CrossResults points: men vs women

[sparky33]

I'm stalking a selection of racers on CrossResults.com, and I noticed that the points average (5 best in 12 months) is significantly higher (i.e. a worse rating) for top elite US women as compared to top elite US men.

For example, Katie Compton has ~220 points (4 wins and 1 3rd place in elite races).

That 220 rating would rank her somewhere among male elite racers who regularly finish a little bit outside the top ten... someone like Dan Chabanov, an superb racer but not a regular podium finisher. Tim Johnson, for comparison, has a 150 rating with a handful of podium finishes. I would have expected Compton to have a rating closer to 150...

I would have thought top elite men and women would have similar ratings. Does anyone know why they don't?
I understand that points ranking only matters inasmuch as it orders racers within the same gender, but I am just curious why the stats are not aligned across genders.

[gavingould]

good question, no idea on answer of how the points system applies across gender.
I think 100 points is considered to be the absolute best you could do (Sven Nys is at 102 - he still doesn't win every race but I'll be damned if he's not extremely consistent)

[ChrisG]

One short answer is the depth of the field. It's not just a matter of podium/placing, but the ranking of everyone who happens to be in that particular race. There is simply a greater quantity of men racing at the elite level.

[old fat man]

The algorithm is not meant to be gender specific. Plenty of elite women line up with the men in the 3 race, or the single speed race. The algorithm is predicting the racer's expected placing if everyone (Sven all the way through to a 10 year old junior) were to be racing in the same field.

[sparky33]

That makes sense. Thanks.

[nooneline]

One short answer is the depth of the field. It's not just a matter of podium/placing, but the ranking of everyone who happens to be in that particular race. There is simply a greater quantity of men racing at the elite level.

ding ding ding. the ranking disparity is the algorithm incorporating field size.

[MarkJ]

The algorithm is based on the current score of each participant in a field, which is the average of the top five of their last ten (in the past year, I think).
The winner gets 85% of the average current score of the top five placers.

Given an infinite field, each of the top five would get 85% of their average, and if they all went in with equal scores they would all come out with a score closer to zero. Therefore, the algorithm can trend toward zero in large fields.
Given a field of five, they would be placed on a line from 85% of their average to their average, and if they all went in with equal scores and placed in an identical order over many races the winner would trend toward zero (but never reach it) while the loser would stay fixed at their initial score. Therefore, with a field of five the algorithm will tend toward zero for some and no change for others.
Given a field of three, they would be placed on a line from 85% of their average, through their average, to 115% of their average. Therefore, in small fields the chance that you gain anything for showing up is less.


So, as nooneline said, it's about the field size.

So, small fields suck. What else?
Well, what happens when nobody in a race has any results from which to calculate their current average?

Looking back to the 2006 data, comparing multiple fields in multiple races, it looks like there is/was a biased range of points for races that cannot be based on prior results. Multiple races showed the most elite mens field distributed from 187 to 253, first to last, B men from 340 to 460, and A women from 374 to 495.

Based on this it would seem that the women were screwed from the get go.

Aside from all this, the algorithm is a competitive ranking system based on relation of riders in the same fields. Riders from no separate gender groups, or even fields for that matter, can be compared in any way unless there have been participants that cross from one group to another. The fewer the points of crossing, the less meaningful the link.

So, after all that rambling, it would appear that points are only good for ranking like quantities (even though whoever setup the initial value distributions had hopes that it would be otherwise).

[gavingould]

i've gamed this system to garner a better starting position. hasn't really helped, i'm still slow.

[BSUdude]

i've gamed this system to garner a better starting position. hasn't really helped, i'm still slow.

The guy who won the 1/2/3 race at hopkins started on the back row.

[gavingould]

The guy who won the 1/2/3 race at hopkins started on the back row.

ha!! yeah, that guy is slightly quicker than i am. or will ever be. he also lapped me a couple times at Norge last year in Brian Matter's wake.

importantly, Maloney started back row at Hopkins because he didn't pre-reg. ChiCrossCup only stages by points those who pre-reg. day-of gets back of the pack.
Barry Wicks has been a victim of that once or twice, it didn't slow him down either.