Thursday, September 01, 2005

The Indefensible

I once proposed this defense, partly seriously, but it hasn't been implemented yet. It's just a variation on the offsides trap.

A possible new Defense for the O

Premises, all true to varying degrees
  • The O has trouble getting turnovers against any reasonably skilled team.
  • Even reasonably skilled teams have weak links
  • All passes have some risk to them, and “good” passes have an acceptable rate of return for a given risk.
  • A plan that is good is theory is bad in practice if the players you have can not possibly execute the plan.

THE DEFENSIVE IDEA: Encourage the turnover-prone throwers on the other team to make risky throws. For example, let a bad hucker on the other team get a somewhat open long cutter so he can overthrow him.

RATIONALE: We all have (or should have) developed a set of risk/reward curves for our throws, with a different curve for each opponent or level of difficulty for opponent or game situation. Another way to look at this is that you’ll consider many things when considering whether to throw a particular huck: how good the other team is, how windy it is, how good the deep defender is, how good the receiver is, how open the player is, where on the field you are, what the score of the game is, etc. A situation can change just a little, but the risk changes from “worthwhile” to “unacceptable”. So, if your offense is so good that you will certainly score if you make short passes, then it makes no sense to try a huck that has a 90% chance of completion (disregarding for the moment such factors as fatigue, long-term strategy, practice, confidence, etc.).

I have to say that occasionally, the best chance that the O has at getting a turnover is if the other team misses on a huck. We are banking that the other team fails to realize that the risk/reward curve has changed.

Defensively, we need to make only small adjustments to the faceguard/last back defense.
“Last back” has to delay on the switch so that the pass gets thrown. When you switch/poach, you can either try to prevent the pass by switching as soon as you see it (and so that your teammate can also switch), or you can try for the block by switching as late as possible so the thrower doesn’t realize you’re going until it’s too late. This D would favor waiting. We get burned now by switching too soon and they get a free underneath cut.
We need to have some way to let the team know whether the player with the disc is a good hucker or a bad one. The two ways we could set this up would be either “if players A or B get the disc, we want them to huck” or “if anyone except X or Y gets the disc, we want them to huck.” We would want some guys on Furious to huck, but not Cruickshank. Perhaps the player covering A or B yells “FORCING!” when it is apparent the player is going to catch it (but not when he’s about to catch it, as that’s unsportsmanlike conduct).
We probably will want to force a few players back to the disc. So, perhaps A and B get forced back to the disc, while the others are faceguarded and forced away.

Another variation would be to encourage certain players to try difficult break mark passes.


Jon said...

Hey Jim,

It might be time to turn off anonymous comments.

In response to the post, I wish some teams would try this defense against us. I'd get more practice hucking.

dix said...

Jim, in another blog I read on in order to post a comment you have to enter a few letters or numbers that show up as a graphic which any human can read but machines can't. Not sure if that's offered as a service on blogger but it would get rid of the spam.

Dennis said...

Way ahead of you Jim...sort of.
The strategy is extraordinarily effective in mid-level ultimate where, often an enormous disparity separates throwers (and the choices they make).
T-cup is sort of based on the same principle. The high percentage throwers are covered -- and, also, have to work through the psuedo-zone-- while the lower percentage throwers are the ones forced to work the disc.
Another thing I often do -- just on "normal man-to-man": I line up against the turn-over machine and blatantly poach. I almost play short-deep (or three in clam). The only thing I don't do is let my player get deep.
It's amazin' what a flow-killer that is for the other team. The three hot-players could have been working it easily for a while. Now, suddenly their flow is stopped because of the extra-poacher just standing there, and their machine is going to get it two or three times a point. I then move over to him and mark hard. Any time there's a break in action, I tell the three closest defenders to front like effin' crazy. It's enormously effective.
(PS. I usually play this way without detailing the strategy to my teammates. So I'm often yelled at when the machine gets the wide open dump or swing pass -- and there I am jogging over to him. But unlike 1988, say, I'm granted more leeway now. They whisper, "Hey, he knows Jim Parinella so maybe he knows what he's doing."
When you think about it, of course, it is absolutely insane, on a percentage level, to play in-your-pants deny-the-disc D on a machine. All you do in that case is turn the game into a 6-on-6. Worse, you turn it into a higher quality 6 on 6, of course, because the weak link of the opposing offense has been removed -- thanks to you. The opposing offense should buy you drinks.
"Hey, our machine's throwing at 75% and looks off dumps. But you stopped him from getting it and forced us to work it with our best players. What are you havin'? Jack and Diet?"
This strategy is significantly less effective in higher-level ultimate, of course,-- but something like you suggest could work.

Justinsucksatultimate said...
This comment has been removed by a blog administrator.
Justinsucksatultimate said...


What is most amazing about the man-to-man strategy you describe is that I explain what you are going to do before nearly every point and people still don't get it. I mean, I actually say before the pull:

"Dennis is going to hide in the back of the stack and wait for someone to throw long. That means someone will always be open, usually the person that just threw the disc. If you are not being covered, stay open for a short pass instead of running into the middle of the stack."

However, Portsmouth ultimate just seems unable to adjust for two reasons. The first is that I REALLY suck at coaching people. It doesn't matter that the advice I give is good. People don't listen.

The second is Portsmouth is still a small pond in terms of ultimate. The defense you are describing requires that there be a player who is too stupid to know what to do when they are left uncovered. Unfortunately, such players are in good supply on sunday pickup and summer league. But I am not sure that defense would work consistently against a more intelligent ultimate team.

Justinsucksatultimate said...

on the main topic, playing defense seems a little like counter espionage. create the appearance of a weakness to get the offense to play to your strength, or make a move that you anticipate.

as a soccer goalie, i almost never played in the middle of the goal on a penalty kick. i would deliberately be a step off-center in order to entice the kicker to kick to the more open side. the key was balance -- move too far off center and you can't recover in time, or worse, the kicker realizes what you are trying to do.

that same concept is really at the heart of zone defenses. i suspect that people who make bad mistakes against zone D have little understanding of how statistics are applied to behavior. interesting... i wonder if you could find a correlation there?