USFA membership/website/rating system update

[Sciurus-Rex]

Originally Posted by MyrddinsPrecint

... Taking power away from the people...

And once again the true issue rears its ugly political head: the metaphor that some illegitimate, faceless Authority is attempting to control the individual's destiny or otherwise control a vaguely sensed resource.

Seriously, how do you define "power" in this particular situation, and how is it being taken away?

[Mr Epee]

Originally Posted by Sciurus-Rex

Seriously, how do you define "power" in this particular situation, and how is it being taken away?

Knowledge is power - Francis Bacon

If the point formula was secret, then it would be very hard for competitors to "game the system" ... but the side effects are the possibility of abuse, and a loss of trust when things seemed fishy... even if the formula was working correctly.

[Peach]

My problem with any secret formula is that, while I'm not particularly paranoid about the abuse of power, I have tremendous faith in the general incompetence of the human race and the ability of something--anything--unanticipated to go wrong. When something is covert, it's much harder to fix.

[Sciurus-Rex]

Honestly, I don't think the formula should be kept a secret either. Full disclosure and transparency all the way, baby!

I just hate to see alarmist, knee-jerk cliches tossed off to rile up the villagers with their torches and pitchforks.

[KD5MDK]

For that matter, it would only be a secret formula that prevented gaming as long as no one outside the National Office knew how it worked. As soon as one high level coach figured it out, either by renting a statistics major, or by having dinner with the right person, they suddenly have an unfair advantage over all other coaches.

[telkanuru]

And all I can think of is "So not only do I have a Secret Plan to fight inflation, but now you're not in favour of it?"

The problem I can see with any non-rolling points system is that it wouldn't degrade via inactivity. The problem I see with rolling points (and the current NRPS) is that it favours those with the time and money to fly to many competitions.

[Fechter1]

Originally Posted by telkanuru

The problem I can see with any non-rolling points system is that it wouldn't degrade via inactivity. The problem I see with rolling points (and the current NRPS) is that it favours those with the time and money to fly to many competitions.

As a possble solution to that, I would like to reference the German points system (go figure...):
At the end of the season, point standings are divided by 10 (i.e. you lose 90% of your points), and you earn points as before.

Something that could work in our case (just spitballing here), might be to assign tournaments a possible point scale based on the strenght of the attending fencers, and every now and then (once a month, or once at the end of the season, or ... or ... etc) you lose a percentage of your accumulated points.

Of course, anything like this would mean a lot more computing to be done for each tournament, and since there's an inevitable delay between getting the point changes to the national office, we could see a lot of erroneous postings. Unless, of course, all the number crunching is done by the national office (for all sanctioned tournaments, every week, etc), which is going to increase their workload by an order of magnitude, I would imagine...

It'd be interesting to see just what they finally come up with...

[oiuyt]

Originally Posted by telkanuru

The problem I can see with any non-rolling points system is that it wouldn't degrade via inactivity. The problem I see with rolling points (and the current NRPS) is that it favours those with the time and money to fly to many competitions.

Non-rolling systems COULD include time-degradation if you're willing to allow lots of complexity. Have a exponential or hyperbolic decay of the impact that results have on an athlete's score.

With re: favoring fencers who make it to the competitions... uhm, yeah. It's next to impossible to come up with a system that can fairly rank those who don't compete (or who compete minimally). Depending on how the system is structured it could count the top N scores (similar to the current NRPS), which helps remove most of the benefit of flying around the country looking for competitions every weekend.

It might be worthwhile to check out how other countries/sports handle numeric systems.

-B

[oiuyt]

Originally Posted by Fechter1

Of course, anything like this would mean a lot more computing to be done for each tournament, and since there's an inevitable delay between getting the point changes to the national office, we could see a lot of erroneous postings. Unless, of course, all the number crunching is done by the national office (for all sanctioned tournaments, every week, etc), which is going to increase their workload by an order of magnitude, I would imagine...

A lot more computing but a lot less work for the NO. Once a structure like this comes into play the computations almost have to move to an automated system. Various oversight and spot-checking controls would need to be put into place, but entering results into the system can be farmed out to the tournament organizers, and the computations don't require any human intervention.

-B

[seak]

To respond to Allen, I am nothing more then a coaches college attendee I think I was fairly open about that.

That being said, I asked him about the rating system (as I had heard rumors on here) and the way he answered the question it seemed to me (and others in the room) that a numerical system in the future is a done deal, but that they do not know how the system will be worked yet. or what form and shape it will take.

I could of course be wrong, and this being the USFA just becuase they say something is going to happen doesn't mean it will

[keropie]

Yeah, it could also just be that they wanted the field in the database just in case, so they didn't have to add it later (though generally speaking adding a field is trivial, especially compared to populating it with meaningful data).

[arc]

When I see the proposals for numerical point systems, I have strongly mixed feelings, and while I'm not sure that these are even intelligent comments, I'll toss them out there anyhow:

First, I have great trepidation about any system that relies on the tournament organizers to enter results & updates directly into the national database with QC occurring after the fact. For most tournaments I've been to, I wouldn't have a problem, but there are some I've become aware of where this could simply be asking for abuse ...

Second, in any large organization, I'm a big fan of the KISS principle -- and a numerical system with formulae to determine relative ranking movements doesn't seem to mesh with it ... while the current system can be occasionally misleading, inasmuch as not all fencers of the same ranking are the same quality (weak vs. strong tournaments, etc.), and the criteria for earning a particular rating are a bit arbitrary, there's a certain charm in being able to post that one-page sheet which lays out 'if the tournament ends thus, and you place thus, then you earn the following rating' without having to field questions from every fencer as to how the results will change their ranking.

That's not to mention the potential costs in having to send out new membership cards to everyone who fencing in a tournament and shows movement in their rankings, or alternately, the potential nightmares in logistics of trying to deal with those five people who didn't preregister and aren't from your division who show up at the door of a tournament expecting to get seeded with the points they earned the previous week at an A4-equivalent tournament ... *shudder* I just see nightmares with trying to administer this at the local level ...

[telkanuru]

None of the numerical proposals are difficult if you have a national points database online. You simply make a form where results are entered online and the program auto-calculates points to a live-update serchable web page, from which you pull point listings for use during check-in. That's not the problem.

I do like the idea of a chess-like rating system where one loses points at a fixed rate as time elapses without activity.

[AAAI]

.....

[oiuyt]

Originally Posted by telkanuru

I do like the idea of a chess-like rating system where one loses points at a fixed rate as time elapses without activity.

Mmmm, sorting out a reasonable way of doing time-decay on an ELO-style system other than just losing points periodically seems non-trivial. With the zero-sum nature of the points calculations it seems that just losing points isn't a particularly good solution either (although if enough newbies are joining the system with some base amount of points to offset both the constant drain and losses due to points of retiring fencers the system could still remain functional).

-B

[Feltan]

Originally Posted by oiuyt

Mmmm, sorting out a reasonable way of doing time-decay on an ELO-style system other than just losing points periodically seems non-trivial. With the zero-sum nature of the points calculations it seems that just losing points isn't a particularly good solution either (although if enough newbies are joining the system with some base amount of points to offset both the constant drain and losses due to points of retiring fencers the system could still remain functional).

-B

The time decay function (which I think is a good idea) can be implemented and tested six ways to Sunday before they go "live" with a system. However, that requires vision and some planning -- they had better be doing it now if this is getting rolled out anytime in the next year or so.

The actual equations are trivial once you figure out what you want to do. The annual (post-Nationals) decay seems most reasonable, and is a freshman computer science homework project to implement.

Regards,
Feltan

[misha]

Originally Posted by telkanuru

None of the numerical proposals are difficult if you have a national points database online.....

Yeah, I can see how it will look like:

USFA will ask someboody they know to build a website
No quality control, no guarantees, no warranty and no support

Website will be down half of the time...

Good luck seeding your next competition by ranking you can not access !



.

[mrbiggs]

No matter what, the numerical system will probably be very far off, because first we need to have a working, dependable website, then a working, dependable, accessable fencing database (actually, we have one, but the USFA isn't using it), then we need to get the actual system tested and through the beaurocracy.

[telkanuru]

Originally Posted by misha

Yeah, I can see how it will look like:

USFA will ask someboody they know to build a website
No quality control, no guarantees, no warranty and no support

Website will be down half of the time...

Good luck seeding your next competition by ranking you can not access !

Note my qualifier "If you have a nat'l points database online"

[JEC]

Originally Posted by Peach

My problem with any secret formula is that, while I'm not particularly paranoid about the abuse of power, I have tremendous faith in the general incompetence of the human race and the ability of something--anything--unanticipated to go wrong. When something is covert, it's much harder to fix.

If you take into account the USFA's track record on keeping and updating the USFA website, proposing a far-reaching numerical system with weekly changes due to a multitude of regional tournaments does not sound too promising!