Should there be a level of USFA tournament between the current X2 and X3 level?

[KD5MDK]

I'm looking at the tournament rating chart on AskFred, and it appears to me that there's a significant gap between the C2 to C3, or B2-B3, etc level tournaments. To go from a requirement of 25 to one of 64 fencers is a bit of a jump, and one that most tournaments don't make that I see in my section. In particular, because the Southwest Section is so Epee heavy, it is difficult to achieve an upper level tournament in Foil or Sabre. I'm thinking that a a theoretical C2.5 B2.5 A2.5 tournament with say, 40 fencers required makes a logical fill in the gap between the current offerings. It would give tournament organizers something to shoot for below the B3 level that currently is unobtainable in some events, and encourage people to attend tournaments that stand a chance of making that mark.

(For my examples, I'm using the 2003 Longhorn Open and the 2004 Longhorn Open, the 2004 Crescent City Open and 2005 Crescent City Pre-Pays , the 2004 Rose Condon and 2005 Rose Condon, and the 2003 SabreCave and 2004 SabreCave. I picked them as big tournaments in the area I'm familiar with.)

Here's an example: In the SWS, a good sabre tournament will be a B2. There is only one A sabre fencer who comes to any local tournaments, and so in the year I've been fencing here, I've only seen an A sabre tournament when the Rose Condon flew in another A from New York. The gold medal bout of that tournament was incredibly good to see.

But outside of that, the best we'll ever get is a B2. Because there simply aren't 64 sabre fencers who will show up to a single event, and I don't see that changing too much soon. We can have a tournament I'm sure where 4-6Bs,4- 6Cs and 4-6Ds and 40 total fencers show up. But I don't think we can do 64 and 24Cs+12Ds, even with a Sabre only tournament (SabreCave).

Foil is in roughly the same trap.
Here's the Tournaments and their ratings:

2003 Longhorn Mixed Foil: 89 fencers. 1A, 8Bs, 7Cs, 14Ds, 15Es. A B2 Event.
2004 Longhorn Mens Foil: 70 fencers. 1A, 8Bs, 13Cs, 11Ds, 12Es. A B2 Event.

2004 Crescent City Mixed Foil: 143 fencers. A2. (I don't have time to add them up from AskFred)
2005 Crescent City (Pre-paid) Mixed Foil: 86 fencers. 7As, 6Bs, 14Cs, 17Ds, 15Es. Possibly A2/B3 event.

2004 Rose Condon: Mixed Foil: 66 fencers. 1A, 3Bs, 6Cs, 13Ds, 14Es. A B2 Event.
2005 Rose Condon: Mens Foil: 45 fencers. 4As, 5Bs, 4Cs, 9Ds, 8Es. An A2 Event.

Sabre:

2003 Longhorn Mxd Sabre: 39 fencers. 4Bs, 5Cs, 2Ds, 9Es. A B2 Event (This is where a U won the tournament and got his B)
2004 Longhorn Mens Sabre: 40 fencers. 1A, 4Bs, 5Cs, 5Ds, 7Es. A B2 Event

2004 Crescent City Mixed Sabre: 47 fencers. 1A, 5Bs, 7Cs, 3Ds, 11Es. A B2 Event.
2005 Crescent City (Pre-paid) Mixed Sabre: 47 Fencers:*1A,*3B's,*4C's,*7D's,*12E's Could be a B2 event.

2003 SabreCave: 34 fencers. 3Bs, 4Cs, 2Ds, 8Es. A B2 Event.
2004 SabreCave: 43 fencers. 1A, 3Bs, 4Cs, 4Ds, 5Es. A B2 Event.

So, to recap my argument:
There's a level of competition occurring between the B2 and B3 levels, and which could also happen inbetween the A2 and A3 levels, and the C2 and C3 levels. It is bigger than a 25 fencer, only 6 rated fencers needed barebones B2, but not at the level of 64 fencers, 24 Cs B3. I think that it would be fairer for there to be a tournament rating established to recognize this gap.

Or maybe I'm just whining that to advance above an E in my section at many events you must beat a C to place in top 8 and earn your D (if ratings were determining placement).

[Inquartata]

It's not an uncommon state of affairs. Even larger sections like Pacific Coast don't often manage A events, though they sometimes get the 64 entrants. And as for the more sparsely populated sections, well...we're lucky to manage 15 sabre fencers all told. Plus we only have Cs by virtue of a few people who travel out of the section to bigger events. ( And you're complaining that you "only" have B2 events! We are envious! )

Actually, I think the whole idea is NOT to offer many opportunities to upgrade ratings. That would mean more people qualified to enter Div I and Div II events, which the USFA is trying to retard in order to keep numbers down...

[mrbiggs]

I've always thought that since so many of tournaments are now computerized, it is practical to have an equation to determine the ranking of a tournament. The top rankings would stay as is; it takes an A to make an A, a B to make a B, etc., for the most part. But the number of B's, C's, D's, E's, etc., would change.

For example, a tournament with 3-4 A's, 6-8 B's, 10-12 C's, and a few D's E's and U's should, in my opinion, give out more rankings than it does. A U needs to finish in the top 12 (in other words, come out ahead of an A or B) just to get an E, even when a finish in the top 20 is a performance that certainly deserves the letter. (It would mean the U came ahead of a C, B, or A...and a real U would not be capable of doing that)

[Jarentz]

Also, why aren't there catagories for the National or other large tournaments for 128 or even 256 fencers?

[telkanuru]

National tournements have pre-set levels. By and large, though, points are the purpose of national tourneys, not ratings.

The argument seems to be that you want your tournements to be artificially inflated to give out higher ratings, quite frankly. There was an E and Under here last weekend with 51 people in MF. The winner got a D, of course, but just because there were a lot of (not so good) fencers, dosn't mean that winning such a large tourney should have got him his C.

Also, even when a division has As, you may not get what you want. For example, in the NE division, the Hangover Open Mixed Epee had 6As, but only one made it to the L8, and therefore it was a B2

[KD5MDK]

What does "artificially inflated" mean in a system arraigned in a relatively arbitrary manner? It's not like I'm saying that we should abandon DEs or switch to pure indicators scoring or something. I'm just suggesting that the tournament rating scale might be adjusted a bit.

The argument using an E and under is not at all the same, because that's a rating limited tournament. A fencer who defeats a number of Es, or potentially could have, ought to be rated a D. In an E and under, there's no evidence he could have beaten a D, so no reason to give him a C.
The point I'm bringing up is that it's quite possible, even likely, that at a B2 tournament, the top 1-4 can be A and Bs, the top 5-8 be Cs, and so even if a fencer defeated many Ds and even Cs, still not earn above an E.

Suppose there was a tournament where 11 As and a B were the only participants, and the B came in 3rd. Wouldn't that indicate that the B deserved to be an A? However, the event ratings chart disagrees. Yes, it's a contrived example, but it illustrates the problem of an excessively rigid chart.

Levels of Competition, # of fencers more than previous level, # of rated fencers required:
E1, 6, 0
X1, 9, 6
X2, 10, 6
X3, 39, 36
A4, 0, 36
(D1 ignored as a special case)

Doesn't the big jump stand out to anyone else?

My suggestion is that there be an X2.5 tournament (ok, I'd say call it X3 and move all the other numbers up 1 until A5) and have its requirements somewhere around 40 fencers, 4X, 4Ys, 4Zs. Awards? 1-B, 2-4 C, 5-12 D, 13-16E for a B2.5, adjust one letter up or down for the rest. . Or something else. I was hoping to work on this during discussion.

It would make the middle of my chart look like this:
X2, 10, 6
X2.5, 15, 12
X3, 24, 36

That looks reasonable, but all the numbers are subject to discussion. I just want to see what people think.

On another benefit of this, it would enable good fencers from weaker sections to keep from being kept out of the top spot. Thus, it would be more reasonable to restrict Div I to Bs and As, and not allow Cs. That ought to make the national level happy if they want to restict attendance.

As for all the As being knocked out and breaking the tournament, them's the breaks. It's not exactly as if a Hangover Open is going to be a demonstration of everybody at their best, given how literally I think most fencers would take that title.

[peet]

Without expressing an opinion either way on the event classification question, I'd just like to say:

It's really cool to see the data available on askFRED.net used this way to help understand what's going on in our fencing world, and to think about how it could be improved.

That's one of the main reasons I do the FRED thing. Glad to see it's working!

Rock on, everyone!

-p

[telkanuru]

Let me restate your argument to make sure we're on the same page. You are in a situation where you do not have enough persons of sufficiently high rating to achieve an (X+1)1 or an (X+1)2 and not enough bodies to achieve the X3, ie. holding a B2 because you lack 1 A to make an A2 and the 64 people needed for a B3.

If the above is correct, I would first point out that you are in a rating limited environment, so the E and under example would apply. Second, that the current rating criteria is based off of how many people of theoretically increasing levels of difficulty you must beat to end up in a certain place. It seems to me that any 2.5 system would be inherently easier to earn a rating because of this. Example:

At a B2 w/32 ppl, assuming pool results approximate initial seeding and 1 A, a couple Bs, a few more Cs, and a smattering of Ds, Es, and Us for good measure, the person (3 or 4th seed) attempting to earn a C should not face the A until the L4, and 1 or 2 Bs in the L8 and the L16, and fence a weak fencer in the L32. The high level fencers in the quarterfinals serve as buffers for pool flukes while defeating an A in order to earn your B is not extreme, and nor is defeating a B in order to earn your C.

At an A2, it is much the same, only you should have an A on either end of the bracket weeding out weaker fencers before the semifinals along with the Bs. Again, you will most likely have to defeat someone 1 rating higher than the one you achieve in order to obtain it, but this is not an extreme challenge.

At a B3, you have more (and higher) ratings because there are extra buffer levels and more fencers of higher rank, theoretically, meaning the lower fencers do not make it into the L16 or even the L32.

In short, 2 methods are utilized to ensure the quality of high-rated fencers: the need to defeat one very high rated fencer or multiple medium-high rated fencers. It seems to me your 2.5 system does neither, allowing more randomness and fluke chance into our already rather unstable system.

On a lighter note, I had to go through 3 Bs to get my C, so I don't see why you all should have a break

[acaba]

Originally Posted by mrbiggs

I've always thought that since so many of tournaments are now computerized, it is practical to have an equation to determine the ranking of a tournament.

There was discussion in a past American Fencing article about a numerical rating system similar to the one the Chess Fedreation uses. (I couldn't google it. Does anyone have a link to it?) The main objection to this is that EVERY tournament would have to be entered into the database both accurately and quickly. Luckily, I'm in the VA division and things run really smoothly here, but I have heard of horror stories from other divisions where things are not as good.

While many of the tournaments are computerized, enough are not that I think there would be a problem getting the less advanced/organized people to submit their tournament data.

Back to the subject, I think there may be a gap in the rating system too between 25 and 64 fencers. Or you just need to get another fencer to go out and earn an 'A'. This happened when the second men's epee fencer in the VA div earned his A a few years ago. Over the next 2 years, the upper level competitive fencers got a free upgrade of one letter.

[oiuyt]

The PdT Open MS last year had 57 fencers including 2 A's and 11 B's. The two A's did their job, finishing 1st and 2nd. TEN of the eleven B's failed to make it into the round of 8, resulting in a B2 competition. Weird stuff happens.

KD5MDK- I like your suggestion for an X2.5 level. If anything it looks like a high requirement given that it doesn't change anything for the upper level ratings, but merely extends the D/E range. Perhaps drop an email to Kalle Weeks as chair of the Tournament Committee just as a suggestion for consideration as a future change?

Originally Posted by telkanuru

ie. holding a B2 because you lack 1 A to make an A2 and the 64 people needed for a B3.

If the above is correct, I would first point out that you are in a rating limited environment, so the E and under example would apply.

Really? Sounds like an open to me. How many ratings-restricted events do you know that have 40+ people and 1 A, 4+ B's (or ANY A/B fencers... granted I have once seen a competition that was limited to B and below (stupid))?

Aaron-
The numeric system was abandoned before the approved test year for a number of reasons including, in no particular order (and likely an incomplete list) the additional overhead (the USCF charges tournament organizers a fee per game played to enter the games into the system), lack of transparency, and the fact that almost no one at the national level with the exception of George Masin understood it. Having been an active competitive chess player growing up who got interested in how the rating system worked back, the proposed system, as presented in AF made sense. If you don't have a feel for how these things work, generally including having gone through a few specific example cases, and/or don't enjoy applied math, the system looks scary and confusing. This is a very good way to guarantee a lack of adoption. Added to that is the fact that, while an improvement, our current system works fairly well most of the time. There isn't a large enough incentive to add a noticible overhead for a slight improvement.

I was originally a huge proponent of the numeric system. The said, over time I have come to shift camps completely. I understand the system well, I like the system, I think it would be a noticible improvement over our current system, which I see a number of flaws with. I don't, however, feel that the numeric system is enough of an improvement to offset the added cost (in time, money, and confusion, any one of which is, alone, enough of a cost to make the change not worthwhile in my opinion).

Chess basically just tells people not to worry about it, to accept that winning games will improve your rating, losing them will hurt it, and that how much depends on the relative ratings of the players. And that works fairly well. People are reasonably happy with that. Most tournament organizers don't know how the system works. They just send in the results and are content to use the quarterly reports as "good enough" to seed their tournaments.

Our system is "good enough" that systematic changes, especially major changes, need to be spectacularly better. Minor tweaks (allowing age-restricted events to award classifications, then limiting the younger ones to only C1 or better levels to do so, KD5's X2.5 idea, etc.) can be accepted with a much lower threshold -- merely improving the system enough that having a new chart each year or so isn't too confusing that the added confusion costs more than the improvement.

-B :)

[acaba]

oiut, I remember that the numerical system looked like a big PITA to administer. I think the ABC system we use works 'well enough'

I think that adding a X2.5 category would be a good improvement to the system. I don't see how awarding extra C's, D's, and E's would have a negative effect on the system. We were worried about 'ratings inflation' a few years ago here in VA, but after further review and comparing the numbers of ratings of competitive fencers against the national numbers, we found that we had a relative LACK of C's, D's, and E's.

Handing out more ratings for mid level tournaments would also be a good way to encourage people to stick with fencing. Right now it takes the students in my college club about 2-3 years to earn their E. This is a long time before you can demonstrate to your parents that you have been spending their money wisely on 'that fencing stuff'.

Has anyone proposed adding more letters to the clasification scheme? I know that Karate has about 10 levels of 'belts' you can earn. Maybe adding more letters would add to people's enthusiam.

[telkanuru]

Originally Posted by oiuyt

Really? Sounds like an open to me. How many ratings-restricted events do you know that have 40+ people and 1 A, 4+ B's (or ANY A/B fencers... granted I have once seen a competition that was limited to B and below (stupid))?

Clarification: the event is rating-limited due to the fact that few people of high ratings show rather than on the direct part of the organizers. If you frequently have tourneys where noone higher than a B shows, then you effectively have a B and Under.

[KD5MDK]

The real key factor that makes an "E & Under" or similar event different is that the winner is given a rank in excess of that of his opponents. If you win a tournament of 15 fencers, you can get a D despite never fencing against one. Your suggestion that the winner of a large E and Under should get a C under my proposal skips a rating level entirely. I never suggest that to get a rating in my new level you don't have to beat someone of that rating. In order to get a rating, you ought to beat someone of that level. It takes beating an A to get an A, B to get a B, etc. I'm aiming to preserve that. What I'm saying is that the current upper level events are too big to be reasonably expected in many areas, but the lower ones are too small to capture the level of fencing going on. If the top 8 fencers at a competition are Cs, there will be no one with Ds if you never go above a B2. I think that if you're aiming at seeding pools using ratings, and nothing more, than whether a person fences at the level of D, an E or a U has some importance. Or we could just abolish those ratings as obselete.

[mrbiggs]

Originally Posted by KD5MDK

The real key factor that makes an "E & Under" or similar event different is that the winner is given a rank in excess of that of his opponents. If you win a tournament of 15 fencers, you can get a D despite never fencing against one. Your suggestion that the winner of a large E and Under should get a C under my proposal skips a rating level entirely. I never suggest that to get a rating in my new level you don't have to beat someone of that rating. In order to get a rating, you ought to beat someone of that level. It takes beating an A to get an A, B to get a B, etc. I'm aiming to preserve that. What I'm saying is that the current upper level events are too big to be reasonably expected in many areas, but the lower ones are too small to capture the level of fencing going on. If the top 8 fencers at a competition are Cs, there will be no one with Ds if you never go above a B2. I think that if you're aiming at seeding pools using ratings, and nothing more, than whether a person fences at the level of D, an E or a U has some importance. Or we could just abolish those ratings as obselete.

I agree that a fencer should have to finish higher (not necessarily beat) than a fencer of the letter he's earning in order to earn a letter, except in the case of D's and E's, for two reasons. One is that it would be very difficult to get D's if people could only win them at D & Under or higher tournaments, same with E's. The other is that D's E's and U's are not all THAT separated in terms of skill, so the winner of a 20 or 30 person unranked tournament should get a D, even if he hasn't beaten another D.

[KD5MDK]

Sorry. I agree with current policy in that regard. I was responding to telkanuru's suggestion that I was advocating the winner of a large E & Under getting a C.

You should never earn more than one rating ahead of people you've finished ahead of. With enough Cs and Ds, you can have a B3 tournament and earn a B, despite no Bs being present. But you have to come out ahead of a lot of Cs for that to be true. In general, this is not the case.

In any event, what my project is really trying to do is:
1) Make there be an intermediate level of tournament, mostly as a target for organizers and attendees to shoot for, to increase the size and prestige of otherwise marginal events.

2) Give more opportunities for earning lower ratings (D and E) so that fencers above the unclassified level will be better distributed through pools. The ability to show families and fencers that progress is being made is a nice bonus. Es are not precious ratings to be hoarded carefully. It's entirely possible to get an E without winning any DEs.

[glowstix]

Originally Posted by KD5MDK

It's entirely possible to get an E without winning any DEs.

its also entirely possible to get a D without winning a DE also.. . in my limited experience with the ratings system, i've come to realize that it all evens out in the end. you'll eventually get what you deserve. i've beaten tons of C's and a few B's in DE's at times at tournaments yet i'm still the dreaded D. i once went 5-0 in pools and still drew 2 B's in succession for my first 2 DE's. i trust that eventually i'll move up but it doesn't matter really. high rating or not, you still have to fence the pools and DE's. if having a high rating means you skip the pools altogether then i'll be more obsessed about it.

if you're using ratings only to measure progress and to "justify" why you fence then perhaps the X2.5 would work; i agree that its nice to have something to measure you skill level but in reality you'll never have a perfect system.

[dekko]

Since the subject was broached, has there been a Div II event at nationals in the last five years that didn't have a B or higher rated fencer in it? Seems as though everyone ends up with a B or A rated fencer in it. The Div II NACs usually have no Bs or As but seems as though nationals always has Bs and/or As in the Div II events. I know they get them at sectionals or other events.

As for the points v letter ratings system, remember the following coaches and club owners have to be able to explain the rating system to parents of new fencers and new fencers themselves. Those who really like the points systems proposed already have them, their called national and international points.