Ranking

[graphixaddict]

the best system i have found is the USCF system (united states chess federation) take a look here http://www.64.com/uscf/ratings/
for the most part i think its automated, and on the hacker note...just keep it independent of the true ranking files, keep it on another server. for the most part a computer that holds the rankings of every fencer in the world should not be connected to the internet anyyways its just to risky... it would be very easy however to just backup the files to a tape drive each night..the ones from the independent computer and update the files on the internet accesible machine.
Graphix

[Dave]

The other thing that the british system does is that it accounts for the quality of the fencers at the event.

This works out roughly as

(a factor depending on position in the competition)*(the number of people in the top 50 attending)=number of points

Only the top 64 (or 75% whichever is lower) of the competition get any points.

Your national ranking is the calculated by adding your best 6 scores.


This means that you can get far fewer points from winning a small open with no real competition than you can from scraping into the 64 at the national championships.


Seeding at competitions is then done on you rankings


details can be found here
http://www.foilcommittee.pwp.blueyon...02-3-final.htm

[oiuyt]

Gav- Depending on where fencers live, fencing across state and/or division boundaries is not always rare. This season, exclusive of national competitions, I've fenced in six divisions in either five or six states. That's just counting local events which I drove to. Am I a standard fencer? No, not by a long stretch, nor is my area of the country typical. But there's a LOT of cross border fencing around here.

EDew- The gain and the loss in the chess sytem are zero-sum. The points that the gainer receives are equal to the points that the dropper loses. How does this limit the top players? How does it discourage competing? (also as a side note, if the difference between the ratings of the players is sufficiently large the outcome is either not used or is capped as if said threshold difference was the actual difference, I forget which) The whole theory behind the system is that with a ratings gap of X points the expected outcome can be probabilistically predicted. The difference between the actual outcome (0, .5, or 1 wins for each) and the predicted outcome (0..1 in a binary distribution where 100 ratings points == 1 standard deviation) is the scaling factor for the number of points shifted. If player A has 100 more points than player B and wins player A moves up ~11 points (numbers are made up but are reasonably close, but I haven't played competitively for ~8 years so I might be off a bit) and player B moves down ~11 points. If player B wins, he moves up ~22 points and player A moves down ~22 points. If the game is a tie, player A moves down ~5 points and player B moves up ~5 points. If player A and player B played 100 games, it would be expected that player A would win (counting ties as half a win) ~67 of the games and player B would win ~33 of the games. Over 100 games the ratings change would be expected to net 0 change to either player. The source of change is when the ratings DON'T accurately reflect the relative difference in abilities of the two players. When this happens there SHOULD be a change and there is in the chess system. Yes player A is risking more is s/he loses than player B is, but player A is less likely to lose in (assuming the ratings are accurate) to an exactly proportionate degree as the added risk.

Such ratings are used to seed tournaments. And to limit who various tournaments are open to. Of course all that really matters is at the end of the day(s) who has won and who has lost to take home the hardware and the money. Sounds pretty much like our system. Now go try to convince competitive chess players that the ratings mean nothing other than a convenient way to seed. Why should fencers accept that our system doesn't have additional uses? Granted our system is more flawed than theirs so it doesn't work as well or as effeciently (in the sense of moving people to an accurate level quickly rather than in the amount of overhead sense). Does that mean that it can't be used as a rough gauge or that it shouldn't be, or that it isn't? Of course not. Pretending that the classifications don't have meaning other than for seeding purposes is stupid.

Do I think we should impliment a chess-like system? No. Much as I would love it, I think that the cost would be more than is justified by having better ratings (better == more accurate). In the chess world something like $0.50 goes to maintaining the rating system for every tournament game played. That would add a couple of dollars to every local tournament entry fee. Not much, but enough that I'd prefer to stay with a variant of our current, flawed system. I'd like to see improvements to remove the easy to remove flaws and to fix inconsistencies, but I'm willing to live with something very close to what we have (good thing given that I have little power to change the system).

-B :)

[Chris]

Originally posted by Gav
As I understand it, you guys have different organisations that deal with each state. Why couldn't each state deal with it's own ranking list?

Well, it's not QUITE that simple...many states DO have their own Division, yet some larger states (either in area and/or fencing memnbership) have several, while some smaller states (mainly membership-wise) are combined: Manhattan Island alone is its own Division!

Yet all these divisions are subordinate components of the National organization...they have certain autonomy and freedom, but, must also operate within certain constraints. The classification system is one of them...certainly a division COULD make up some rules of their own, but the national would then disregard the results, which would most likely lead to fencers NOT coming to the avent again.

If you want an American-wide list you would have to establish a system of comparing the rankings across the US.

The National DOES run list of points earned at the Natioanal event throughout the year, which accounting for age groups, etc, means about 4-5 events per year, HOWEVER, only a small portion of membership participate in these events.

How often do people meet to compete inter-state?

Quite a lot! In addition to 5-6 events in my local area, I expect to travel to 5 different states for 6* other events this year, and that doesn't include our sectional (multi-state regional) event which has rotated to my home state this year, plus our Nationals. *(next year it may include 2-3 more) THis is not abnormal for people in many parts of our country.

You would, of course end up with a league system where fencers would be ranked in their own division then those who were able to compete inter-state would then be entered onto the national [Premier?] League. It's a multi tirered system but I don't see an easy way away from that . The US is the size of a continent.

Well, first of all, the U.S is far too egalitarian, and libertarian, to have that fly, and if we DID have that, well the debate about our classification system would be all the more intense, as it would, in fact, be literally fencing us in!

However, in a sense, we already have this, as in the late 90's, the USFA instituted a Division 1, and Division 1A.

To qualify to the 1A Championship event, one must qualify at the Sectional Championships; to qualify for the Division 1 Championships, one must score 'points' (place in top 48, or top 40%) at a National Circuit event* - to enter a National Circuit event, one must now be classified as 'C' or higher*. *(there's a few loopholes for top juniors, and a couple others, but you get the idea...)

Database Administration is fairly easy it's just the will to implement a scheme that seems to be holding you back.

Well, trust me, there's plenty of will rolling around, it's just not all rolling in the same direction!

[oiuyt]

Originally posted by Chris
Well, first of all, the U.S is far too egalitarian, and libertarian, to have that fly, and if we DID have that, well the debate about our classification system would be all the more intense, as it would, in fact, be literally fencing us in! ;)

Actually we're likely moving to something very similar to this. The USFA is now getting to the size where regionalization and a tiered system to qualify to all national events is going to happen sooner or later. The RYCs are a test bed to get kinks worked out before rolling it out for everyone. Eventually you'll have to do well regionally to even be eligible for national competitions.

As to the debate, no. The HPC will slowly add more and more to the plan until we're there. Boiling a frog slowly and all that.

-B :)

[135711]

[i don't want to mess up thread, but someone told me to take off my name, so, i thought about it and i did] thanks.

hopping into other folder - keep the world safe while i'm gone


wasn't that the way it once was? you had to qualifly to meet at nationals?

[Chris]

Originally posted by oiuyt
Actually we're likely moving to something very similar to this. The USFA is now getting to the size where regionalization and a tiered system to qualify to all national events is going to happen sooner or later. The RYCs are a test bed to get kinks worked out before rolling it out for everyone. Eventually you'll have to do well regionally to even be eligible for national competitions.

There's a pretty big difference in the two: currently, our stratified system only limits your ability to fence up and down the strata: you are restricted from the upper levels, unless you qualify, in a specified manner, conversely, if you have 'proven' yourself skilled to some certain degree, you are held out of the lower level(s); (IIUC) his idea completely limits your ability to fence OUTSIDE you geographic area, unless you have risen to (somewhere near) the top.

As to the debate, no. The HPC will slowly add more and more to the plan until we're there. Boiling a frog slowly and all that.
-B

Yes, and there's a goodly few of us who aren't quite too happy about what they have done so far!
Granted, some of this IS sour grapes, as I always liked the idea that I could (still*, sometimes) hack my way through sectionals to THE National Championships, but perhaps, if the sport is to get larger (which I guess I DO want), this is not a workable model. C'est la Vie!

*(time WAS when you qualified for Open Nationals directly from your Division; which is why they are no longer called Open Nationals! BOY, you should have heard the 'elites' complain about THAT! 200+ guys in MF; A's, B's & C's - no D's, or E's [although a C, then was like a D1 now, except you needed 18 vs. 15], 2 flights of ~20 pools, and it took a whole day of pools, just to get down to 48! Of course, is helped, a little, that REGULATION time was, then, 6 minutes, with UNLIMITED overtime!)