Tuesday, July 05, 2005

learning the game

A loyal reader writes:
who have been the major influences in your ultimate career? from whom did you learn? were they players you emulated? opponents that forced you to get better? teammates who actively taught and coached you?
or is your style all your own, sprouting from your head like athena from zeus?

The short answer is, no one taught me nuthin', though I have learned from people, and that's why I am how I am, for better or worse.

One of the frisbee accomplishments that I'm most proud of is that I feel that I helped to create an open, almost academic atmosphere of knowledge sharing. Although some teams still guard their "ideas" and playbooks, you'll now find a lot of high-level players freely (or for $19.95 from Human Kinetics!) passing down whatever they know about the game to whoever wants to listen.
You kids today don't know how lucky you have it. I can't remember a single piece of direct instruction or advice from the first 10 years of my career. (Of course, there have been times in my life when you couldn't tell me anything.) There was always some general team-level strategy, but it never translated into "you need to do this or that, which you can do by implementing the following."
The two things that did help me were playing with and against some top players, and hanging out with Dennis and Alex (or maybe that's four things, depending on how you count it).
Playing with top players allows you to see repeatedly the things they do and the decisions they make, in a way that getting schooled 15-2 twice a year at tournaments does not. It also helps you get over your fear of them, and lets you think, "Hey, I can do that." You're also able to separate the truly top players from those who are simply good complements who would not stand out in weaker surroundings. One such standout that sticks in my mind as I write this is Jeremy Seeger, who was fundamentally perfect.
Playing and hanging out with the Tea Party in the years 1989-1994 was also important, as these were really my formative years in my understanding of the game. We talked about frisbee, we figured things out, we were able to try out the ideas on a good team. Even if we were just rehashing things others already knew (but were keeping secret!), it felt like we were on the cutting edge of technology, and it was exciting.
And I guess there is a third (or fifth) thing that made me me, and that is the writing. Having to describe what it was that I did forced me to think about it more, and to learn from it. You really need feedback in order to improve, and if it doesn't come from a teammate or opponent, you have to do it yourself.
Having said all that, I think some of the drills we've been using the last few years (but not that one you all know I hate) are good at teaching.


dix said...

Jim, you left out Dan Powers teaching you the clam and Duncan teaching you to value the disc.

parinella said...

That's funny. But you do touch on something else important in my development, the stats. It forced me to look at every single turnover I made and reevaluate my throw choice on those passes (although in retrospect I should have also looked at the throws I didn't make), and it made me realize that I should get more involved in the offense.

dix said...

You know, looking at the game statistically seems so obvious now, it is easy to underestimate the resistance to any form of quantification back then.

parinella said...

But you know, some of them may have inadvertently expressed some genuine concerns in their fear. There is no question that what we measured was of value. But was what we measured as important to determining value as what we didn't measure? For instance, we never did a plus/minus stat. Suppose Captain Intangibles always drew the other team's top defender plus two poachers, and as a result he rarely touched the disc but the team always scored when he was on the field. We wouldn't have captured that. Or Mr. Fill always gets open at stall 9, so _others_ benefit by not having as many throwaways. Or the defender who truly shut down his opponent and never got blocks. Or the guy who is overvalued by counting touches because he never breaks the mark or advances the disc.

I think the flat-earthers looked at these exceptions as "proof that the stats don't work", rather than taking what we had and trying to improve on it. For our part, we should probably have realized that what we had was probably only the equivalent of the triple crown stats (batting average, HRs, RBIs), and that there was a lot of value left unmeasured.

dix said...

True, much like advancing the runner in baseball or smart baserunning. We never did track the stat of 'complaints about stats' to check for inverse correlation with completion percentage. Nate leaps to mind.

parinella said...

Perhaps most of the complaints that were made were of the "Productive Out" type, but there could have been complaints of the "Moneyball" type.

We didn't measure break passes, pass distance (other than huck/non-huck), defense (other than blocks), and we did little to adjust for context (man and zone were added together, no adjustment for playing time or quality of opposition). And we added the results together in an arbitrary manner (it made sense, sure, but it was still arbitrary). Those are all valid complaints.

dix said...

Virtually everything in baseball can be quantified in some way ala Bill James. Ultimate's a lot harder. It always bugged me how you guys would pad your stats dumping against the Dartmouth zone while guys like me got no credit for clearing the fuck out of the way against the good teams.

I thought the stats were broken down into 'tough games' and other. Playing time could be somewhat derived from total chances though I'll probably tell my granchchildren I was hurt a lot that year.

The only adjustments I remember were those to make Dennis look better.

Wicks said...

"We never did track the stat of 'complaints about stats' to check for inverse correlation with completion percentage. Nate leaps to mind."

You don't mean me, do you? My completion percentage was, I imagine, very high, but I still complain to Jim about his stats all the time. I would say my high completion percentage would make me seem, to a casual stathead, more valuable than I really am (was?). What I mean is...it's not like I had a high completion percentage AND threw lots of hucks. I just threw a lot of completions (though I suppose I would get some points for breaking the mark and for the double-assist, right Jim?).

But I am pretty sure you meant a different Nate, so nevermind.

parinella said...

It's a different Nate, from Earth Atomizer days.

I've been mentally preparing a graph that tracks the "expected scoring percentage" pass-by-pass and second-by-second through a possession. It starts off at, say, 40%, drops to 35% when the pull lands in the endzone, goes up to 37% when the first uncontested pass to the goal line is thrown, rises to 37.05% when that pass is caught, drops to 35% when the Man cutter doesn't get open, goes back up to 37% when there is a swing pass, drops to 30% on a shitty throw, rises to 42% on a nice catch, etc. For each event, assign the change in % to the responsible parties, and sum it up to get a "Goal Probability Added" for each player for each point. In the above example, the Man gets a -2% for getting shut down (and his defender gets a +2%), or if it's partly the thrower's fault, give each of them -1%.

This is a long way of saying that not all passes are equal, that not all yards gained are equal, and that generally most of the credit for a pass belongs to either the thrower or the receiver. In the first pass example above, virtually every player in ultimate would be equally effective as the receiver, and most players would be effective as the thrower, so even if the pass gains 20 yards, there is little credit given for it, and what there is goes to the thrower.

dix said...

It seems like many of those are not easily objectively measurable. You used to mock the stat-criticisers by saying 'why should I use numbers and statistics when my feelings and impressions will do'. Some stats that might be more objectively measured might be saves, defined as receiving a disc released after stall eight or transitions defined as a pass completed which was released prior to stall two. Sure, a lot of bogus plays will get counted but no stats are perfect. Manny Ramirez leads the league in outfield assists. nuff said.

parinella said...

I still mock them for saying that. And I wouldn't want to be responsible for taking all the stats I outlined above, but that isn't to say that they aren't actually measurable. "Saves" would be a worthwhile, doable stat to add to the triple crown stats we kept (and we actually did have, for a brief time, "Good Plays" and "Lame Plays" which prevented/almost caused turnovers).

Re: Manny. To say that Manny has the best arm because he leads the league in assists would indeed be ludicrous. However, he has added value/prevented runs by throwing those guys out. Of course, to get his true contribution, you'd need to account for the extra bases runners have taken on him.

Wicks said...

What stats are you going to take for the women's team you are starting?

parinella said...

Did my wife hack into my google account again? Dammit.

I can think of several relevant stats for women.

sack said...

dix said...
" Jim, you left out Dan Powers teaching you the clam and Duncan teaching you to value the disc."

don't get me started...