The Importance of Seeding, or Why Can't I Win Any DE Bouts?

When I had only been fencing about a year, I started to become a regular competitor. I went to all the tournaments I could find. I did OK (not great) at the novice and E-and-under types of tournaments, but it seemed like I couldn't win a direct elimination bout at any major local tournament. I always seemed to get somebody for my first DE whom I couldn't beat.

As I gained more competition experience (as well as fencing experience and skill, enabling me to place higher in tournaments),I began to realize some of the reasons for why this happened. This article addresses one of those reasons, namely the seeding into the direct elimination. It’s directed mainly at intermediate fencers and beginning serious competitors, who have the ability to place in the middle of the field with the right approach.

I do not discuss repechage in this article, mainly because by the time you get to repechage, you are probably in a field of very strong fencers and the probabilities about whom you'll fence aren't very useful anymore.

What is the purpose of a tournament and a tournament format?

The ultimate goal of any tournament event is to rank all the participants in order of how good they are. If all the participants could be ranked discretely into best, second best, third best, and so on, then a successful tournament format would always place the participants in the final results in that order.

So, one easy answer to why a beginner can't seem to win his first DE bout is to look honestly at his ability on the tournament day (i.e., take into account whether he's having a good day or not) relative to the rest of the field. If more than half of the fencers in the tournament are fencing better than fencer X that day, then fencer X should finish in the bottom half of the tournament. All the fencers whose ability on that day is below the median (assuming no byes in this case) should lose their first DE bout.

What's with the weird assignment of bouts in the DE table, anyway?

At first glance, the pairing of numbers in a direct elimination table may seem random. In fact, it's structured very carefully.

Assuming no repechage, every direct elimination bout eliminates one fencer from the competition. Each round of direct elimination eliminates half the participants (less in the first round if there were any byes). Those eliminated fencers earn places in the bottom half of the size of that round. For example, in a DE of 32, those who lose their first round bout place 17-32. Those who lose their second bout place 9-16, and so forth. Ranking of fencers within those ranges is based on their initial seeds into the table. For example, in a DE of 32, the only difference between the 17^th and 32^nd finishers is how they did in the pools – 17^th place goes to the highest seed who lost in that round, while 32^nd goes to the lowest.

Clearly the best and second best fencers should be kept apart as long as possible, ideally until the final bout (round of 2). If they met in the round of 32, for example, then one of them would have to be eliminated and would place 17^th, while other fencers who are not as good would place higher. Similarly, the top four fencers should be kept from fencing each other until the round of four, to ensure that they’re not forced out too early, and so on.

To accomplish this, in each round the top seed fences the last seed, the second fences the second from last, and so on. And from this, we can determine that the initial seeding into the table is very important in determining whom you will fence. This is the real purpose of the round(s) of pools before the DE – to get an accurate seeding so that the top fencers will be kept apart as long as possible and will hopefully survive to the late rounds.

How your initial seed affects whom you'll fence.

Consider a fencer whose ability is exactly at the median of the field of the event – half the fencers in the tournament are better than he is, the other half aren’t. If ratings are distributed evenly through the field, perhaps he's a C. Assume the field is 32 fencers, so there will be 5 rounds of DE bouts.

Suppose this fencer performs very poorly in the pool and ends up seeded last – 32^nd. His first DE bout will be against the top seed. Even if our fencer was just slacking in the pool and can really fence at an ability level of about the median of the field, the top seed is likely to be a much better fencer – say an A. Our fencer’s prospects are somewhat bleak.

Suppose our fencer performs at his expected level in the pool and ends up seeded in the middle – say 16^th. His first DE bout will be against the 17^th seed, who had a similar performance in his pool and consequently is probably of similar ability. This bout is a toss-up – either might win, but neither is likely to get stomped. However, whoever wins this bout will face the top seed in the second round. If our fencer can defeat his first opponent (who is of similar ability), he will most likely get eliminated by his second opponent and finish 16^th.

Suppose our fencer performs well in the pool and ends up seeded fairly high – say 5^th.His first DE bout will be against the 28^th seed, who is probably far below his ability level and should be easy to beat. His next bout will be against the 12^th seed (or the fencer who upset the 12^th seed). Since our fencer's ability level is about equal to the 16^th seed, the 12^th is not likely to be a vastly better fencer. Our hero may be able to beat this opponent on a good day, and presumably he’s having a good day if he did that well in the pool. Only in the third round will he meet someone such as the 4^th seed, who is probably much better. Still, in this case he would finish around 5^th.

Get the best seed you can.

What conclusion to draw from all this? Although (or perhaps because) DE bouts can be more difficult to win than pool bouts, you should do everything possible to make it as easy as possible. One way to do this is to get the highest seed you can. The higher your seed, the lower the seeds of the opponents you will face. This means that with a much higher seed, you have a good chance of surviving one or two rounds longer before you end up against someone whom you have little chance of beating.

In the pool, everything contributes to your seed. Victories are of course the most important thing. The first criterion for ranking fencers based on pool performance is the ratio of victories to bouts fenced. If you win all your pool bouts, you will seed among the top fencers in the tournament; your worst possible seed is equal to the number of pools. Conversely, by losing all your pool bouts, you guarantee a seed near the bottom of the field.

However, even each touch can have some effect on your seed, because it's likely that you will be ranked against other fencers in the tournament who have the same ratio of victories to bouts fenced. The next criterion is the “indicator” – hits scored minus hits received. Every touch that you can score on an opponent (even an opponent who eventually beats you) and every touch that you can prevent an opponent scoring on you (even an opponent whom you beat) will improve your indicator. This is especially important if not all of the fencers will be promoted from this round of pools. I have made it to the second round of a tournament based solely on indicators, making the cut by only a couple of touches.

So, in general you should try to get every victory and even every touch that you can. Never assume that an opponent is unbeatable. Particularly in the first round of pools, some of the top fencers may not be fencing their best, either because they don't feel the need to warm up as much as the devoted intermediate or because they're saving their energy for the finals; or, you just might have a quirk which they can't seem to deal with in the space of 5 touches.

It’s equally important to fight for every touch against beginners. If I’m fencing in a tournament that I want to do well in, and I have a pool bout against a ten-year-old who’s been fencing for a month, I will beat him 5-0 if I can. I probably will not fence any more aggressively than necessary to accomplish this (i.e., I probably won't fleche on every touch and leave Adidas tread marks up the front of his jacket) but I will not let my own seed suffer out of sympathy for an opponent, lest my final placement suffer gravely.

An example:

I was presented with a good example of this theory at the 1997 Leon Auriol Open in Seattle. The names of the protagonist have been eliminated for genericity, but they know who they are!

In the men's épée event, there were 33 fencers. I, a C97, did fairly well in my pool, and I ended up seeded about 5^th. My first DE was against a relative beginner, whom I beat fairly easily. My second DE was against an experienced but not very mobile fencer, who is comparable to my own level – I can beat him if I’m having a good day, but he can beat me if he's having a good day. As it happened, I was having a good day that day and he wasn't, so I won that one. My third round bout was against one of the several A’s in the field, and he trounced me. I finished 6^th.

One fencer whom I fence against often, comparable to my level or a little better (also C97), complained to me afterward about his relatively poor performance – placing 15^th. It turned out he was eliminated in the second round of the DE by his coach (a very experienced B fencer). I asked him what had happened in his pool. He said that in one bout against a junior fencer of significantly lesser ability (but definitely no door mat!), he decided to play around and try to get toe touches. He tried them about three times and got hit each time. Because he gave up those touches and was perhaps a bit over confident, he lost that bout.

He committed a minor sin by giving away touches for no real reason. He committed a major sin by losing a bout which he should have won. The result was a lousy seed, and in the second round he encountered a fencer whom he had little chance of beating and got eliminated. Because I managed to do fairly well in my pool, I made it to the third round before encountering somebody I couldn't beat, so I placed much higher.

Exceptions and caveats.

Obviously this is not a 100% perfect assessment of the problem, only a reduction of one specific aspect of it. There are several factors which can throw these things awry:

• A higher seed does not necessarily always beat a lower seed. If that were guaranteed, there would be no point in running the DE table at all. Nevertheless, it's fair to say that your chances of beating the lowest seed in the tournament are probably a whole lot better than your chances of beating the top seed, and the longer you can avoid fencing one of the top few seeds, the better.


• DE bouts are more difficult to win than pool bouts. Because they last longer, they require greater mental and physical endurance. Also, they tend to require a larger repertoire of techniques and tactics. A trick that can get five touches on the average opponent can win a pool bout, but only gets you one-third of the way through a DE bout. If that's the only trick in your book, you just might lose 15-5. Still, the strength of your opponent is an important factor.


• The above assessment assumes each fencer is a fencing machine who can perform exactly at a particular skill level indefinitely. Of course, fencers are people who have endurance limits, mental patterns, and so on. If winning all your pool bouts completely exhausts you, that top seed may not do you much good – you might be too tired to win more than one or two DE bouts, whereas if you conserve your strength you might last longer. Or, perhaps you don't stay warmed up and focused if your first DE bout is too easy. In that case, being the top seed might be counterproductive.


• It's not necessarily a mortal sin to sacrifice a touch if you have a good reason and you are capable of dealing with the consequences. For example, if you can sacrifice a touch in a pool bout to gain valuable reconnaissance information which will enable you to win the bout, that may be a valid strategy for you. However, if you can gain the same information without sacrificing a touch, that's better.


• In a DE, the final score doesn't matter, all that matters is who wins. If you win 15-14, you still face the same opponent in the next round whom you'd face if you had won 15-0. Still, giving away touches in a DE bout can be a bad idea. A run of three or four touches where you’re just playing around might be enough to weaken your concentration and give your opponent a morale boost, and turn the tide.

Conclusions...

Always fence with intent to win the tournament. Afterwards over dinner, you can discuss your final placing with your coach and decide whether it was a good result or not.

Barring any strategy to the contrary, always fence for every bout and every touch. Never throw away touches for no reason, and never ever throw away victories for no reason. It's better to lose 5-4 than 5-0. It's better to win 5-0 than 5-4.

Yes, some of the seeding into the DE is luck. If the 14^th seed is somebody you've always been able to beat easily but the 15^th is somebody you've never been able to figure out, you have to hope that your seed doesn't end up being one that faces the 15^th at some point – which one of the two you’re likely to face eventually is not something you have a lot of control over. Nevertheless, encountering an A fencer for your first DE has very little to do with luck. Putting forth your best effort in the pool can often postpone that meeting for one or two more rounds, and you'll finish a lot higher because of it.

Prevot Greg Jones is a coach and the owner of Rain City Fencing in Bellevue Washgton. Copyright © 1998 by Gregory A. Jones. This article may be reproduced freely, as long as it remains unmodified and his copyright notice is included.