beevo

February 4th, 2012 at 8:31 PM ^

This isn't a spot-light lover such as some of the others we've dealt with in the last couple of days.  The kid was solid on Iowa and then there have been a couple of seismic shifts in his recruiting world.  I think he is doing absolutely the best thing by not making a quick decision.  He really needs to study incoming classes and depth charts...offensive system....line coachses...OC...head coach....campus.....academic programs....and so on.  I still like our chances here better than I did with Diamond.  He would round out a superb OL class.

Joined: 05/13/2011

MGoPoints: 1120

Muttley

February 5th, 2012 at 11:24 AM ^

A "Convert-on-Second-Down". What else was there but a "Convert-on-First-Down"? And then, the coup de grace of all Conversions, the sinister "Score-on-Special-Teams-And-Defense".

joeyb created a slight breach of etiquette by skipping the "Convert-on-First-Down" and going right for the throat!

Joined: 07/07/2009

MGoPoints: -73245

HeyGuy14

February 4th, 2012 at 9:07 PM ^

So Kozan?!?!?! whats the deal? when is he committing? I want to start getting excited about the 13' class but for some reason cant get closer yet!

Joined: 12/13/2010

MGoPoints: 280

cozy200

February 4th, 2012 at 9:24 PM ^

Real sports with bryant gumble did a piece on this. About a high school coach who never punts and always onside kicks. Math supports going for it. His team by the way scored 3 tds off of onside kicks in one game and was up 21 - 0 in the first 3 minutes. Guy had like 4 different onside kick packages. Yes my grammar is terrible

Joined: 01/02/2011

MGoPoints: 2430

T-town Wolv

February 4th, 2012 at 9:26 PM ^

are getting a little too hyper about recruiting for football.  Hoke & co will do their jobs just be patient and don't get too antsy. 

Also all that 4th down jazz is about math not science. Most coaches have probability sheets sitting next to them at all times. Even Mathlete did a chart for us last year. http://mgoblog.com/diaries/never-punt-denard-fourth-down-strategy-revisited

Sidenote: My 100th post!!!!!!

Joined: 12/01/2010

MGoPoints: 175

The Barwis Effect

February 4th, 2012 at 9:28 PM ^

The January 24, 2012 edition of HBO's Real Sports with Bryant Gumbel featured a story on a Arkansas high school coach that never punts and onside kicks after every score.  Here's a synopsis:

Kevin Kelley, head football coach at Pulaski Academy in Little Rock, Ark., uses his own research to back up the mantra that his team should never punt, should always kick onside kicks after a score and should never return punts.  If this sounds too unorthodox to be effective, Kelley's theories have resulted in a 104-19 record since he became head coach at the high school, and statisticians have supported his theory on the relative value of punting.  REAL SPORTS correspondent Jon Frankel visits the 42-year-old Kelley in Little Rock as his team prepares to seek its third state title under his leadership.

http://www.hbo.com/real-sports-with-bryant-gumbel/index.html#/real-spor…

The weird thing, as you may have read, is that he doesn't have his team even make an attempt to return a punt -- they literally just run off the field as soon as the ball is away. I guess he just figures that his defense did their job in getting the other team to give his team back the ball and he doesn't want to risk a penalty, or, worse yet, a fumble.

Anyway, it was a really interesting piece. It's available for viewing on HBO On Demand until 2/26/12 if you want to check it out.

Joined: 06/30/2008

MGoPoints: 10175

michelin

February 5th, 2012 at 10:14 AM ^

I am inclined to believe that coaches do not go for it on 4th down often enough. Yet, I have always wondered if biases in the statistics used to support the “go for it on 4th down” argument may weaken that argument. The problem is that these statistics let us know what happened in the cases when one went for it on 4th down, but they do not tell us what would have happened if the teams “went for it” in the other cases. Let’s say that the teams that “go for it” usually run the ball and they do so more often if they have a strong running game, or if their opponent has a weak run defense, or if playing conditions are more favorable to the run. Thus, if one’s decision to go for it is based on certain favorable playing conditions, one cannot really know what would have happened regardless of the conditions. So, if your team does not have a strong running game or the defense is strong against the run but you start “going for it” always on 4th and one anyway, your success rate may change dramatically. For example, as detailed in the footnote below*, say your success rate previously was 95% but you went for it only 10% of the time. But say your success rate would have been only 50% on the other 90% of the times you had previously chosen not to go for it. Then, if you decide to always go for it, your overall success rate would turn out to be less than 55%--a far cry from the 95% success rate you thought you would get. Footnote *Suppose V+ denotes the fact that you did go for it—and can verify (V) its success(S). We can define that success however you like---eg ideally, your success would not just depend on how often you get a 1st down or score but according to the average number of points scored by you minus those scored by your opponent afterwards (calculating that difference, of course, presents many other difficulties, which could further affect the actual success of “going for it”; however, for simplicity’s sake, let’s suppose that we can measure such “success” ). A problem then arises if, however we choose to define success, it increases depending on certain conditions eg if strong running game (R+), weak run defense (D+), weather favorable to run (W)+, etc The problem is that the measured P(S/G,V+) is not always= P(S/G), the chance of success you would have in ANY conditions—eg if you always went for it on 4th and one. For example, suppose P(S/G,V+)= 95/100 Suppose V- denotes the cases when you did not go for it eg if the conditions are (R-, D-, W-) Success in such cases is unknown but suppose it would have been P(S/G,V-)=50/100 So if the fraction of times going for it Pr(G)=p(V+) = p P(S/G)= p(P(S/G,V+)+(1-p)(P(S/G,V-) Say, for example, p=.1 then P(S/G)=.1(95)+.9(.5)=.545, which is far less than the apparent success rate=.95

Joined: 09/22/2009

MGoPoints: 2541

joeyb

February 5th, 2012 at 10:58 AM ^

Your point is valid to an extent. On-side kicking every time might not become beneficial once teams start expecting and practicing the on-side kick. IIRC, the numbers point to a 28% success rate for expected on-side kicks and 62% for unexpected. However, if you have 12 different onside kick formations, it should tilt the number slightly back in your favor as the other team doesn't know what to expect.

With going for it on 4th down, yes, the numbers might change, but going for it on 4th and 1 from your own 9 yard line is probably still a huge advantage regardless of which style of offense that you run as long as you are competent. What might change is that instead of 4th and 8 being the cut off for going for it between the 40s, maybe 4th and 6 becomes the cutoff once there is more data to work with. Although, a lot of times, statisticians will use 3rd down as a measure of success for 4th down because 3rd down is a do-or-die situation most of the time like what 4th down would become. If you watch the clip on HBO about it, he points out that he only has a 50% chance of converting the 4th down at their 9 yard line, but the odds are still significantly in their favor if they go for it as opposed to punting.

Joined: 10/12/2008

MGoPoints: 16845

michelin

February 5th, 2012 at 11:30 AM ^

The statistics I use are just hypothetical and illustrate one potential problem. As you indicate, many other biases may occur; and possibly, one bias in the "the 4th down success" estimates may cancel another bias and the "go for it on 4th down" argument may be as valid as ever. However, you cannot know that; and systematic overoptimism does commonly occur when we can verify only what happens in a partial sample (eg only when teams actually do "go for it on 4th down"). In some cases, one can actually correct for such biases; so one could possibly attain a more definitive answer to the question: "should I go for it on 4th down) (for those interested in exploring the question further, I suggest doing a literature search on "verification biases").

Joined: 09/22/2009

MGoPoints: 2541

turd ferguson

February 5th, 2012 at 12:09 PM ^

I agree that bias is very likely, but I have a hard time getting a sense of the magnitude and direction.

The point that you're making is certainly one source of bias.  Another possibility, this one going in the other direction, is that teams tend to go for it on 4th down when they're losing and desperate (i.e. being outplayed and maybe panicky).  So maybe the teams that go for it tend to be worse than their opponents - and tend to do so in really compromised situations.

Another possibility is that teams are using their very best 4th down plays when they go for it now.  However, if you multiply the frequency with with they go for it by a very large X, they'll probably have to dip into a less ideal group of plays.

There are about a dozen different stories that I could imagine.  I've seen some people use third down conversion rates, which I think is better but still far from perfect (since teams so often play for a field goal or field position on 3rd down, figure that they have 4th down if they need it, etc.).

Long story short, I agree that teams probably don't go for it nearly enough and that the estimates for how often to go for it are probably really unreliable, but I'm not sure how to answer that question accurately.

Joined: 12/09/2009

MGoPoints: 26493

Mgrad92

December 16th, 2015 at 11:52 AM ^

This went up today:

The High School Football Coach Who Never Punts

Pulaski Academy football coach Kevin Kelley never punts. He always kicks on-side. His teams play … weirdly. Now the Arkansas squad is going back to Dallas to face its rival, Highland Park, a team that has won 84 games in a row at home and outplayed Pulaski Academy last year. Will Kelley’s data-driven plays lead to the win? The latest film in FiveThirtyEight and ESPN Films’ short-documentary series Collectors, “Undefeated,” directed by Jamie Schutz, follows Pulaski as the team adds a new wrinkle to its playbook and tries to end Highland Park’s streak.

Joined: 01/13/2011

MGoPoints: 514