2014 SN Diary: Travel Day, June 20

I’ve set this post to be published at around the time I board my first flight Friday morning. It’s rare for me flying anywhere for fencing besides Portland or Phoenix to get a flight that leaves later than 6:00 am, which means I get up at 2:30 or 3:00 to get myself to the airport in time. I learned long ago not to try to work on my way to tournaments—flights are for traversing the wormhole into the alternate universe of Tournament Time. My main goal is to stay awake long enough to board my flight, and then sleep. If I can’t sleep, I’ll watch the landscape out the window, and if it’s cloudy and I can’t sleep, I’ll resort to a crime novel or science fiction. The point is to avoid anything like reality, because that transition into the unreality of Summer Nationals is an important factor in how I survive it.

Over the last couple of weeks, in addition to working on the BC staff schedule and other SN prep (plus reviewing material for tomorrow’s board meeting), I’ve been creating the beginnings of posts for every day of SN. I think I’ve finally figured out a way to blog SN as it happens that I will be able to keep up with through the ever-increasing cognitive dysfunction.

The pre-written sections are a look at what each day’s competition is expected to look like from my point of view, and includes what I think of as the “big grid” in the strip planning spreadsheet we use. So this post will provide some basics about how that spreadsheet works and how we use it.

Here’s the big overview of a typical competition day:

whole sheet

The big grid on the left is the overall view of strip usage by the half hour through the whole day. The righthand two-thirds are columns for each event. What we do is enter the data for each event in the appropriate columns, and from that the formulas generate the grid view.

Here’s a look at the information on a single event:

eventThe category, sex, weapon, and number of entries are entered in the blue boxes at the top. On the left side of the column, we can enter the time. Traditionally, we enter the time as a half-hour later than close of registration, to allow for the withdrawal of no-shows, resolution of conflicts, and assignment of referees. With the new setup, using the referee auto-assign, we can now start as soon as 5 minutes after close, but we have chosen not to adjust the spreadsheet for that yet, since slight overestimating the time required works in our favor.

Below the number of entries, the worksheet gives us the range for the number of pools possible. For national events, we almost always choose the smallest available number because the rules require maximizing pools at 7 members whenever possible. (Veterans are the one category allowed to have pools of 5 even when larger pools are possible, but we don’t like to do that unless we’re really pressed for time, since the Vets, like most fencers, always like having more bouts.) Also in this section, we can set the promotion rate and whether the pools should be flighted, and the worksheet shows how many strips that will use and how long the pools can be expected to take.

The next section shows the DEs and how many fencers are promoted to that round. After indicating whether there is repechage (always and forever at national events, I hope that box shows “N” because a “Y” means 2-3 hours longer for that event to occupy 8-16 strips than without repechage. Again, we set the number of DE strips and the worksheet shows the time required.

If there is repechage, the next section down shows the strip usage from the round of 32 or 16, as applicable.

Finally, the last section shows the timing for the round of 8. The formulas have not yet been updated to show the effect of replay, so we sometimes bump up the “seeding minutes” between rounds to allow for the extra time replay takes.

What we do when we create the schedules for national events is enter the data for all the events for each day, and then we start playing around with the start times, the number of DE strips, and whether pools need to be flighted in order to make everything fit into the overall day, as shown in the big grid. Sometimes it’s a pretty easy process. Other times, there are long stretches of time filled with exclamations like “Ack!” and “Yikes!” and “Crap, that didn’t help at all!” and “Oops, so not going there!” as we try various options.

Note, by the way, the number in the top upper left of this column of information. That’ll tell you which column in the grid shows this event.

 

Looking at the whole-day picture, we can find additional useful information:

grid

In the main part of the grid, each column represents a single event, with the weapon indicated at the top of the column, and each cell shows how many strips are used during that particular half hour. On this day, there are 5 individual events and 2 team events (those are further to the right, in the “T” columns). So looking at our event from above in column #2, we can see that the pool round will use 22 strips from 8:00 am to 12:30 pm and the DE round will use 16 strips (4 complete pods) from 12:30 pm to about 4:00 pm. The actual time will vary according to how fast the fencers fence and when the complete round of 16 can move over to replay. So the timing shown here for the final rounds will be only an approximation of the eventual reality.

Further to the right are three columns under “Weapon,” which show how many referees are needed in each weapon during any given half hour, and at the bottom, the maximum number required simultaneously.

The last column on the right shows the number of strips in the hall, and further down, ominous red numbers that indicate strip deficits. As a general rule, red numbers here are bad, but some are worse than others. A -1 or -2 when a final of 8 is finishing isn’t really a problem, because by then that 8 is probably in its semis or gold medal bout and not using all 4 of its allotment. And the -17 on this day is what I think of as a squishy negative—many of the 32 DE strips used by the #1 event will already be free as the last few finish up, so not too many of the pools of the afternoon events will need to be delayed. Unless, of course, that #1 event has a major injury or equipment problem that delays the last pool coming in, so that the DEs start late and then finish late, and suddenly we’ve got a competition day that’ll run later than projected.

Down at the bottom left, we can change the times allotted for each kind of bout and match. They’ve been where they are now for the past couple of years. It used to be that foil durations were somewhere between epee and saber, but that changed with the foil timings. It also used to be that saber generally ran faster than projected, and foil and epee more slowly, to the point that if the point weapons finished within an hour after their projected end times, I considered it in good time. Since we’ve started using replay, saber tends to run at or slightly later than projected, and epee often runs as much as an hour ahead. Foil these days consistently runs more slowly than both other weapons, though we haven’t quantified it well enough to change the timings yet.

So that should give you the basics of how to interpret the “how hard was today supposed to be?” grids I’ll be posting throughout SN.

Layout & Strips, Part 6: Repechage, Plus Some SN Considerations

I’ve been fond of repechage ever since I figured out how it works. As a moderately geeky, detail-oriented BC person, I find it an interesting intellectual exercise to calculate all the backside table reseeds and swaps, and it’s always fun to try to get it done before the computer operator hands me the printed version. But much as I personally like the double-elimination format, I feel strongly that it no longer belongs at national competitions, but should be relegated to camps and fun tournaments along with fencing-16-to-all-places (which can also be thought of as quadruple elimination).

Here’s why I (and most other BC people I know) are not happy about the return of repechage to national events. Let’s take the hypothetical tournament we’ve talked about in the last two posts and add repechage to the ME event. The pool round would be the same, of course, with the 23 pools of 7 and 6 pools of 6. The DEs start out the same: 158 of the original 197 fencers would be promoted to the DEs. Again, that’s an incomplete table of 256, on 8 pages in XSeed.

But around 1:00 pm, when the tableau has been fenced down to the 32, the double elimination kicks in, and two extra repechage tables are added to the process. If the round of 32 is fenced on 8 strips, as is typical, the event will finish around 6:00–6:30; if fenced on 16 strips, it will end a bit sooner at 5:30–6:00. Repechage adds between 2 and 3 hours to the length of the event.

But that’s not its only effect. Don’t forget that WF event, expecting to be able to go wide on 24 strips for its DEs. Depending on where the ME is when the WF pools are done, those 24 strips might not be available for the WF DEs. Because the ME uses more strips for more hours, the options for other events are reduced. For this relatively simple hypothetical tournament day, the consequences are not too serious. For a more complicated day, such as those we will face at the January NAC, where not only will we be adding repechage back to all the epee events but we will be adding six Cadet events to the usual Junior and Division I events, those consequences could well cause delays that cascade through the whole schedule. Each day could end up even longer than just the two to three hours added by the repechage itself.

Think about that for a bit: 12 large events over 4 days become 18 large events over 4 days. Instead of 3 events each day, we’ll probably have 4 events each on Friday and Monday, and 5 events each on Saturday and Sunday. And spread thorughout that assortment will be 6 large epee events with repechage making them 2 to 3 hours longer than they otherwise would be. Those days will have SN hours—with fewer events, they won’t be such complicated days as at SN, but they will be SN numbers and SN hours. It will be an interesting tournament.

There’s another aspect of repechage events worth talking about, and that’s the question of when the inevitable waiting around should occur. At every repechage event, there’s always at least one group of referees which runs their quadrant significantly faster than the rest and inevitably gets frustrated at not being able to keep going as far into the DE table as they want to.

One of the irritating quirks of XSeed is that it cannot print out the tableau or bout slips for the round of 32 and beyond until all 32 competitors are known. Sometimes the BC uses this to our advantage—if we need to reduce an event from 16 or more strips down to 8, we’ll stop at the 32 to move everybody to 2 contiguous pods. But this drives those hyper-efficient referees batty—they want to keep going and get all those frontside bouts fenced down to their 4. Sometimes, if we’re not moving to different strips, we’ll let them keep going, but to do so, we have to print out blank bout slips and start transferring bout results—both front- and backside—to a handwritten paper tableau. Eventually, once the whole 32 is in, XSeed will print the round of 32 tableau, but until then, keeping track of both the front- and backside bouts is a tricky and often frantic process for even the most experienced BC staff.

And the thing is, when those rapid referees finish off their quadrant 30 or 40 minutes ahead of everybody else, they get to sit around waiting for everyone else to catch up before the rep tables and the round of 8 can be fenced. Why waiting around then is so much preferred to waiting around earlier may be one of those referee mysteries I’m doomed never to comprehend.

If this glimpse into BC thinking about strip management has you thinking that strip planning doesn’t seem all that complicated, keep in mind that my imaginary tournament day was a relatively simple one. Some NACs—and JOs—are nearly as straightforward, but others are more challenging, such as the March and April NACs, with multiple age-level Veteran or Youth events, even though most of the individual events are not huge.

The monster of all challenging tournaments is, of course, Summer Nationals. With 6 to 14 events each day and daily entries ranging from 350 to nearly 800, figuring out which events should go where on  66 strips is a decidedly nontrivial jigsaw puzzle not found anywhere else in the fencing world.

Layout & Strips, Part 5: DEs

At the end of the last post, we’d reached 11:00 am on our imaginary tournament day. All the pools from our ME and MS events have finished, and the promoted fencers are waiting for their DEs to be posted. While the scoresheets are printed and the referees assigned for those events, the 112 WF fencers whose event closed at 10:30 are reporting to their strips for their pool round.

As you may recall, the MS DEs will be fenced on 8 strips, in 2 adjacent pods of 4 at one end of the hall. Because the WF event will be the last to finish, I’m going to put their 16 pools of 7 on the middle 4 pods. Assuming that the BC stage is roughly in the middle of the room, putting the WF in the middle, too, will mean they won’t need to move to a different area for their DEs. The ME DEs will go on the remaining 4 pods at the other end of the hall. For the next few hours, every strip in the hall will be used for competition.

How do we decide how many strips to use for DEs? That’s determined mainly by the total number of strips available and the number and size of events to be fenced. In this case, we have 345 fencers in our 3 events: 158 promoted in the ME, 75 promoted in the MS, and 112 fencing in the 16 WF pools (from which 90 fencers will eventually be promoted to their DEs). That means we’re looking at a not-very-full table of 256 for the ME (8 pages of tableau in XSeed), a not-very-full table of 128 for the MS (4 pages in XSeed), and a half-full table of 128 for the WF.

Normally, we aim for the DEs to take somewhere between 90 and 120 minutes to fence down to 8 finalists, and we use our planning spreadsheet to see how many strips it takes (8, 12, or 16, usually) to achieve that. Sometimes we are forced into smaller numbers than we’d like because of a crowded schedule or a shortage of referees (fairly common in saber, where we might have 12 strips available but only 8 referees).

In this hypothetical, the ME is projected to take about 2-1/2 hours to fence down to the 8. Typically, the point weapons run about 30–45 minutes longer than our spreadsheet projects, mainly because we haven’t bothered to adjust the spreadsheet bout durations to align with the actual current durations. If the ME reaches the 8 by 2:30 or so, we’ll consider it on time. Since this is a relatively simple day, where no other groups will need any more strips, the only time pressure is the desire of the fencers and officials to eat their dinners at a moderately reasonable hour.

Since the ME is on an 8-page table, each pod will have 2 pages of the tableau. Depending on the preferences of the head referee and the number of referees available, we might send this out as four groups of referees and 2 pages per pod, or as 8 smaller groups with a single page for every pair of strips. The single-page-per-pair grouping would make the paperwork easier to track, but the whole-pod grouping would minimize delays caused by one or two strips being slower than others, since the larger groups can pick up the slack more easily to compensate. (The larger groups also make it easier to avoid referee conflicts, since there are more referees available to use.)

Similarly, the MS might be assigned with half the tableau (2 pages for each half) to each of the two pods of 4 strips, or with each one-page quadrant assigned to a pair of strips. A saber DE of this size would typically take 60–75 minutes to fence down to the 8.

Once each event reaches the 8, it could continue to fence straight through to the final, or we could pause to move the round of 8 to a single pod. On a relatively leisurely day like this one, we’d more likely take the pause and assign the 8 to the single pod closest to the BC platform, and then hold the gold medal bout on the designated finals strip, if there is one. If not, we’d choose a strip in the pod used for the round of 8 that allows the most space for spectators (and preferably a good view of the fencing and the score box for the BC).

The round of 8 typically takes 90–120 minutes for the point weapons and 45–60 minutes for saber. (Not only are saber bouts shorter, but it’s not uncommon for saber fencers—especially the younger ones—to opt not to take the whole ten minutes allowed between DE bouts.)

If the ME finishes as projected and there are enough foil referees, we’d be able to take the WF DEs wide and put their DEs out on 24 strips, 3 in each of 8 pods. With 90 fencers in a 4-page tableau, that would be a half-page for each of the 8 groups. That might save 20–30 minutes over the 16 strips originally planned for the WF DEs, not crucial for this day but perhaps highly desirable on a long crowded Summer Nationals day.

Under this hypothetical schedule, the MS would likely finish between 2:00 and 2:30, the ME around 3:30–4:00, and the WF around 5:00–5:30.

Next time: Repechage and some thoughts on SN complications.

Layout & Strips, Part 4: Pools

From here on, I’m going to be talking mainly about national tournaments, which are different from any other category. Even large sectional championships or regional tournaments like Duel in the Desert or Pomme de Terre have a completely different feel from NACs and SNs. Mostly it’s a matter of scale: the quantitative differences—more days, more entries, more square footage—are big enough to make real qualitative differences. While the general principles are the same for local and regional tournaments, their application varies with the specific venues, events, and entry fields.

So to pools. At national tournaments, the pools for most event categories are maximized to 7. This means that pools always consist of 6 or 7 fencers, except where the number of entries makes that impossible, such as a field of 15, where 3 pools of 5 are allowed, or a field of 9, which is usually run as a single pool on 2 strips. A few categories, such as Veterans, allow pools of 5, 6, or 7, in which case the bout committee may opt to use all pools of 5 and 6 instead of 6 and 7, but more on that later. In any case, the pools should be as evenly divided as possible, and there should never be more than two sizes of pools (except occasionally where an injury or expulsion knocks someone out of one of the smaller pools—always a frustrating occurrence).

Let’s say for our theoretical tournament that we have 40 strips, laid out in 10 pods of 4 strips each, with men’s epee, with 197 entries, and men’s saber, with 93 entries, closing at 8:00 am, and a women’s foil event with 112 entries closing at 10:30 am. We’ll keep it simple—80% promotion rate, no repechage. (I’m saving repechage for a later post.)

Our 197 ME entries give us 23 pools of 7 and 6 pools of 6. We could just put these 23 pools in order on strips 1 through 23, but that would mean that all the pools of 6 would be grouped together. Once they finished, there would still be those other 23 pools working through their bouts, and only a few of them would have one of the now-empty strips from the 6s close enough to use to double-strip.

We could assign the ME pools three to a pod, and put saber pools on the fourth strip in each pod. Since the saber pool will only take about half as long as the epee pools, that fourth strip could be used for epee bouts once the saber is out of the way. But wait, there are only 10 pods, which means that there wouldn’t be enough strips that way for the 14 MS pools (9 of 7 and 5 of 6) that we’ve got—we still need to get 4 more saber pools out. So the MS will have to be flighted.

But what if both flights of saber finish before the epee pools are done? That would mean we’d have to assign the DE bouts for the saber to strips scattered all over the room instead of to adjacent strips, or we’d need to hold the start of the MS DEs until enough adjacent epee pools have finished to give us the 2 pods we’ll need for the saber. Neither of those options is likely to make the saber people happy, especially since they’re already flighted as it is.

Ideally, what we like to do as much as we can, is put pools out on the same strips that each event will use for their DEs. So we’ll put out 7 or 8 pools of the MS on 2 adjacent pods, probably at one end of the room, and the MS will simply stay there on those 8 strips for their DEs, too. With even or almost-even flights like this, it’s not usually worth trying to place 6s next to 7s so the 7s can double-strip when the 6s are done, and since saber runs so relatively quickly, it wouldn’t save all that much time anyway.

(While I like to run events efficiently, I’ve never seen much point to rushing everybody just for the sake of going as fast as we can. On a relatively simple day like this theoretical day, there’s no real need to rush the saber—rushing it would not make the overall day any shorter and there is no other event waiting to use the same strips.)

With the saber pools out on 2 pods, that leaves 8 pods for the ME pools. We could arrange those pools so that 6 of the 8 pods each had one pool of 6, so that once the 6s were done, the extra strip in each pod could be used to help speed up the remaining 3 pools of 7.  We’d then have 3 pools of 7  and 1 empty strip on one of the last 2 pods, and 2 pools of 7 and 2 empty strips on the other. If we’re lucky, the slowest epee pools will occur in the pods with empty strips, but that hardly ever happens. More likely, 2 or 3 pools at opposite ends of the building will be the last to finish, as much as half an hour or 45 minutes after the first pools finished.

In any case, with this arrangement, the MS will be able to start their DEs as soon as they finish their pools—they won’t need to wait for any of the epee to be out of the way. And once all the ME pools are done, their DEs will be put out on 16 strips, and the remaining 16 strips will be turned over to those 16 pools of 7 in WF who’ve been waiting for the ME to finally be out of their way so they can get started.

But we’ll save that for next time.