close games; quite to the contrary, given their status in the Big Ten and college football as a whole, they are extraordinarily bad.

THE TWO-MINUTE DRILL

At Big Ten Media Day in Chicago, Kirk Ferentz was asked specifically about the two-minute drill and its chance of success (I think the question was from Morehouse, unsurprisingly):

INTERVIEWER: Last year, fourth quarters, when it comes down to a two-minute drill, I don't know what the percentage is, 50/50?

COACH FERENTZ: It's not 50/50. Historically, I'd say it's probably less than that.

INTERVIEWER: More like 25?

COACH FERENTZ: Yeah, it's closer that way than 50, so you know, and I'm not minimizing that. But you know, one thing you want to do is just be careful about not oversimplifying what the issues are. Sometimes it's two minute drills. Sometimes it's two minutes defense. One minute. And then I'd also say, there have been a lot of games, a couple that are very prominent in my mind right now, and we'll share with the team. And a lot of times it's what you do in the first half. You squander some opportunities that are there for you, but you're not ready to go. And in the last two minutes you're thinking back to those first couple of series where you had where you just really did a poor job. You know, we wouldn't be in this situation right now if we had done better there. And I can think of two games right off the top of my head like that. So it's, and there are a lot of things that go into it. But I can tell you we've spent a lot of time this season looking at that. And clearly, we have to do better. And you know, from every standpoint, coaching and playing, it's critical.

As a quick aside, in the aftermath of last week's loss, my favorite part of that quote is the last two sentences.  They've obviously spent a lot of the offseason looking at that, knowing they have to do better, and their grand answer is to simply not run the two-minute drill at all.

ANYWAY, on his proposed percentage of two-minute drills completed, Kirk is mostly right: In his twelve seasons at Iowa, Ferentz's teams have now run eight series starting inside the last two minutes with a chance to change the outcome of the game.  They have been successful in two of those eight (Tate to Holloway at the 2005 Capital One Bowl and Stanzi to McNutt in 2009 Michigan State).  In the other six instances (2005 Northwestern, 2008 Illinois, 2010 Wisconsin, Northwestern, and Ohio State, and 2011 Iowa State), they failed to convert.  Likewise, Iowa has had to defend against a two-minute drill that could change the outcome of the game four times in the last twelve years, and has successfully stopped three of those four drives (2002 Purdue, 2004 ISU, and 2009 Michigan); only in the 2006 Syracuse game, where the Orange completed a pass, picked up a 15-yard penalty, and knocked through a 41-yard field goal to tie in the last 28 seconds, did an opponent convert a scoring drive in the last two minutes.

Where Ferentz's "percentages", and Iowa's performance, breaks down is in the three-minute and four-minute drills.  On the offensive side of the ball, Iowa is hardly better with the additional time.  Ferentz's teams have converted three three-minute drives (the aforementioned two last-second wins plus Banks to Clark against Purdue in 2002) in eleven opportunities (the previous eight plus a three-and-out against Wisconsin in 2007, the non-drive against Pitt in 2008, and Stanzi's ill-fated winning drive against Northwestern in 2009).  That's a 27% conversion percentage.  In the four-minute drill, Iowa is 6-for-18 under Ferentz, for a rather pedestrian 33% conversion rate.

Where Iowa breaks down, though, is in three- and four-minute defense.  While the Hawkeyes have  been successful defending two-minute offenses in all but the strangest of circumstances, they have been embarrassingly bad when the offense has slightly more time.  Opponents are six for twelve against Iowa in the three-minute drill and eight for seventeen in the four-minute drill.  That's a conversion rate nearing the half that the interviewer postulated.  It was also the primary cause of defeat for Iowa in 2010 (we haven't seen it yet in 2011 -- ISU's final scoring drive started with just over 5:00 to go -- but it doesn't look promising).

THE SPECIFICS

The inevitable question is, what of the great seasons?  If Ferentz is really so bad at late-game strategy, how has Iowa been successful?  The answer to that question reveals the true "percentages" that Iowa exploits.

In 2002, Iowa was 1-for-1on late-game offense (Purdue) and 1-for-2 on defense (successful against Purdue's counterattack, not so much against Penn State).  The PSU three-minute drill tied that game, a game Iowa eventually won in overtime.  The only two other close games on the schedule that season were Miami (OH), where Iowa led by 12 with 10 minutes to go (remember this; it becomes a theme) and Iowa State, where the Hawkeyes lost after a late touchdown drew them within 5.

In 2003, Iowa only played two one-possession games (those were the days, right?), wins against Michigan and Wisconsin.  Against Michigan, Iowa built a two-possession lead in the fourth, only to lose a score back to the Wolverines.  Against Wisconsin, the defense stopped the Badgers at the 5 yard line with 21 seconds to go.

Iowa played five close games in 2004 and won them all, but only one -- Tate to Holloway -- required a four-minute drive.  In two games, Purdue and Minnesota, Iowa held a two-possession lead in the fourth and let the opponent land a late score.  In the other two, Iowa successfully ran out the clock.  And the fifth, that LSU win, was less a two minute drill than a Hail Mary necessitated by Ferentz not realizing the clock started after a false start, the first of many late-game clock catastrophes under his watch.

The worm turned in 2005.  After a successful four-minute drill to send the Michigan game to overtime (an eventual loss), Iowa was 3-for-3 since 2002 in late game drives. The defense was 5-for-6 in stopping such drives over that stretch.  The turning point came against Northwestern, where Iowa blew a 13-point lead in the last 10:58, then was unable to mount a drive of its own to retake the lead.  That game set of a string of late-game futility that continues to this day; since that loss, Iowa is just 3/14 on late-game offense, while allowing opponents to score on 6 of 10 attempts.  For whatever reason -- my guess would be that other teams started realizing that Ferentz's abilities in one-possession games were the soft underbelly of Iowa football and stopped trying to trade punches with Iowa early -- the bloom came off Ferentz's rose that day.

2006 and 2007 were a comedy of late-game errors: The Brodell bobble against 2006 Indiana, the fifth field goal from 2007 Iowa State, that debacle against Syracuse.  There were also late-game drives that went nowhere against Wisconsin in both seasons.  The 2008 team played five close games and lost every game but one.  For a brief moment in 2009, it appeared as if Daniel Murray's kick had lifted the Northwestern curse.  Iowa stopped UNI under ridiculous circumstances, held back a late Michigan drive, and countered MSU's successful three-minute drill with one of its own.  Yet, for all the close wins in that season, the fact is that Iowa typically followed a pattern in getting there: Build a two-score lead and bleed the clock dry.  The Hawkeyes only converted one late-game drive, and were just 2-for-3 in stopping opponents from doing the same.

If Ferentz had fixed his flaws in 2009 -- on close examination, he hadn't -- they reappeared in 2010.  Late-game drives got nowhere against Arizona, Wisconsin, Northwestern, and Ohio State.  The only successful late-game drive, against Indiana, was again less a drive than a bomb to McNutt, and Indiana should have converted a three-minute drill of their own to win had it not been for a dropped touchdown pass.  Worse still was the return of Ferentz's botched clock management.  The last series against Arizona stalled when Ricky Stanzi was sacked 47 straight times and Ferentz inexplicably hesitated to use his timeouts.  Against Wisconsin, the offense looked like the Keystone Cops, as Iowa ran the ball twice for first downs without having a set play ready for the next snap and Stanzi inexplicably checked down to receivers with no chance of getting out of bounds.  Against Northwestern, Iowa lost after leading by two scores with less than 10 minutes to play for just the second time under Ferentz, capping its implosion with a disjointed two-minute drill where, in a college game where the clock stops on first down, the offense somehow ran 15 plays and completed a pass for more than ten yards exactly once.  And against Ohio State, needing a field goal to tie and with two timeouts to use, the offense sputtered to a halt, ending with a completely characteristic 19 yard completion on 4th and 25, both timeouts still firmly lodged in Ferentz's back pocket.

THE CONCLUSIONS

There are two inescapable conclusions to draw from this: Iowa is nearly unstoppable when it holds a two-possession lead with 10 minutes to go.  Since 2002, Ferentz's teams are 9-2 in one-possession games where they held a two-possession lead at that point.  Furthermore, this analysis only considered games with a final margin of eight points or less; there are plenty more where the opposition never got that close again.  The grinding offense and bend-don't-break defense are excellent tools for finishing out a win in those circumstances.  However, where Ferentz doesn't hold such a lead, where the defense is forced to make a series of stops in the last ten minutes and the offense has to do more than bleed out the clock, Iowa is bad and trending worse.  These are the situations where a coach with sound clock management, offensive variety, and a defensive strategy that doesn't allow an opponent the opportunity to move into position to make a game-changing play win games.  It's also where Ferentz repeatedly fails or, as we saw Saturday, simply concedes his best opportunity to win.  This has nothing to do with the "percentages," because on the whole (as HEC correctly pointed out) the percentages are in his favor in that situation EVERY TIME.  This has everything to do with the fact that Kirk Ferentz simply doesn't know how to properly manage his team in general -- and, frankly, his offense in particular -- late in games.  Saturday's surrender at 1:17 wasn't Ferentz showing he had no faith in his team.  It was Ferentz showing he had no faith in himself.

The strangest fact is that Ferentz has quite clearly moved toward more conservative game plans and play calls over the last decade.  In his early days at Iowa, his teams didn't play a particularly large number of close games, and those it did play were largely of the two-possession lead variety.  His philosophy has clearly shifted, though; since 2007, Iowa has played at least five close games every season.  They played a staggering thirteen one-possession games in 2009 and 2010.  Even when the on-field results are a mismatch, Iowa still routinely wins by 10-20 points; the 2010 Orange Bowl win over Georgia Tech might well be the most lopsided 10-point win in the history of the sport.  It has been covered ad nauseam, but Ferentz's philosophy has apparently shifted to keeping the game close and playing for the close win, primarily by building a lead, milking clock on the ground, and relying on his defense.

But isn't this exactly the worst philosophy for a coach who has clearly and repeatedly shown a complete inability to understand late-game strategy or clock management?  It might keep Iowa close against the Ohio States and Wisconsins, but it also virtually guarantees that Iowa will not build a sustainable lead in those games, requiring the team to again rely on Ferentz's late-game strategic and management weakness.  It also keeps inferior opposition in play while Iowa sits on the ball like an egg.  This is what happened Saturday, where Ferentz, in quite possibly his dumbest decision in a game of dumb decisions, built a 10-point lead early in the second quarter -- let me repeat that: THE SECOND QUARTER -- and tried to run out 45 minutes of clock with Marcus Coker.  Only when ISU scored on back-to-back drives, tying the game at the half and bringing a muted Ames crowd back into the mix, did he give up his delusional strategy.  If this strategy was sustainable, and I would say that it hasn't been since the 2009 Minnesota game where it made its first real appearance, it is no longer.  This isn't statistics or percentages, because the statistics and percentages run opposite to it.  This is a flawed philosophy that took down a 2010 Iowa team loaded with talent and threatens to destroy a 2011 team that isn't.  The problem isn't the players.  The problem is Kirk Ferentz.