Monday, May 28, 2018

Down By Two

One of my favorite activities to do at a practice is something called Game Flow.  I won't go into the whole thing here, but it ends with a game of Down By Two.

You'd be surprised how many games come down to one team being down by two goals near the end of the game.  A fantastic example of this is the UNC/Oregon Men's College Semifinal that happened yesterday.  The conceit of Down By Two is obvious: one team is down by two (12-10 or 13-11) in a game hard to 15.  There are two levers to play with: who is down, who is pulling.  The pulling really matters.  Being down two and pulling is a harder game to play.

So why play this game?  I said that many games come down to this, but many games come down to 4-3, so why don't we practice that?  The idea is that a late game differential is emotionally different than a mid-game differential.  You have to think and behave differently when you are down by two late in order to win the game.  Prior momentum is irrelevant, as teams can quickly develop or kill momentum with a single goal at this stage.  Everything feels amplified, and that is something teams need to practice.  Pulling with a two point lead feels different, as the game is about emotionally maintaining balance and confidence as you close out a lead.  Receiving down two is a little different because you know you are going to have to get a break eventually to win this game.  I could talk about this for a while, but the point is that the mental states are different and worth practicing.

Watching the UNC/Oregon game reminded me of that.  Often I feel like I make up things and try them with little feedback on how they go.  Sometimes players tell me things, but I don't really trust them to give me honest opinions in the moment.  Once a player commented that Nethercutt gets his teams to play 4v6 often.  It was incredibly gratifying because I have no idea if that game is actually beneficial, but at least one person smarter than me thinks it is worth doing, so that is something.  So seeing the score 13-11 and watching Oregon lose reminded me that Down By Two is a pretty important game and that skill is valuable.  Now to figure out if my crazy red zone is really worth it.

Saturday, February 10, 2018

Defining Space and How We Access It

As I get older I feel like my thoughts then to be broader.  While my youth was spent dissecting the minutia of how to play better handler defense by slightly changing the position of my body, now I spend more time thinking about broader strategy elements like how to structure and effective poach.  I think part of this comes (at least for me) with a reduction in ability.  Who am I to say how to best position your body to stop an up-line cut?  I'm old and have creaky joints so I can barely stop anyone (and more importantly have to fall back on tricks that younger players might not have to).  I guess this doesn't have to do with the specific topic at hand, and there is probably more to get into in this topic.  But the point is that I've spent a good bit of time thinking about space in ultimate and I wonder if I can codify some terms and ideas regarding the offensive use of space in this game.

First, let's talk about roles in defining space.  For the most part we are going to be talking about active space as an area that cutters use in an attempt to get the disc.  There are lots of ways to access active space (which are covered in the cutting tree that I never finish), but all of them require a particular active space.  There are other types of space, but we aren't going to be too concerned about those in this discussion.

Offensive structures are the first attempt to define space (active space being their primary focus).  A vertical stack splits the field in two vertical active spaces, with one (the open side) being the most active and the other being used.  It then finds success by switching which active space is being used as quickly as possible.  More on that later.  Horizontal stacks split the field into a near and a far space, both of which are accessible from the start.  However, as this offense (or any spread offense) continues the spaces shift and it becomes a larger, central active space (closer to the original position of a split stack).  Side stacks choose one lane and maximize it as much as possible.  The point is that offenses try to define active space with their structure.

Defenses understandingly try to limit defensive space with their positioning, marking and poaching schemes.  Let's use a vertical stack as a structure to talk about.  In a vertical stack the offense is try attempting to create two vertical lanes ~17 yards wide and space at the back.  The mark attempts to stop one of those lanes by "forcing" a direction, and the defenders likely play a yard or two off their person, narrowing the lane to maybe ~12 yards.  Maybe they bracket the last person to reduce the size of the deep active space.  In the case of a side stack, maybe they throw a person into the active lane to poach, reducing the active space.  Not to mention that once a cutter is in motion the defending player then attempts to reduce the space for you to throw the disc based on their proximity and speed relative to the offender.  The point is that defenses attempt to reduce space through schemes and athleticism.

Now let's get to the real point of the article: how do teams try to access space.  It is different for every offensive structure, and beyond that it is different for every team.  There are may ways to be successful at accessing space (some might call this "getting open" but since the theme is space I'm going to call it "accessing space).  Likewise there are many ways to fail at it.  But I think in general we can break down all of these methods into three main categories.  You can access space through cutting prowess, through throwing prowess through shifting the location of the disc and by shifting the landscape of the field through player movement.  The last is the one I am most interested in, but we need to explain the others as a reference.

First, cutting prowess.  This is how many offenses, and especially most side stacks, operate to use space.  Put a great cutter out there, let them juke and confuse their defender and eventually explode open.  It helps if you can offer two readily available active spaces (or partition your space into two sections like under/deep) so the cutter can threaten one thing and take the other.  But in the end, it is about the ability of your cutter to get open.

Second, throwing prowess.  This was on display (I think, I should watch it again) during the 2017 USAU Mixed National Finals where Amp was consistently using short breaks to the front of the stack to open up the offense.  The idea is that a person with the disc is also critical in defining the active space.  Your horizontal offense might attempt to open up initial deep strikes, but if your center handler can't throw it deep then you haven't successfully opened that space.  This happens near the end zone, where many offenses (looking at you CFS) will just throw a pass to open space and have the cutter run on.  This also happens when Brodie is trapped on the flick side of a vertical stack and just throws a hammer to the break (and undefended) lane.  We are pretty comfortable with the idea that certain throwers change the spatial environment on the field through breaks, hucks and just gutsy throws.

Third, shifting the location of the disc.  Let's call this what it is, swinging the disc.  The idea is that vertical space gets more congested the longer a disc is in a third of the field, which inherently means that other space is opening up.  If we aren't all throwing Brodie's hammer to the breakside, then moving the disc laterally is another way to access what was previously blocked space.  That isn't the only way this can happen.  In a stiff headwind your horizontal stack might not attack the deep space as effectively as in a neutral wind.  But running your resets upfield (almost like and up-line cut) allows your handlers to shorten the distance for the huck and then opening up that space.  We talk about moving the disc to change landscape often.  "High side" is probably the thing that I yell most from the sideline (thanks Nancy Sun and Alex Snyder).

The fourth one is the one that I don't think we spend enough time thinking about.  We think about cutter movement as accessing space and clearing space, but often those are still the active spaces in the general structure of the offense.  There are instances that are maybe not thought enough about, where the offense opens up new space because of native player movement.  Many of these instances are involved in plays.  A sweep in a horizontal stack is designed to vacate space on one half of the field while a player from the other half enters that space.  A split (or "red sea" as we called it in high school) in a vertical stack takes the front/back two cutters out to either side so the third person cuts in/deep.  The very premise of the split stack I run is based on how this concept works on a mixed team.

The concept, to be more concrete, is that defenses understand the spaces you are trying to open up through your structure.  They position to limit your cutting ability.  They mark to reduce your ability to throw to spaces they don't want the disc.  All of those things can happen to limit your space, and we are good at them.  But all of those things require an understanding of the active space.  But by moving cutters you can open up space that wasn't there before, and therefore the defense might not have been ready to defend.  As a result cutters, who might understand this shifting space, can easily get open there in a new way.

Let's take a specific example that isn't from a set play.  The game of adjustment and counter-adjustment in the side stack is fun to play since it has been a dominant offense for the past 5-8 years.  One thing that started to happen is that when the lane is crowded the disc will swing and the stack will sweep to the other side.  This takes the space that was previously occupied by the stack and turns it into the active space.  Let's now think about the person defending the back to the stack.  That person is clearly defending deep because there is a lot of clutter in front of them and defenders that could help if their person cuts under.  But why would they?  That space isn't even open because all of the offenders are there.  But when they move that space opens up and that offender in the back of the stack is open by player movement rather than cutting or throwing (ok, technically there was also disc movement and the counter to this is easy because you just keep your poach towards the top of the side stack in place to cap the cutter and then you are done . . . but there are more examples where it is all player movement but this one happens natively rather than in a play).  This method (similar to disc  movement or throwing prowess) allows a person to get open without being a great cutter, but also is hard to defend.  That is especially the case when this happens in the middle of offensive flow rather than a set play.

So what does long diatribe mean aside from I need to go back, edit it and add pictures to further this idea?  It is worthwhile to think about how your team accesses space and whether or not it makes sense for your personnel and team ethics.  It is worth knowing that some modes of accessing space require different skill sets.  It is important know ways to use all four methods for accessing space in different situations for different offensive structures.  Not all four are easy at all times, and some structures really limit different modes of getting open.  I'm growing increasingly partial to the fourth because it works regardless of your cutters and throwers ability.  The flip is that it requires people to be really aware of the field.  But everything has its drawbacks.

I'll try to clean this up at some point, but I had it on my mind and wanted to get it out before I forgot.

Friday, February 02, 2018

Further review on PT totals between Revolver, Ring of Fire and USX

Jacob's comment in the last post started this, and after spending some time trying to figure it out I decided that I can't add pictures to a blogger comment.  So I started a new post.

The gist of the conversation was that I would have liked a more even PT distribution between players during the finals (although I understand why it wasn't flat and commend Tallis Boyd on doing a great job with a difficult task).  Jacob asked if that was all that uncommon, so I look at particular games that he suggested.  Here are the graphs, where the y-axis is the number of points played.  For reference, the Revolver-Ring game was a slow bleed where Revolver won 13-10.  The USX final was a closer affair until the very end, when we won 13-11.  Graphs (Revolver, Ring, USX, respectively).

From the looks of it, Jacob is right that the USX point distribution isn't that different from the others. A better question would be if it was different from our semi-final against Canada.  One thing that might be worth paying more attention to is where the ramp starts.  For USX our minimum point were 3 out of 24 (1/8 of the points).  For Revolver the minimum was 3 out of 23 (stupid prime numbers), while the Ring one was 1 out of 23 (ignoring the 0's which might have been injured players).  It would seem that the starting position of the ramp would likely determine how steep it could be.

One other thing that is really worth pointing out, is that a more flat distribution isn't "better."  In this case it was something that we might aspire to, but that doesn't mean we are playing a better form of ultimate or have a higher chance of winning.

Saturday, January 27, 2018

A few more numbers for USX '18

I ran a few more numbers for the finals again Japan and I thought I would share them

Point distribution among players wasn't as flat as I would like.  Two players only played 3 points, while four players played 10+.  The ramp between those two levels was pretty even, but I wish playing time was a little more even.  It is a difficult ask as those sort of things tend to fall away in the finals and during a tournament when there are many ofter things going on.

Possessions per goal for each team were decent, but not excellent.  USX needed (on average) 2.77 possessions to score and Japan needed 3.18.  It is always true that the team with the lower number during a game wins, so there isn't much behind a comparison of these numbers.  Comparing it to previous college and club teams it falls within the range of expectation for college teams (where playing 2 is elite college and 3+ isn't unheard of) and I wouldn't expect to be at the front of that number since there isn't a ton of time to gel and the competition is better.  It is pretty bad for a club team, although I haven't looked at those numbers for the mixed division in particular.

The number that is most interesting to me is percentage of possessions ending in an unforced error (defense doesn't touch the disc).  USX had consistent numbers in the first and second half around 43-45%, which again isn't that bad (but could really be a lot better).  The impressive thing for me is that in the 2nd half Japan had a number of 21.4%  That is very elite.  That means 1 out of every 5 possessions ends in an unforced error (which includes hucks that go too far).  If a team I am coaching has that number then we are either winning, both teams are playing lights out, or they are getting blocks.  The latter was the case this time, as we were able to slightly grow a lead despite the Japanese playing relatively error free ultimate.

Wednesday, January 24, 2018

Gender Contact Ratio (3rd cycle)

This past January I was fortunate enough to have another wonderful experience coaching the USA U24 Mixed team (USX).  We won a gold medal, like the previous two years, but this cycle was very different from the past two.

This was the first iteration with no founding players.  Last cycle we had returners that had shaped the teams existence and this was the first official passing of a torch to new leaders for whom this was already an "established" thing.  This time we took a smaller roster, slightly eschewing the "more people means more fun" mentality for the possibility of a tighter knit team.  This cycle also was captive by the times and far more interested in gender equity than teams in the past.  All of these things are probably worth talking about at some point, but for now we will divert from the last one to look at the same thing I have looked at the past two, gender contact ratio.

The same caveat applies as before: there isn't a "perfect" number that represents the most "equitable" style of play.  We can't make hard conclusions about the involvement of both genders from this number, but it does help us check our perception of reality and talk about broad topics.

The context of the game was similar to last cycle.  It was windy, but not too oppressive.  Japan used their women more effectively (at least in my opinion) than Australia did in 2015.  Our team was more conscientious about using our women, and we had more female handlers than we had ever had.  In 2013 we could argue having 3 female handlers, with really only Sophie being a true center handler in the bunch.  This group decided early on that we had really good female center handlers and threw the first pass to them without any instruction from us.  This time we had Hardy, KJ, T-Lo and Anna all acting as handlers and all taking turns getting centered to.  Japan also utilized a zone that required a lot of patty-cake from our handlers (or maybe it was the handlers that we had at the time) because we weren't going to swing around their 3-3-1 as easily.  This patty-cake often featured male handlers getting lots of touches.

The numbers for this game were pretty surprising.  I felt like we were more equitable with our distribution, not as a mantra but rather because we were opening up good space and willing to throw passes.  I often talk about how my anecdotal measure of success for a mixed team is how willing they are to throw a 20-30 yard under to a woman who is well guarded.  It shows a level of trust, openness of space, and also a recognition that small spaces are ok sometimes.  I felt like this team leaned in to that principle well and was often throwing big gainers to their women.  Those women didn't turn around and boost it as much as I would like, but at the same time they weren't always resetting the disc either, so I think it is a structural problem in how we were running the offense.

To the numbers.  In this game there were a total of 417 touches (as compared to 252 and 180 in the past two cycles . . . patty-cake).  Of those 417, 157 were by women.  This breaks down to 37.6% of the touches or a contact ratio of 1.66 times as many touches for a man as compared to a woman.  If we look at the past two cycles we had numbers of 3.00 (2015) and 2.53 (2013).  I know I said that this number doesn't well correlate with better play, but I can't help but feel that we threw to women more often and this number supports it.  No doubt part of it is getting people in places to shine, but we were consistently throwing difficult passes to small windows and hitting women.  We also did a better job of clearing out space downfield so that strings of female passes could occur before a male defender and offender crowded the picture.

It is worth noting that these numbers include a few points against that Japanese zone where we had a high number of touches and since much of that was between three men (Brett, John and Matt) it skews the numbers a bit.  Women were still key players in those points, but not in the patty-cake that was most of the touches and rather in the key outlet and through passes that actually moved the disc.  One such possession had 27 touches by women as compared to 52 touches by men.  Another was 18/47, respectively.  I thought about taking these points out, but they are real points that affected touches, so I felt they should stay in.

On thing that is worth looking at is a comparison of the final to the semi-final.  After the semi against Canada there was a clear feeling by members of the team that we "hadn't used our women well."  While that is amorphous and certainly isn't completely embodied by this ratio, it was the feeling by the players and we were a little surprised.  Looking back at the film the first time it felt like there were plenty of places where only men were touching the disc, but that it was largely a structural issue (cutters too far away and reset defenders able to get too close) than a broader failure to throw to women.  Watching the film again looking at contact ratio, the semifinal was less balanced than the final (showing that we cleaned things up) but not terribly so.  The ratio was 2.24 touch for men for each by a woman.  That is better than both other finals, although it is markedly worse than our final this cycle.

So what does this mean?  I feel like I got this job in part because I was able to well describe a way that I thought we could best showcase mixed ultimate, not by ignoring gender and treating everyone the same, nor by over leveraging particular advantages (see Bad Larry mid2000s who ran 4 women so their men had more room to homey) but by saying that if we put people in the right places and throw to them in a system that creates space for everyone, we can all get better.  It is a "total is greater than the sum of the pieces" approach.  I feel like we did that in the past, and while watching these two games again more critically I do feel that we made many refinement errors (didn't run our redzone particularly well, suffered from handler creep, etc.) we did a better job of getting everyone space at times and throwing passes to everyone.  That, combined with a better balance of genders in each position, and a focus on gender equity, is likely part of the reason we ended up getting more people involved.

Again, that doesn't mean that the road to victory is having a ratio of 1, or that this is the only way to win these games.  But as a person who hopes to live up to values of inclusion, and recognizes that there are times that I fail, looking at the Gender Contact Ratio for this cycle makes me think that we did a pretty good job.