Thursday, July 07, 2005

QoTW - Kick or Receive

Assuming wind is not a factor – what do you prefer to do when you win the flip – kick or receive? Traditionally, I’ve been a fan of taking the ball. I’ve always thought of winning the flip and playing D as something college teams did. However, last year at Nationals (no wind) I think every team that won the flip against me selected to play D first. Anybody have any thoughts on this?

17 comments:

jason sweeney said...

For me, I'd rather go into the second half with the advantage rather than the first.

This, of course, is determined by the level of wind. If it's not too windy, I would rather start the second half of the game receiving. If it is windy, I'd rather start the second half playing downwind.

Only if it's really windy, and it's one of those games that might not get to a second half, I will try to get the downwind start, whether it's receiving or pulling.

Of course, I don't always get what I want.

heacox said...

This has been a point of discussion in our open summer league. Ever since Stu told me to take the wind if there is wind, and otherwise take the disc, I have been choosing to receive* when I win the flip.

A lot of guys on my team think we should start on D if we win the flip, because it's easy to get a turn on the first point of a game (if this is applicable to high-level play, you could say that the offense has not yet established a rhythm), and you have the added bonus of receiving to begin the second half.

However, I'm not sure how much this matters if there are a few turnovers per point and wind is not a factor. Therefore I would choose to take the disc and hope that more possessions** would lead to more goals over time. Another factor is that I don't like having to play offense out of the half-time break, although as I think about this, I realize it is for the same reasons that other people choose to start the game on D--it's easier to get a turn out of a break (at least at the level I've played most of my games).

*Or start with the wind.

**Or more downwind points.

parinella said...

Mathematically, it's pretty straightforward that you should take the disc (or wind), since the game might be capped and whoever starts with the ad is more likely to have the ad at the end. If there is no "win by 1" cap, it doesn't matter mathematically.

You have to invoke psychological arguments (while ignoring others) to make the other case. Heacox suggests that the O doesn't have a rhythm on the first point. That might be true, but wouldn't that apply to the first point they play regardless of whether it's the first or fourth point of the game? You would have to think that the O develops its rhythm by watching the D play a point or two.

You could also argue, as Alex sometimes does now, that it would just mentally kill the team if the O gets broken to start the game.

An interesting way to play the game would be to do away with the pull, but you let one team pick the yardline that the O starts at, and the other team gets to choose whether to play O or D. You could also use this to tweak the O/D balance without changing the other rules. If it's a swill-fest, let the O start at mid-field every point. If it's a great O day, start the disc 10 yards into the endzone. You could also do sudden-death this way. The ref or TD chooses a direction, the flip loser picks a yardline, the flip winner picks O or D.

Anonymous said...

i have never understood the thinking behind electing to start on D. my thinking is that you always want to give your team the best chance to score first, whether that is choosing to receive or to go downwind.

last year in the open division of nationals with no wind, there were many games decided by one point. so it's quite possible that teams could trade out to the end, and the team that starts on offense will lead the whole way. strong upwind/downwind games are also likely to be one point victories.

more importantly if you need to win by two, the team that starts on O just needs to get one break (or be up one break) at any point in the game.

the team that starts on D can still win by only being up one break, but they would have to get it in the first half. however, if the teams are on serve at half, and that team starts the second half down 8-7, they must get two breaks to win by two.

i think that's right. hope that makes sense.

as to the idea about it being easier to get a block on the first point, in my experience that has not been the case in high-level play (my sl team sucks, so i can't speak to effective sl strategies). my feeling is that d-teams need to work themselves into a game more than offenses due to needing time to figure out matchups and the other teams strategies and just getting warmed/fired up.

b-lo

Alex de Frondeville said...

Clearly if there is a significant amount of wind, the choice is to take the downwind direction. The other team will almost invariably choose to pull (if they receive, the wind might die by half-time, giving a large advantage to the team that started downwind). Back in the '90s, DoG invariably wanted to start on offense because of the mathematical arguments that Jim espouses, and because we felt our offense was unstoppable. Secondary consideration for me was that if you assume your D will get scored on to start the game, AND the O gets broken, you would prefer not to start in a 2-0 hole. Much better to use the beginning of the game to get the O on track. If they get broken to start, you're only down 1-0, not 2-0. And the best part was, everyone else wanted to pull at that time, so we almost always ended up with both choices, receive AND side, because if they won and chose pull, we picked side, and if we won, we picked side because they would choose pull.

Now that DoG's O isn't quite as stellar, and we have dropped back into the pack, my thinking has become a little more defensive on the topic.

Of the four possibilities, which are the best/worst?
1 Starting on D and taking half
2 Starting on D and not taking half
3 Starting on O and taking half
4 Starting on O and not taking half

I would argue that 4) is monumentally worse than the other 3 choices, giving the opposing team the chance to score two straight points on O. 1) is by far the best for the same reasons. 2) and 3) are basically neutral. So given the cost of 4 versus the benefit of 1, for a team whose offense is equal to its defense, I would argue that starting on D is better.

aj said...

Jim:

I think I remember reading something you wrote where you made the same argument (although, I can’t find it at the moment, so potentially I’m just making it up), and I agree with you. For this reason, I was surprised when Alex elected to play D in our pre-quarters game against you guys. Like you say – from a mathematical standpoint it just makes sense to play O, but in a game like that one it does make some sense psychologically to start on D. Your team has more experience and is much less likely to be nervous in a pre-quarters game than Chain. By starting on D you put all the pressure on your opponent to score the first goal and perhaps you can take advantage of their early game nerves. I think it would have made a lot less sense to do that if you were playing a more experienced team who is less likely to be rattled.

On a totally unrelated note…here’s something else you guys did strategically that impressed me – It’s generally the policy of our O that if we have a long point and don’t score or give up 2 breaks in a row, we put in our second O team. When we put in our second O team you guys ran a full-field clam on them. It was the only point you ran it all game. We eventually scored (after a turnover), but I thought that was a great idea – our second O team is made up of primarily D players who pretty much never have a chance to see junk. It seems like throwing junk at them is obvious, but no one else did that to them all season. Anyway, not related, but I was thinking about it, so I thought I’d mention it.

Anonymous said...

alex wrote:
Of the four possibilities, which are the best/worst?
1 Starting on D and taking half
2 Starting on D and not taking half
3 Starting on O and taking half
4 Starting on O and not taking half

I would argue that 4) is monumentally worse than the other 3 choices, giving the opposing team the chance to score two straight points on O. 1) is by far the best for the same reasons. 2) and 3) are basically neutral. So given the cost of 4 versus the benefit of 1, for a team whose offense is equal to its defense, I would argue that starting on D is better.


i think it's not quite fair to say 2 and 3 are neutral because it depends on whether the team starting on offense is ahead at half just from trading or if they got a break.

if Teams X starts on offense, let's consider three scenarios at half. a) teams are on serve: score's 8-7. b) Team X is up one break: score's 8-6. c) Team Y is up one break: score 7-8.

in scenario a, Team X can trade to cap or get a break to win by two. Team Y must get one break to win at cap or get two breaks to win by two.

in scenarios b an c, the teams are in the same situation to win the game. the team that's up can trade in the second half and win by two. the losing team must get two breaks just to win at cap, three breaks to win by two.

starting on offense and giving up a first half break and then getting scored on to start the second half to make it 7-9 is not much different than being on defense and giving up a first half break and having a halftime score of 8-6. atleast mathematically speaking.

in my mind, the argument to choose to receive is based on scenario-a. the argument to choose to pull is that first half breaks are more important than second half breaks, and choosing to pull gives you one extra shot on D in the first half.

that's probably more confusing than my first post.

b-lo

Rahil said...

For the record, in a perfect game of ultimate(no turnovers), the only way you can lose is if you don't recieve the opening pull.

Tarr said...

I always go with side if I can perceive any discernible advantage. Otherwise, receive. I've never noticed any special weakness in O on the first point of a game, at least not above the middle-tier college level.

Alex's "4 scenarios" are something of a numerical mirage. Yes, half is when a team that starts on D takes the two point advantage if they got a first half break. No such dramatic moment exists if you start on O, since you have the two point advantage as soon as you score on D. Unless you think there is something especially psychologically crushing about giving up two straight O scores, it doesn't matter.

An interesting way to play the game would be to do away with the pull, but you let one team pick the yardline that the O starts at, and the other team gets to choose whether to play O or D. You could also use this to tweak the O/D balance without changing the other rules. If it's a swill-fest, let the O start at mid-field every point. If it's a great O day, start the disc 10 yards into the endzone. You could also do sudden-death this way. The ref or TD chooses a direction, the flip loser picks a yardline, the flip winner picks O or D.

If you lost the flip in the, uh, quarters of nationals, on a calm day, where would you set the yard line? 1 yard in front of the end line? Just curious.

heacox said...

I didn't think Jim was serious about this suggestion, but then yesterday he started a new topic on his blog about it. It strikes me much more like a game theory exercise than a viable alternative to the pull. However, I still can't make complete sense of it, even after reading Jim's other post about it numerous times?

In a sudden-death situation, if wind is not a factor, doesn't the loser of the flip have the advantage? If the winner picks O you could stick him on the back cone, and if he picks D then you can start in the red zone (even on the goal line)?

Am I missing something? Do both teams bid in secret (which is what I originally imagined)?

parinella said...

In my little game, the flip loser decides where on the field the disc is to be put into play, without knowing whether they will be O or D. They will pick the place where they feel it's 50/50 whether the O or D will score. Then the flip winner gets to make their own decision on whether to start on O or on D.

It's a variation on the problem of how to cut a piece of cake so that both parties think it's fair. One person cuts the cake, and the other gets to choose the better piece.

In Tarr's what-if, I'd probably put the disc on the line on the forehand side of the field about 5 yards from the back of the endzone. If the other team lost the flip and put the disc on the back cone, I'd choose defense.

You could make the decision more strategic by requiring the teams to put out their players before the flip.

Tarr said...

Interesting. Do you say that because you think it's a break-even point for all teams, or do you think DoG has a disproportionate advantage in this case? That is, do you think some teams would say "screw this, we're putting it up to (MG/Chase/Damien/Brandon/AJ/Danny/etc.) and if he can't come down with it we take our chances on D"? Whereas DoG would actually feel comfortable working it around as long as they have enough space for one reset pass?

I think getting rid of the pull in general is inelegant and unnecessary, but I actually like this idea for sudden death.

parinella said...

It's not a break-even point for all teams, but I just assumed a typical Nationals quarterfinalist on a calm day. Receiving a pull, they would expect to score about 2/3 to 3/4 of the points (3-5 breaks per game).

I felt that the forehand side would be easier to get out of, and that DoG might have a relative advantage compared to other teams, but otherwise this estimate was not DoG-specific.

One of these days I'm going to make a set of charts of field location vs scoring probability, upwind and downwind, with and without a mark (or in flow and at a stoppage). Someday, actual data will exist that can verify it.

Flo said...

There are a lot of fascinating aspects about this new idea of sudden death.It is clearly more fair than a flip for the last pull. Just a little theoretical comment:
Would you really give the winner of the extra flip the choice of o or d? is it not better to pick the spot and the playing direction?

here is my reasoning:
if both teams are exactly equal, there is surely a spot on the field where scoring from is 50/50. From the opponents goal line, o has the advantage, from the back corner of its own endzone going up wind, o is most likely in a disadtvantage (and if this is not enough, you can throw in extra handicaps like a stall count at 8). Assuming the chances change continuously with field position (a pretty safe assumption), there must be a spot with 50/50. In fact, there are multiple spots on the field where it is 50/50. The difficulty of finding this spot aside, the team choosing the spot thus can always make it a 50/50 affair.

But what about the assumption that the teams are exactly equal? This is probably not true. If your team has big difficulties getting out of a trap, you would most likely choose a spot further to the middle of the field further away from the scoring endzone, same with preferences of flick or backhand side. Or upwind/downwind.Since teams strengths are different, there should be a spot on the field where you have a >50% scoring chance from, no matter if you are on o or d.

parinella said...

Ok, so two amendments to the sudden death:
1. Let the flip winner either pick the spot or "defer" and pick O/D.
2. Instead of a flip, give the choice to the team that would have been receiving.

And I also like the idea of using a game of skill (whoever can throw it closer to a far-away cone) instead of the flip.

Edward Lee said...

A little bit off topic, but is there some reason the official UPA rules refuse to use a coin flip to determine choice of pull and side?

If disc flipping must be done, it should be a sequence of N flips, where N is large. The team that is calling it specifies whether they think the number of discs landing face-up will be odd or even.

Jon said...

If disc flipping must be done, it should be a sequence of N flips, where N is large.

Not necessarily. You can use the same procedure that you would to make a coin of unknown bias "fair":

1. Flip the disc (or coin) twice. Define the sequence up-down as up and the sequence down-up as down.
2. If the sequence down-down or up-up occurs, start over.

No matter what the bias, the flips are independent, so down-up and up-down are equally likely. If the disc is really biased, I guess you could end up doing a large number of flips, but probably you'll only need 2 or 4.