USFA membership/website/rating system update

[telkanuru]

Originally Posted by Inquartata

Why? If they really are better than the "lesser" fencers, they're going to beat them in any case. It doesn't matter if they're seeded 1st on points or 5th as a random A06. Either way they're going to be top seed in their pool, they'll come out with good indicators and probably rise in the seeding for DEs. Where is this clear benefit? I just don't see it.

I will ignore your comments about epee for they are obviously the creation of a weak and addled mind.

When you have 120 As at a competiton (Men's Foil, Epee), then it's not the difference between 1 and 5th, it's the difference between 1st and 120th. If you had random seeding within all As regardless of points you could have, say, 2-4 people in the top 10 of sr. points in the same first round pool fairly easily. Sr. points provides a good deal of granularity in a place where a good deal of it is required.

[MikeHarm]

Numerical system might work good in an automated system.

I could see fencers go to the tourney, the tourney folks punch in their id numbers etc, laptop connects to a ratings server downloads the current info on their stats, autoseeds the event then after its over they upload the results to the ratings server and you could show everyone what their new rankings are.

The current system seems pretty good for a snail mail way of mailing things back and forth but thats getting more and more obsolete every year.

[Inquartata]

Originally Posted by telkanuru

I will ignore your comments about epee for they are obviously the creation of a weak and addled mind.

Of course you will. The famed epee blinders work very effectively.

When you have 120 As at a competiton (Men's Foil, Epee), then it's not the difference between 1 and 5th, it's the difference between 1st and 120th.

No, it isn't. No one on the national point list is very likely to be an A05, A04, or A03, who are going to make up the bulk of the As. They are going to be A06's and thus seeded ahead of all lesser A's anyway. So your example exaggerates the possible spread---it isn't anything like 1 to 120.

If you had random seeding within all As regardless of points you could have, say, 2-4 people in the top 10 of sr. points in the same first round pool fairly easily.

But we don't. We seed A06s first, then A05s, etc. This is a straw man argument...

Sr. points provides a good deal of granularity in a place where a good deal of it is required.

Except you haven't shown that, at least yet. I may be convinced, but you will have to show me that points add something useful to our existing system, not to a cruder version of our existing system.

[Sciurus-Rex]

Originally Posted by Inquartata

((In response to, "If you had random seeding within all As regardless of points you could have, say, 2-4 people in the top 10 of sr. points in the same first round pool fairly easily."))

But we don't. We seed A06s first, then A05s, etc. This is a straw man argument...

Firstly, you keep using "straw man argument," but it doesn't look like you fully understand the term: A straw man argument is a logical fallacy based on misrepresentation of an opponent's position. To "set up a straw man" or "set up a straw-man argument" is to create a position that is easy to refute, then attribute that position to the opponent. ... This didn't happen in the example you cited by telkanuru; he never implied attributing his position to anyone else, and it only seems easy to refute because you act as though you're refuting it -- although you're not.

In fact, you're setting up something of a straw man yourself by misrepresenting his position. He has given a hypothetical example of large numbers of As potentially being lumped together in inappropriate pools by nature of their letters alone, but you quickly dismiss the example with two faulty points: First, that it doesn't happen now, and second, that we seed by year identifiers.

The fact that we seed by year identifiers as well as letter grades wouldn't matter in telkanuru's example because his sample of 120 fencers is bound to include many As with the same year identifier; so it's still possible that randomly sorted pools could include multiple top-10 fencers. ... That it doesn't happen now is an empty argument because high-level events also apply national points for seeding purposes (and organizers can also manipulate pools by hand to reduce inappropriate random outcomes).

In other words, telkanuru's hypothetical is still a valid example for discussion, and your allegation of "straw man" logic to dismiss his example is somewhat strawlike itself.

[Inquartata]

Originally Posted by Sciurus-Rex

Firstly, you keep using "straw man argument," but it doesn't look like you fully understand the term: A straw man argument is a logical fallacy based on misrepresentation of an opponent's position.

And that's what it is. He's arguing as though my position were "A numerical point system offers no real benefit over a system seeding all As randomly". When of course my real position is that a numerical point system offers no real benefit over the system we actually use now".

And now you might as well go teach yer granny to suck eggs.

[Sciurus-Rex]

And you're full of baloney. But I used a smiley, so that's OK.

[telkanuru]

Actually it seems you use "straw man" to mean "I don't agree with you". Woo.


Suprisingly enough, not agreeing with Inq (argumentum contra inquartata) is not a codified fallacy.

[Inquartata]

Well, an understanding of logic has never been either widespread or popular. I guess I can't expect you two to be any different. I lift my palms and shrug.

[Insipiens]

In case you are interested in a non-US perspective, I fence competitively regularly in the UK. We clearly have fewer competitive fencers than the US, and you can get from one end of the country to the other more easily (up to a point).

We have a numeric points system for each weapon and each gender. Points are awarded depending on the strength of a competiton (calculated by reference to the number of fencers in the top 10, top 20 and top 50 of the rankings at the start and half way through the season, plus good foreigners or if there are not enough strong fencers there the total number of entrants) and your place in the competition (with a staggered multiplier to take account of the L32 vs L16 afect). You only get points up to 64th place or better if there are few competitors (something like top 75% but I cannot remember).

You can also get points from international competitions based on some even more arcanely complicated system.

You can count points in the system from a maximum of 6 competitions of which a certain number (possibly 3) can be international competitions. If you do more than 6 competitions, you automatically get ranked on the six comps that give you the highest points. The national championships have an additional multiplier to the points.

Only open competitions count. The rankings are updated monthly and points last until the end of the same month as the competition in the following year; i.e. the points from a competition in September 2005 will drop off when the rankings are calculated at the end of Sept 06/

We use the national rankings for seeding, ranking the strength of competitions and selecting fencers eligible to do FIE competitions.

Many fencers also use them as goal-setting and achievement measuring: I have made the top 100 or the top 50 or I want to ensure my end of season ranking is higher than last year. Many people like to be able to say "I am the x ranked fencer in the country" to their non-fencing friends and family.

Problems with the system include:
*if the top fencers do not turn up success in a competition makes not much impact on the rankings (but they amended the weighting of comps based on the number of competitors recently so it makes more) - arguably this is fairer because without the top ranked fencers the comp is easier.
*fencers who take a year out and come back still being very good have no points so no rnaking and mess up the seeding.
there are doubtless more.

[parrythis]

Good news!!! This week - the week of August 21st - I received my USFA magazine with the cover that informs me that my membership is due to expire the end of July 2006.

[ChubbyHubby]

See this thread: http://fencingforum.com/forum/showthread.php?t=7745

on the UK forum on similar issues with ranking systems / databases etc.