Friday, November 03, 2006

Chain at Nationals

There will undoubtedly be those who consider Chain’s success at nationals a fluke. Those who think Chain’s run was legitimate might still say that a single tournament is too small of a sample size to draw any worthwhile conclusions. I think both of these criticisms are at least potentially reasonable – let that serve as a caveat for what follows.

I think Chain’s success at 2006 Nationals was primarily the result of two strategic principles. The first is that relentless (perhaps reckless) aggression is a powerful strategy. The second is that the point at which the replacement value of a fresh tier two player exceeds the value of a fatigued tier one is somewhat closer to 60% of points played (for the tier one player) rather than 50% of points played.

In an earlier post, I defended the value of relentless aggression. There my argument was basically two-fold. First, I argued that The Rule demands that all things being equal, a player should throw a longer pass rather than a shorter pass. 2) I also made the more controversial claim the advantageous implications of the huck extend beyond the calculation of giving your team the highest percentage chance of scoring this goal. This claim is based on the fact that once a team believes you are crazy enough to huck at any point they will over commit to protecting the deep cut opening up the underneath. The nice thing about this second point is that it can extend to later games in a tournament/season without your team even having to bear the burden of the additonal turnovers early in a game. I think FG and Sockeye both currently benefit from (2) due to their reputation as mindless huckers.

In any event, we (the blogosphere, although not sure I’m still a member) have had this discussion and I don’t mean to rehash old territory. Another facet of our relentless aggression was an insistence on forcing the disc upfield whenever possible. A dump has to be considerably higher percentage than a 20 yard gainer if the dump is to be justified by The Rule. Honestly, I still don’t think HnH is optimal in perfect conditions, but this year’s nationals was far from perfect conditions. I think that anyone arguing that a possession style offense is optimal in this year’s conditions (with the exception of Sunday) is either bad at math or simply not being honest with themselves.

The second strategic principle can basically be restated as “you should play your studs more than you are now.” Jim posted on this subject with some fictional numbers a while ago, I’m too lazy to find the link now, sorry. At early tournaments this year we had three considerable comebacks at the end of games when we just put in our top 7 for several points in a row. This led us to make the conscious decision to play our studs more at nationals. It sounds pretty obvious, but the current dominant strategy of splitting O/D has led a lot of top players to play only about 50 percent of the points. In most situtaions, your studs could play more than 50 percent of the points without fatigue impacting their play to the extent that it would make sense to put in your next tier of player.


parinella said...


Welcome back, and congratulations. I didn't see any of your games at all, so can't comment about whether you were pure HnH or just SLG (Sensible Long Game).

Efficiency curves is the post you were thinking of.

Regarding "relentless aggression", it has to be "successful relentless aggression", or else the D won't overcommit and will just allow you to throw it away deep repeatedly.

I am very curious about the studs theory and would be interested in exploring it more.

richard whitcomb said...

As far as the studs theory goes ...
What was your impression from Nationals of how much (many?) legs you had left after 3 games Thursday, 2 Friday and 2 Saturday?
I saw almost all of your Quarters from 11 - 10 on, and only a few points of your semi, but from what I saw of the semi, some of your 'studs' looked tired.
Seems like Chain didn't have any blowouts so there was not much of a chance for anyone to rest.
I always played O and never really got tired, but I was never good enough on D to improve on the 7 we could put out there. As a 'two way' player, how much COULD you play if necessary?


aj said...


My sense is that our views are not that far apart. It seems that we both agree that The Rule should govern decision making on the field. We just have different intuitions about what The Rule requires more often in the current environment. In the abscence of reliable stats to the contrary...i'm gonna stick with the "chicks dig the long ball" strategy.


Personally, I think I can play well for something in the neighborhood for 55-59% of the points over the course of a day. I think that is probably less than most players as I tend to cut more/less efficiently than a lot of players. I was in considerably worse shape this year than in past years due to some early season injuries that kept me from training like I might have liked.

My guess is that some players could effectively play in the high 60 percent range.

Given how limited practice time is, there is definitely some value gained by strictly maintaining O/D lines. It's an interesting balancing act to try to figure out how to meld the two ideas in a way to optimize efficiency.


parinella said...

Max % is limited by both time between points and the number of points played in a day.

You could reduce the effect of the first by using timeouts between points (or by stalling between points) and by adopting other clock-management strategies (playing studs on first and last points of the half, having a non-stud zone team that can give you a long point). And if you're going to manage it down to the nearest 5% (about a point a game), you need to consider that a hard point might be as taxing as several easy points. Besides the obvious long point = hard point, you need to consider that bunching points together makes the later points harder. A hell point when it's your third point in a row might make you worthless for the next 20 minutes, whereas a hell point on your first point might just require a single point off to recover.

For the second, I guess you need to use your studs 50% or less in the less competitive games and 60-75% in the more competitive games.

otoh, the more competitive games are going to have more points.

aj said...

Another thought is hiding/protecting your studs by either giving them favorable match-ups or playing a zone and putting them in the back. As we've discussed before, players tend to be more defined by their offensive skills - the more O players you can hide among the grunts to run around on D the better.