Is Adrian Peterson a Liability? A running back's fumble rate distribution | Hacker News Reader

Is Adrian Peterson a Liability? A running back's fumble rate distribution

advancednflstats.com

34 pointsprofquail24 comments

Apologies for the tangent, but are there any HN users who are interested in sports analytics?

Absolutely. I drool a bit every time I hear Darryl Morey talk about the MIT/Sloan Sports Conference.

You mean Dork Elvis?

I do :)

Yes, definitely, although I'm personally more interested in baseball than football. I originally saw this article via The Book Blog here: http://www.insidethebook.com/ee/index.php/site/comments/how_...

Yep, but baseball is more my thing. I guess I am just a fan of modeling systems (and heck - if it helps my fantasy league selections - cool).

Yeah - I play in fantast football league - it was a way of helping family keep in touch when my grandparents passed away. My dad, brothers, uncles and cousins are pretty in to football. I didn't really know much at all (probably couldn't really name more than 2-3 current players at the time), but the stats and numbers and everything keep me pretty interested. I get interested in the math behind things and trying to predict future performance (which is very difficult but fun to analyze nonetheless).

I won the league last year and was destroying people this year until the playoffs when my cousin had 3 players put out some of their best numbers of the season (combined they scored more than most teams of 8 usually do) and stop me in the playoffs. I am in the top 3 highest scoring teams every year. I find it alot of fun actually.

Good question -- I am. I created a poll here to see if other are as well: http://news.ycombinator.com/item?id=1101191

Very much so. I've been going through The Book[1] trying to work out some of the calculations using clojure as an exercise in learning both topics. Baseball has an incredibly rich data and deep knowledge ecosystem that is fascinating (and sometimes profitable, just ask Nate Silver).

At my day job[2] we don't do much in the way of advanced analytics, but it's on the wishlist. I'm veering into shameless plug territory now, but if you are interested in this stuff, please check out our Saber blog[3]. Also if you are a smart dev, come work with us :)

[1] http://www.insidethebook.com/ee/

[2] http://www.sbnation.com/

[3] http://www.beyondtheboxscore.com/

Hi Dusty do you have an email address?

Hiya. I just filled in some info in my profile, you can find it there.

Definitely, but I also prefer baseball over football (but football over basketball).

Yep. I also created my own analysis algorithms, which I planned to turn into a small subscription website a while back until I found out that the data services I needed were too expensive for me to turn a profit.

If you're interested in sports analytics, feel free to send me an email and we'll chat.

I've noticed a couple of teams--like the mid-00's Packers and Seahawks--where it didn't seem to actually matter who their running back was. Even their third and fourth string backs would perform well. This is when I started to suspect that a team's blocking--the offensive line and fullback--made more of a difference to the running game than the running back. Not that you don't need a good running back, but that the ability to run fast enough and go through the holes created by the blockers are dime-a-dozen type talents and it's not difficult to find someone to fit in that role.

If this is true, than probably the single most important thing that actually hinges on the running back himself is fumble rate. (On the other hand, Adrian Peterson seems like one of the rare running backs whose actual talent extends beyond, say, an Ahman Green or Shaun Alexander.)

This will be one of the questions that will always follow Barry Sanders. Can you imagine him behind, say, Emmitt Smith's line? Or would hit have mattered, given Sanders style of running (unbelievable field vision, lateral movements)?

This is one of the reasons for Sanders' early retirement.

The Broncos under Shanahan always had strong run blockers and created a bevy of successful rushers.

This past season, a near-requirement for success (a playoff berth) was having multiple starter-quality rushers, underscoring the importance of the offensive line.

Related to this specific article, I wonder how statistically significant these figures are, because the sample size (number of touches) is comparatively small. This has a lot to do with the reason baseball defensive metrics aren't as good as batting metrics: a player gets a lot more chances at the plate than he does in the field, and they're much easier to evaluate.

So the conclusion was that based on the analysis he's not a liability...for the Vikings. But as a pretty serious fantasy football owner, I wonder if — given the fact that Peterson was the consensus #1 overall pick last season — he was a liability to his fantasy owners. You pretty much have to play your number 1 pick (and a homerun hitter like Peterson) every week, but were there less glamorous, middle-tier running back choices that ended up giving their owners a better chance of winning each week (assuming negative points for fumbles lost in the scoring system)?

Yes, I do want to be the Billy Beane of fantasy football.

I owned Peterson, and although he didn't live up to the #1 spot, he did live up to the first round. He was usually the highest scoring player on my team even with 3 fumbles in a game.

Another issue with Peterson is that his yards per attempt have steadily declined: 2007 - 5.6 ypa (2nd among backs) 2008 - 4.8 ypa (5th among backs) 2009 - 4.4 ypa (24th among backs)

In the playoffs this year he really declined (3.6 per attempt)

Source: ESPN ( To qualify, a player must have at least 6.25 attempts per team's games played)

I don't think this is necessarily something to put on Peterson. Over the years, the Vikings have added more offensive weapons, notably a better receiving core and aquarterback. With Brett Favre there this year, the Vikings had a passing threat that was lacking with Tavaris Jackson, so Peterson wasn't relied on as much to run the offense (This is similar in San Diego with LT - his role is reduced now that Philip Rivers has become a better qb). I also think that defenses are more prepared for him than in previous years, when he was first breaking out on the scene.

I think a two-dimensional attack should, if anything, increase his efficiency since defenses can't cheat and load the box with eight guys every down.

The drop-off from year 1 to year 2 can largely be attributed to defenses having had a year to figure out how to stop him. His first year he had some huge games simply because defenses weren't prepared for a rookie to come in and dominate like he did.

Year 2 to year 3 is a little misleading because he got a new quarterback and a new offense, resulting in his rushing numbers being diluted a bit. Note that if you add in his 10.1 yard receiving average, his yards per touch are actually up from last year.

Reply on news.ycombinator.com