Wednesday, January 04, 2006

More O/D crap

Over at the PuPs, I've been spewing stats about O and D value. Here are a few points I want to make about value:

  • Value comes from being better than average or replacement.
  • If the difference in a particular skill between good and average (or best and worst) doesn't translate into a substantial difference in points or turnovers or runs, then that skill isn't that valuable to the game.
  • However, it may be true that the skill is a threshold skill, where you have to achieve a certain level just to play the game. For instance, weight for an offensive lineman in the NFL is largely uncorrelated to value, but you have to be at least 270 or 300 lbs to be able to play. I guess we could also call this a fundamental skill in which simple competence gets you 95% of the value of mastery.
  • We've all been commingling two discussions about value, one being whether an individual offensive player is worth more than an individual defensive player, and the other being whether team offense is more important than team defense. The first issue, I think, is pretty clear for most sports and most players. A top offensive player is involved in a higher percentage of the plays than a top defender. A "shutdown" in ultimate just results in a few seconds wasted (which can of course help create a turnover), while a successful cut clearly increases the odds of scoring. A 20 yard reception from 40 yards out might increase your scoring efficiency from 70% to 80%, but if that cut is shut down by good defense, the efficiency might drop only from 70% to 66.5%. (Here's how I got these numbers. I assume a team is 90% from the goal line and that they lose 5 percentage points every 10 yards. At 20 yards out, they're 80%, and at 40 yards out, they're 70%. For the shutdown, what completion percentage would you expect on a dump where you turn at stall 5? Multiply that by 70% to get your new efficiency. Say it's 95%, then 70% * 95% = 66.5%. You might also need to factor in the yardage loss, pushing it down to maybe 64%.) And you can only cover one player. (Exceptions to this would be a great middle middle in the zone or a great defensive center in basketball who can influence every drive into the lane.)
  • This is almost an aside, but while there may be an expectation that the offense should score every time, the reality is that offenses score only about half the time, which is also about the rate of scoring in the NBA (fluctuations in points per game are as much a function of pace (possessions per game) as efficiency). I maintain that even for the top of the game, a large percentage of the turnovers have little to do with what is considered "hot D" but is just a matter of being close enough to remove some of the margin of error.
  • While I don't play on the D team anymore and I follow the disc when I'm on the sideline, I still see an awful lot on the field when I'm playing O.

Oh, yeah. Offense rulez!!!

11 comments:

gcooke said...

Jim,

I think comingling of the values is part of the point I am making. As you have clearly shown, if you put a D player out on the island of 1 on 1 the O has the advantage (you have also put it the context of a race in which only the O knows the finish line). However, this example, while practical from a stats point of view, places the D player in a context that might be an example of bad defense. Part of the struggle of the D is to try to prevent isolations, so, in this analysis, we are judging two players in a situation in which one is not only at a disadvantage, but maybe even in the context of bad execution.

I think we have judge a D player by his ability to function within the team defense concept.

-G

Anonymous said...

Take also into consideration that an O line can only win the game by itself if they receive first and score with 100% efficiency (never broken). With the exception of that scenario, an O line can only lose the game, whereas a D line can only win the game. O line is expected to score every single time. If they do not and the D line cannot make up for their deficiency, the team will lose. If the D line, however, can create break points with great efficiency, the team will likely win, even with a low offensive efficiency.

- Joe's Brother

gcooke said...

Isn't the O team on the field for only 1 point in the first scenario?

Am I missing something?

-G

Anonymous said...

I submit that, in forming a team, O players in general are more valuable than D.

However
Certainly you have your clear exceptions to this. Specific dominating D players (Nord and a few others today?, Jon Gwertz in his prime, Kaye Nakae before that) can be very disruptive, with the ability to take your best O player out of the game. Yes this may only be for a few seconds, but per count, making the remaining seconds to 10 more challenging. Plus your best O player is your best O player because they desire to be involved a high % of the point, magnifying the effectiveness of the dominating D player.

But lets focus on the good/effective D player. In a game that is significantly weighted in favor of the O player, assuming equal experience and intelligence, dosen't this mean that the effective D player must then be a better athlete (with perhaps more desire), and hence once on O after a turn, don't these same attributes make it easier for them to dominate their couterpart?

Semar said...

Since I know Jim is such a big soccer fan.... there's a reason that offensive players in soccer get paid higher salaries and generate bigger transfer fees than d players... though there are occasional Kaye Nakae's in soccer too.

Basketball too -- Dennis Rodman was a tremendous asset in his prime, but he didn't make the money that big scorers make.

Previous poster's comment (probably in jest?) that great o players can always be great d players is wrong. Plenty of guys who are older and become major defensive liabilities can play huge roles for the O, particular conservative Os, because of what they do with the disc in their hands. It's the nature of ultimate that a smart cutter can get open on much faster and more athletic guys. But this does not mean you want him on d at game point.

Semar said...

Oh -- one point i didn't address. Sure if the D guy has a huge speed/physical advantage on the turn, he might be able to do some damage. But if he's a turnover machine, there are ways to mitigate the damage. (I.e. give him the in cuts).

Anonymous said...

G -

That's only if the D also scores with 100% efficiency. If the D makes no contribution to scoring (i.e. O wins the game by itself), the O line is out there every other point. If the opponent received first, however, the O line still cannot win the game.

- Joe's Brother

Anonymous said...

Given that the prevalence of dominating O players is greater, statistically (since this is Jim's blog, use of this term is mandatory at least once per post)you would be best served to stack your team with as many of these players as you could find (assuming they can play together).

Truly dominating D player are few and far between. When you identify one, by all means offer to subsidize their rent and find them a girlfriend to intice them to move to your city and play with your team.

Anonymous said...

Pardon my ignorance, but who is Kaye Nakae?

parinella said...

Kaye Nakae played for the Condors in the '70s.

I feel that a good defensive showing on a great downfield offensive player would be to shut down 3 or 4 cuts in a game (out of 20 or 25 that a downfield player will make; it's rare in the DoG stats for a receiver to get more than that many receptions (handlers get a bunch of reset passes, so their totals are typically 30)), to make him work for the others, and to limit him to no more than 1 huck reception. If he gets a block on a bad throw, all the better. If indeed the defender shuts down half his cuts, that would be a huge value, but I don't see that happening very often.

Regarding salaries, some will claim that since pro sports is entertainment, fans come out to see goals and offense. While fans will say this, I think that research has shown that attendance tracks winning a lot better than it does offense, i.e., fans dig winners. So, while there may be a premium paid for offense, salary is very closely tied to expected contribution to winning, so you can use salary as a proxy for value (factoring in other confounders like free agency, salary caps, guarantees, signing bonuses, local TV contracts, metropolitan area,...).

And you know, it could be that most of the value in being a defensive player is simply the ability to avoid being toasted up and down the field, and being close enough to set up an immediate mark, to have a chance at an imperfect throw, and to call the receiver out of the back of the end zone instead of having to send it back. Real shutdowns are infrequent between comparable players, and sick blocks are rare (but important).

Anonymous said...

Keay Nakae actually played for the Condors in the 80's. Jon Gewirtz told me that when he first started playing Keay was one of the top players in the game and clearly the best defensive player. Then he hurt his knee and quit playing. JG then subsequently assumed that title. Others (TK, Doug the Bum)have said that he consistently got multiple blocks a game, and not just on weak passes but otherwise perfect throws. Heresay or fantasy at this point 20 yrs later, but someone who apparently met the criteria in your previous post of the rare exception who could shut down the majority of handler cuts.