Simple (simplistic?) counter to rating inflation

[JMcC]

There are too many "ratings" threads to search for this (I fear it would take hours and hours,) so excuse me if this has been proposed previously.

After years of observation, I've seen that:

1. There is excessive rating inflation; disproportionate with the actual skill levels demonstrated in major national competitions.
Example: this season's NAC B Junior epee (I use this because "A" through "U" can enter)
244 entrants
69 "A"s (28%)
10 "A"s finishing in the bottom 50% (or 15% of all "A"s entered)

2. It is dramatically easier to earn a higher rating at a small, local competition than at a larger sectional or national competition.

Given the above, why not:

1. Shift the USFA Classification chart up one level -
Group A4 to use what's currently awarded for A3 (1 through 4 earns A, 5 through 8 earn B), A3 then awards as current A2, and so on down the Groups, eliminating the E1 Group.

2. Drop an earned rating to the next lower if it's not re-earned within three years (or maybe just two years.) Is a recent "B" really demonstrating much skill by defeating an "A" rating that hasn't been renewed in over 3 1/2 years?

[KD5MDK]

Because among a great many other things, we're not suffering from an excess of Es. Or even Ds.

[Fechter1]

Originally Posted by JMcC

2. It is dramatically easier to earn a higher rating at a small, local competition than at a larger sectional or national competition.

Actually, I've seen more A's given out in the larger sectional events... A tournament that barely meets the reaquirements for an A4 seems to be the easiest way to earn a rating here on the east coast...
You can check Askfred on that... We generally have 4-5 A4 rated tournaments, and from those, you can see usually 5-10 As generated from those... And the A1 and A2 tournaments tend to be on the higher end...

[tdwg83]

Originally Posted by JMcC

2. Drop an earned rating to the next lower if it's not re-earned within three years (or maybe just two years.) Is a recent "B" really demonstrating much skill by defeating an "A" rating that hasn't been renewed in over 3 1/2 years?

If the fencer really isn't good enough to renew after 3 and 1/2 years. Is six more months going to make that much distinction before it officially drops?

[KD5MDK]

Not to mention the cries of outrage when someone places 5th at a NAC to renew their B, unless you plan to introduce an A5 category.

[mrbiggs]

That would be only a temporary fix for a few years until the ratings inflate again.

[tdwg83]

Then what?

[mrbiggs]

Originally Posted by tdwg83

Then what?

That's my point. I think that while we're changing it, we should fix it rather than putting off the problem. By fix it I mean implement a zero-sum rating system that would avoid inflation. I personally support the Elo system because it's been very effective in other 2 player games, but I'm sure there are other good systems out there.

[noodle]

Originally Posted by tdwg83

Then what?

Fencing ratings

[Tomas N]

I don't see the trends that are listed in the OP as sufficient rationale for rejiggering ratings. People have bad days, particularly young people. Folks get injured but still decide to fence at a national competition. People's training regimes change due to work/school/life. There are lots of good reasons why an A might end up in the bottom 50%. And the fact that they do indicates that ratings are a good way to seed tournaments but don't solely determine where folks end up. Isn't it good that we actually have to fence the bouts to figure out who wins?

When I look at tournament results, I am always surprised by how well final placing lines up with rating, not the opposite.

For those who want a measure of how well the very top fencers are doing, look at the points lists.

Tomas

[TBean]

Originally Posted by JMcC

Example: this season's NAC B Junior epee (I use this because "A" through "U" can enter)
244 entrants
69 "A"s (28%)
10 "A"s finishing in the bottom 50% (or 15% of all "A"s entered)

While I agree that rating inflations do exist and many individuals who have an "A" might indeed not be that quality of athlete, this particular data does not support your case. This is what you should expect to see.

Anyone can have a bad day, just because you are an A does not mean that you are properly prepared, you can get an unlucky draw because someone had a bad pool or in the case of juniors be adjusting to college and/or a new coach. If the number of A's in the bottom 50% of the tournament signifcantly exceeded even 20% then it should raise eyebrows but this result is reasonable.

[JMcC]

Originally Posted by TBean

While I agree that rating inflations do exist and many individuals who have an "A" might indeed not be that quality of athlete, this particular data does not support your case. This is what you should expect to see.

Anyone can have a bad day, just because you are an A does not mean that you are properly prepared, you can get an unlucky draw because someone had a bad pool or in the case of juniors be adjusting to college and/or a new coach. If the number of A's in the bottom 50% of the tournament signifcantly exceeded even 20% then it should raise eyebrows but this result is reasonable.

OK, take the reverse view then. Look at the results over the last three years for all the national events in those weapons/genders that most people believe are suffering from rating inflation, and count how many "B"s have earned an "A" in them.

My quick look at Junior Men's foil shows only about 5 USFA fencers have gone from "B" to "A" at the last 9 national events (NACs, J.O.s and Summer Nationals.) That's .56 "A"s earned at each national event.

I'd propose that those 5 "A"s were significantly more difficult to earn than "A"s that have been earned at sectional and, more to the point, divisional competitions. Clearly, as structured now, if you're a "B" aiming for an "A" rating, you're chances are far better at the smaller local events that meet at least the current A1 requirements.
_

[JMcC]

Originally Posted by tdwg83

If the fencer really isn't good enough to renew after 3 and 1/2 years. Is six more months going to make that much distinction before it officially drops?

Not sure if you're saying 4 years should remain the time limit because those 6 months don't make a difference, or you're saying dropping to 3 years wouldn't be enough because of those 6 months.

I suppose an argument could be made for a 2 year limit for renewal. If a fencer didn't enter many competitions during those 2 years his/her competitive skills would most likely deteriorate; and on the other hand if the fencer did compete actively and couldn't renew in 2 years, their skills clearly had deteriorated.

[veeco]

Originally Posted by Fechter1

Actually, I've seen more A's given out in the larger sectional events... A tournament that barely meets the reaquirements for an A4 seems to be the easiest way to earn a rating here on the east coast...
You can check Askfred on that... We generally have 4-5 A4 rated tournaments, and from those, you can see usually 5-10 As generated from those... And the A1 and A2 tournaments tend to be on the higher end...

Seconded. The Pacific Coast Championships in Men's Epee comes to mind.

[wahrman]

While that certainly can be true some places. I know that it is false in at least one. While I've not scoured the results to back up my theory, in the 1.5 years I fenced at Bay Cup events around San Francisco I do not recall seeing any As being awarded. And only two Bs that I can think of. At least
in Men's epee.

Those tend to be pretty tough A1s and A2s.

EDIT: On second thought, I do remember someone renewing their A last year.

Originally Posted by JMcC

I'd propose that those 5 "A"s were significantly more difficult to earn than "A"s that have been earned at sectional and, more to the point, divisional competitions. Clearly, as structured now, if you're a "B" aiming for an "A" rating, you're chances are far better at the smaller local events that meet at least the current A1 requirements.
_

[edew]

Originally Posted by mrbiggs

That's my point. I think that while we're changing it, we should fix it rather than putting off the problem. By fix it I mean implement a zero-sum rating system that would avoid inflation. I personally support the Elo system because it's been very effective in other 2 player games, but I'm sure there are other good systems out there.

No zero-sum rating system.

I really don't care what system to use, as long as it more-or-less correlates with the fencer's skill, gives higher ratings to those who do better (i.e., there should be no incentive to try for 8th place when 7th or higher is better), easy to implement, and gives good marketing bang for the buck.

[Shi no Tenshi]

I almost hate to ask this...but...why does it matter? It's just a letter.....

[the ancient one]

After considering NAC points, it may be better to seed fencers based on their worst of N most recent results within a competition level category, rather than their best result in any category over the past 4 years.

A rating set may look something like this:
A4-27
A2-U in this example, no prior history so they use 17 ...
A1-3
etc...

They can age out at some rate like 16 places/year and drop a letter when they place-rank below 64.

[C.I.C.]

Originally Posted by Shi no Tenshi

I almost hate to ask this...but...why does it matter? It's just a letter.....

Pool seeding. At higher competitive levels, generally speaking, being the highest rated fencer in your pool often improves your chances for better results out of the pools.

If there's only one A available to seed into each pool, and you're it, that can be to your advantage.

[HDG]

I'd like to speak for a moment on behalf of my fellow scrubs: those who will probably never make it out of D2 & D3 (or even local) competitions, for whom rating inflation and fine-grainedness of rating schemes are not issues, and for whom letter ratings as something tangible are important.

I know that as any useful indicators of anything, our current ratings are meaningless. I know as well that they really exist only for purposes of event seeding, and that they don't even do that well. I believe, however, that the concerns of this lower tier of competitors, who value and take pride (rightly, wrongly or ridiculously) in their E's, D's and C's, ought to be taken into consideration as well.

Others made this point in the previous ratings related threads; I just felt it needed reiteration.