[WARNING: This post contains discussion of statistics by a non-statistician (last stats class I had was 44 years ago, a political science methods class I wasn’t much interested in even then). The numbers discussed are for entertainment purposes only and have not been analyzed for validity or reliability or any other measure of accuracy or usefulness. Not only that, while anyone is free to run amok playing with the potential implications or flaws in these numbers, I have no interest in doing so myself and will not participate in any discussion of them.]
Last week, while I was looking through old blog posts for something completely unrelated, I came across this Playing with Numbers post, which I’d forgotten about. It was a serendipitous find, since I’d been thinking about entry numbers since I saw the unexpected jump in entries for last month’s NAC in Richmond.
In that March 2011 post, I played with what I then called fencers-per-strip but is more accurately individual entries-per-strip (BC lives and dies by the number of entries, not by the number of fencers), and which I am now naming the Agony Indicator, after Hipmunk’s agony view, where you can see flight itineraries sorted by how awful the times and layovers are.
Back then, I looked at that season’s NACs to see whether the Agony Indicator correlated at all with what I called “perceived difficulty of tournament”—not the quality of the fencing but the difficulty of the schedule (flighted events, delays, strip & ref shortages, late nights). That year, as was often true with the event combinations then, the December and March NACs were significantly easier tournaments to run than the others, and October, November, and January (then mostly either Junior/Cadet or Junior/Division I combinations) were noticeably more difficult to fit into the available days. Here are the indicators I came up with for those NACs:
October NAC: 44.8
November NAC: 51.98
December NAC: 19.7
January NAC: 45.47
March NAC: 34.63
There appeared to be some correlation—the size of the indicator varied directly with the perceived difficulty of the tournament, and the more difficult tournaments had indicators above 40.
Since I had been looking at those Richmond numbers a couple of weeks ago, I wondered how recent tournaments fared with my Agony Indicator. With a bit of research (I still have most of my Masin spreadsheets from tournaments I’ve chaired over the past few years) and a calculator, I took a look:
October 2012 NAC: 49.898
November 2012 NAC: 49.184
JOs 2013: 43.102
October 2014 NAC: 47.143
November 2014 NAC: 55.265
January 2015 NAC: 33.816
JOs 2015: 47.551
Those line up in the same range, reasonably correlating with the challenges. (One caveat: some of these tournaments also had team events, which are not included in the Agony Indicator calculation.)
So what about this season? Just for fun, I ran the numbers for the Gutenberg SYC Alia and I ran last month, and it came out to a not-at-all-agonizing 20.759. That’s the only regional tournament I’ve calculated the Agony Indicator for, though, so I’ve no idea whether my indicator correlates across a range of smaller regional tournaments.
The Richmond NAC last month came out to 57.984, the highest I’ve yet seen, which seems to be in line with what I’ve heard about how things went—multiple flighted events and very late Saturday and Sunday nights. The NAC coming up this month in Kansas City looks to be slightly better, at 57.164, and that number should improve slightly with the no-shows there. Baltimore in December, using the entry numbers as of last week, is in between at 52.776.
What about Summer Nationals? The Agony Indicator for SN calculates out to a completely different range, since the numbers are often three times those of a typical NAC, with only half again as many strips. The 2010 SN in Atlanta was 108.84. The more recent SNs worked out like this:
2013 SN Columbus: 107.369
2014 SN Columbus: 137.4
2015 SN San Jose: 114.344
Note: The entry numbers for 2013 and 2015 were almost identical, but 2015 had 4 fewer strips. Columbus 2014 had as many strips as the year before but nearly 2,000 more entries.
What’s the takeaway? If you happen to come upon entry numbers and the number of strips planned before an event, you can calculate the entry/strip ratio for yourself. If the Agony Indicator works out to more than 40 for a NAC or for SN, more than, say, 105, you can expect multiple flighted events and late nights.
Got any spurious statistics of your own?
[This is my last post about (nonfiction) fencing for the foreseeable future, though I will blog on other topics occasionally while I work on finally finishing what I hope will be a publishable draft of my SN murder mystery.]