I’m sure everyone clicked on this ready to get mad about whatever the answer is. But the joke is on all of you! This is actually going to be an article on just how difficult a question like “biggest” is to define, and how even if you have accurate statistics a question might still not have a definitive answer.
So, What Does “Biggest” Mean?
This question may sound a little familiar, and that is because I asked something very similar in What Is A Normal HEMA Rating?. That premise was “what is average”, because there is more than one way to take an average; the method should reflect what type of question you are asking and what kind of data set you have.
So first let’s look at some possible candidates for an answer to this question. And consider the pros and cons of each.
“Sean, stop being dumb and over complicating. Biggest means the most people there.”– some strawman I just made up.
I’m glad you asked, Mr. Strawman. The difficulties with calculating most people are twofold. The first issue is that while it seems like we’ve defined the question, we haven’t. And the second issue is how do we get the information once we’ve decided what we want. We’ll start with the first issue: who exactly are we counting?
Most People: Most Fighters
One metric we could use is the number of fighters at the event. To break the pattern of me saying something and then saying it is wrong, I’d say that the total number of fighters is a decent indicator of size. And, most importantly, we usually have a pretty good way of tracking this. Through the greatest data aggregator in HEMA: HEMA Ratings.
By this point pretty much every large event is going on HEMA Ratings, so it’s a pretty good indication of how many fighters there were. The numbers are, however, not completely skim-able. Because if we are going by the number of fighters we have to consider the different tournaments. If there is a longsword and a rapier tournament, there is a good chance there is overlap between them. Which is solvable by looking at the people who participated in each, and only counting uniques. A pain in the ass, but solvable.
But that is only IF all the tournaments are on HEMA ratings. Let’s use Longpont 2017 as our example. This is because:
- It made its registration list public, so we have good transparency.
- It’s pretty much the undisputed “biggest” North American tournament of pre-2022 no matter what metric you use. So I’m not going to be hurting anyone’s feelings here.
HEMA Ratings has the following to say:
If you look for the unique fighters we have 193 individuals on HEMA Ratings. But if I go over to the registration list I see a total of… 285 fighters!!!
What is going on here? Well, the thing is that not all tournaments from LP 2017 are in HEMA Ratings.
If you add these the number of entries from the non-HEMA Ratings tournaments we have a total of 266, which is way more than the 92 people we were missing. This makes sense if you think about it, as most of the people in these other tournaments were probably already fighting longsword and were counted there.
Longpoint 2017 is an extreme example, with the table showcasing every one of the common reasons a fighter might not show up on HEMA Ratings. But it’s a good cautionary tale to be sure you understand the data you are trying to use. The moral of that story is that depending on what software a tournament uses, it can be extremely difficult to figure out how many fighters were in attendance! It requires a lot of back-checking to make sure everything lines up. But, with that out of the way, here we have the 2022 North American events with over 100 unique competitors:
Most People: Most, uh, People
Another way to answer the question of most people is just literally the number of people there. While having 100 extra people show up just to take classes doesn’t make you a bigger tournament, it definitely makes you a bigger event.
The biggest problem with this approach is that those records are rarely made available. HEMA Ratings only keeps track of people who actually fought. Most tournament management software only exists to run tournaments, not whole events. And even if the software does, most events don’t bother to enter non-fighting participants.
But, even if this information was dutifully recorded it might not necessarily be easy to parse. Some HEMA events run as part of larger events, making the unique HEMA registrations hard to tease out. Combat Con has a lot of people there for HEMA, but it has a lot lot of people there overall for other things. And it also has people who are there for HEMA and other things.
Further complicating if you’re just trying to answer: “how many bodies can I expect to see in the tournament hall?”. Different events lend themselves to having different numbers of non-participants around: spectators, family, and so on. SoCal 2020 and Longpoint 2016 were close in total registration numbers. But were people more likely to show up and watch in Orange County vs driving all the way to Turf Valley? Or maybe the other way around.
The number of non-participants may seem insignificant, but one last piece of food for thought: at SoCal 2022 everyone who entered the venue had to go through Covid screening. And as such these unique circumstances meant that the organizer had a count of the exact total of individuals who entered the building. While the total event participants was ~250, there were over 500 individuals entering the facility across the 3 days of the event!
Most People: Longsword Fighters
So getting reliable figures for total attendance is difficult, and factoring in weapon crossover makes the question harder. Good thing modern HEMA only cares about longsword[everyone who doesn’t do longsword]. So what were the biggest longsword tournaments?
If you’ve got any sense of pattern recognition you’re going to assume I’m going to say “we need to clarify the question first”. But fortunately this isn’t a cop-out; one of the methods is clearly better than the rest.
The issue to clarify is what to do with longsword divisions. If you have a 50 person Tier A and a 50 person Tier B how big do you say the longsword tournament is? I think this has a clear answer, in that they should be combined. Because we are trying to answer the root question: “how many people are here fighting longsword”? If you argued that it is easier to win a 50 competitor Tier A than a 100 competitor Tier A+B you would be…. Correct! However that is an issue with Quality of Competition, which is something discussed a few sections down. This is purely for the count of attendance. (Though I will note here that if you’re realistically going to win Tier A the inclusion of Tier B doesn’t make your life much harder. You now get to have throw-away matches in your pools against much weaker fighters, rather than getting to have more meaningful fights.)
So we are counting the total number of longsword fighters. But we have to be careful, and make sure no one is double counted. There are several tournament formats where people can simultaneously be in multiple divisions, and should only be counted once. If you have a 20 entry beginners tournament, and they all also enter into the 60 person open tournament, you do not have 80 longsword fighters, you have 60. 20 who did beginners + open, and 40 who did open only. This also comes up a lot when dealing with things like pools sorting into brackets for multiple tiers.
With those clarifications out of the way, these are the North American events which had more than 60 open* longsword competitors:
In this case “Open” refers to the gender division not the skill division. Whether to lump women’s longsword into the same bin for the total longsword is a bit ill-defined. On one hand it best fits with the goal of saying “how many people are here fighting longsword”. On the other hand, people aren’t used to seeing numbers represented like this, making it very counter-intuitive to see it combined. For now I’ve elected to keep it separate. Though, of course, there are always many women fighting in the open who are part of these competitor counts.
For women’s tournaments with 15 or more competitors we see similar events, even if the order is shuffled a bit. Looking at the 2022 data on its own doesn’t give us something radically different from prior graphs, however the tournament sizes appear to be a significant increase from pre-pandemic levels. But Kari is putting together a whole follow-up to her original women’s tournament participation article, so I won’t steal her thunder by discussing it in more depth here.
What most people often mean by “biggest” is often not literally a measure of size. The Women’s C-2 500m in the 2020 Olympics had 14 crews qualified, which is less than the Canadian Championships Under-16 division had the prior year. Yet if one were to ask “What is a bigger sporting event, the Olympics, or a U16 National Championships” everyone would answer the Olympics.
So it makes sense that someone asking the question “what is the biggest” might actually be asking about prestige rather than attendance. Which could be based on a few different indicators.
Most Prestigious: By Reputation
Lawl. Good luck to anyone who wants to quantify this. Not only is it subjective, the perceptions of events are going to vary wildly depending on who you talk to. I travel a lot and talk to a lot of people. From anecdotal experience I can tell you that people have vastly different perceptions of how important a given event is. In addition to reputation there is another way that someone might think about prestige: how high level is the fighting.
Most Prestigious: By Quality of Competition
For a more concrete way to address “most prestigious” we can turn to a measure of the Quality of Competition – aka how good the people who attend are. This is a much more answerable question, but it still has difficulties:
- How do you quantify “quality”?
- How do you factor that into a measure?
There is only one reasonable answer to #1: you use HEMA Ratings data. Everything else is far too small a data set, and far too subjective to individual bias. You will note that I said “reasonable answer” and not “accurate answer”. The influence of island effects on HEMA ratings is something of which the HEMA community is well aware. In general if you are looking at the HEMA Rating of two individuals in a match it is a good predictor of victory because people are mostly facing people from their same “islands”. So it does a pretty good job at pool seeding. Once you start factoring in the different regions compared against each other it’s not as applicable. (Not as applicable. It’s still a decent approximation.)
But let’s say that we decide we will use HEMA Ratings, accepting the inherent strengths and weaknesses of the number. Now we must get to difficulty #2, and it is much harder than people think. Usually the first idea that comes up in conversation is “average rating”. But that is definitely not a good idea.
Imagine you have a tournament of 10 people all rated 1500. Your tournament would have an average rating of 1500. Now let’s say that 10 people rated 1400 show up. You now have more fights that you need to get through to make it to the top, aka more chances to lose. And yet the average rating of the tournament went down. Clearly this is a poor metric.
In short, this is way more complicated than you think. I have more thoughts on the topic, but it’s a whole other article. For when I want to start more inter-regional drama I guess. (You may think that this is a disappointing answer to the section, but it’s far from my worst cop-out on the question.)
Another way to quantify the biggest is more from the logistical side: “How much crap is going on?” This is actually my preferred way to look at an event in terms of what goes into running it, which is a pretty good indication of size to me. (Note: if you’re looking for what tournament is going to be the best value for your money, checking the logistical burden on the organizer is probably not the metric you’re looking to base your decision on.)
Most Logistics: Tournament Registrations
One way to quantify the size is by looking at total tournament registrations. Remember back in the “Most People: Longsword Fighters” section where I said that a 20 person beginner and a 60 person open (that also included the same 20 people from the beginners) only had 60 total longsword competitors? This was because we were answering the question of “how many people showed up to fight longsword”. Now we are interested in “how much longsword fighting is going on” and the answer is 80 people’s worth.
Which in many cases is easier to calculate, as we don’t care about duplicate entries anymore. Though we still need to be a bit careful, as depending on the tournament software used having the same longsword tournament split off into an Tier A and Tier B bracket can result in having the tournament have somewhere between 50% to 100% more counts towards the tournament entry than it should. That said, let’s have a look:
I’ve also color coded the data by the number of days the event runs. With Green = 2, Blue = 3, and Red = 4. And you can clearly see a trend of more days = more tournament registrations. Which should come as a shock to absolutely no one.
While the number of tournament registrations is a good indication of “how much crap is going on” there are some ways it doesn’t tell the full story. Which takes us into a different metric.
Most Logistics: Number of Matches
Ascalon looks like the one outlier of the 3-day events, being relatively low on registrations. But Ascalon also does something else that isn’t normal in HEMA. Ascalon gives people pools of 7 and advances everyone to the bracket. For those of you who don’t organize tournaments, here’s a little challenge. Answer the following question off the top of your head: given that a pool of 4 fighters is 6 matches, how many matches would you expect a pool of 7 to be?
Got your answer? The correct answer is twenty one fights. Which is a lot more than people would normally assume. The number of fights in a pool doesn’t just scale linearly with size, it is geometric. The tournament format will play a big role in determining how much actual fighting is going on.
Said in other words: in the most extreme case a tournament with half as many registrations could lead to an equal amount of fighting. So is matches the best way to look at an event?
That’s the broken record of me saying “it’s not so simple”. While looking at matches vs exchanges gives us more information, in some cases it can make things cloudier. Looking at SERFO and Combat Con “Open Longsword” division we see that they have similar numbers of people fighting longsword, but Combat Con is giving way more matches. Which would seem to indicate that you’d get more longsword fighting at Combat Con, right?
Well not really. While there are many more matches at Combat Con, the number of exchanges a fighter will have is basically the same as between the two. In fact, on a per-capita measure it’s slightly less.
Ultimately, exchanges are probably the best way to evaluate how much fighting happened at any given event. But there is a major hiccup with this: not everyone makes that information available. Currently (to my knowledge) HEMA Scorecard is the only tournament software that shows how many exchanges were fought. Meaning that this data is simply not available universally, and can’t become the basis for comparison.
So, unfortunately, using matches is the lowest level of detail we reasonably have available. In the tournaments that I can see exchange level data there is significant variation in how many exchanges are offered per match; however I don’t have data for everything so I can’t really make any corrections with regards to this. Here is the list of events with over 450 matches. Just as before Red = 4 days, Blue = 3 days, Green = 2 days.
Oh boy, the asterisks.
*Just like with the total tournament registrations I tried to factor all the tournaments in. Given that cutting isn’t the same as a match, I assumed that each person doing a cutting round was the same as a single match. If you’re looking at the time involved for an individual competitor this is probably accurate; if you’re looking at the logistics capacity required to provide it then it is probably a significant under-estimate. Given how fuzzy this match-level comparison is already, I’m not interested in delving deeper.
** MCHO had some King of the Hill and Skills Course happening on the last day. I factored them in as one “match” for every entry, but I’m very certain that is leading to a significant under-representation of what was going on.
I thought that this could be a quick writeup to explain how difficult such a simple question actually is to define. But then I started realizing just how difficult it was to get some of these numbers. Depending on which software was used to run an event, finding the information you want can range from simple to impossible.
We don’t have a Quality of Competition metric, so we have to interpret the question of “Biggest” only with regards to participation. SoCal Swordfight has the most fighters in attendance, Combat Con offers a bigger variety of tournaments at the cost of fewer exchanges in each. Given that they come out to very similar total tournament registration numbers despite all that, I think that SoCal Swordfight deserves the crown.
Remember, Bigger Tournaments Don’t Mean More Fights! If you’re going to a huge event like SoCal to fight in multiple different weapon tournaments, do classes, free spar, meet new people, antique handling, play HEMA games, watch circus performers*, etc you will have a ton to do. If your intent is to fight the longsword tournament matches and go home you will probably find a smaller local event that is much more tuned to deliver that experience.
*I may have made this one up.
And, to put it all in perspective, here are some comparisons of 2022 events to Longpoint 2017. For those who don’t remember (or weren’t around yet).