Compare rankings

[sjors]

Your website is amazing and I'm all for more stats. Surely you need to have the 2k in there somewhere though

Thanks, the queries are not working in the moment. I will have a look

[sjors]

Your website is amazing and I'm all for more stats. Surely you need to have the 2k in there somewhere though

Thanks, the queries are not working in the moment. I will have a look

They are working again.

[boomingaway]

They are working again.

I meant that you hadn't included the 2k in any of your categories (all round, sprint, endurance)

[sjors]

They are working again.

I meant that you hadn't included the 2k in any of your categories (all round, sprint, endurance)

O, sorry misunderstood.
Since I like to have only 4 different events and I like somewhat of logic, I could also opt for:
All-round: 500m, 2k, 10k and HM.
In stead of 4 time-trials 4 distance trials.
Would that be better?

[Tandstad]

I don't know how many times I have done a PB, and spent a good 20 minutes on my phone to see where I compare against times done previous years in my country And by accident, I fell into this thread and see that you have created the perfect lookup-tool for this! Many thanks Siebe, much appreciated.

Brg Jo the stats whore.

[sjors]

I don't know how many times I have done a PB, and spent a good 20 minutes on my phone to see where I compare against times done previous years in my country And by accident, I fell into this thread and see that you have created the perfect lookup-tool for this! Many thanks Siebe, much appreciated.

Brg Jo the stats whore.

Your Welcome.
More to come

[MudSweatAndYears]

The following competitions I could imagine for rowers and skiergs
Sprints: 100m, 500m, 1000m and 1 minute
All-round: 1, 4, 30 and 60 minutes
Endurance: 10k, hm, fm and 100k

I meant that you hadn't included the 2k in any of your categories (all round, sprint, endurance)

O, sorry misunderstood.
Since I like to have only 4 different events and I like somewhat of logic, I could also opt for:
All-round: 500m, 2k, 10k and HM.
In stead of 4 time-trials 4 distance trials.
Would that be better?

It would be very stimulating to have such aggregate scores neatly available in your smart ranking tool!

I like the idea of building an all-round category based on fixed-time events, but probably a distance-based all-round category would be more popular. I'd suggest for rowing to make it as close as possible to the Dutch skating all-round tournaments: 500m, 2k, 5k and 10k.

[sjors]

I'd suggest for rowing to make it as close as possible to the Dutch skating all-round tournaments: 500m, 2k, 5k and 10k.

That makes sense. The distances are more popular.
I was trying to get closer in duration. (i.e. WR 10k Speedskating is round 12:30. WR 10K Rowing around 31:00)

[MudSweatAndYears]

How would you calculate the aggregate score for each category? Would that simply be the arithmetic average of the four splits for each of the sub-events?

[sjors]

How would you calculate the aggregate score for each category? Would that simply be the arithmetic average of the four splits for each of the sub-events?

First every effort will be weighted for age. I'm busy finding the right formula per event for calculating this factor to calculate the a age-independend-score. This factor is not only based on already realised times per age, but based on a combination of
1. studies how age effects performance and
2. the concept2 WR's per age-group.
In this way I try to overcome the issue of a weighted performance that is (too) high because of slightly underperformance in one or more years.
I.e. you see a steep decline in performance in the age-group 40-49 that does not meet the known studies about age related decline. The same applies for some of the age-independent WR.

Some examples:

Event    F/M    H/L    Age-group    Effort    Watt-% agains WR    Decline per year
1:00        F.      L.        40-49         319             85.6%                         1.2%
1:00        F.      L.        50-59         320             86.4%                       - 0.08%

10k        M.      H.        40-49       32:49.8         84.9%                         1.51%
10k        M.      H.        50-59       33:17.1         81.5%                         0.34%


Every event will then have an age-weighted-score (in points I suppose) and then for each competition the sum of these scores is what the ranking will use.
Does this make sense? (English is not my native language)

[MudSweatAndYears]

I'd suggest for rowing to make it as close as possible to the Dutch skating all-round tournaments: 500m, 2k, 5k and 10k.

That makes sense. The distances are more popular.
I was trying to get closer in duration. (i.e. WR 10k Speedskating is round 12:30. WR 10K Rowing around 31:00)

Yes, staying closer in duration to speedskating would be a reasonable way forward. Sticking to distance-only events, this would translate into an all-round rowing competition comprising 500m, 1k, 2k and 5k.

[MudSweatAndYears]

How would you calculate the aggregate score for each category? Would that simply be the arithmetic average of the four splits for each of the sub-events?

First every effort will be weighted for age. I'm busy finding the right formula per event for calculating this factor to calculate the a age-independend-score. This factor is not only based on already realised times per age, but based on a combination of
1. studies how age effects performance and
2. the concept2 WR's per age-group.
In this way I try to overcome the issue of a weighted performance that is (too) high because of slightly underperformance in one or more years.
I.e. you see a steep decline in performance in the age-group 40-49 that does not meet the known studies about age related decline. The same applies for some of the age-independent WR.

Some examples:

Event    F/M    H/L    Age-group    Effort    Watt-% agains WR    Decline per year
1:00        F.      L.        40-49         319             85.6%                         1.2%
1:00        F.      L.        50-59         320             86.4%                       - 0.08%

10k        M.      H.        40-49       32:49.8         84.9%                         1.51%
10k        M.      H.        50-59       33:17.1         81.5%                         0.34%


Every event will then have an age-weighted-score (in points I suppose) and then for each competition the sum of these scores is what the ranking will use.
Does this make sense? (English is not my native language)

I see. Yes, first deriving a score that compensates for age and then aggregating the scores for all events within the all-round category makes sense.

With your tool you can slice and dice the data without limit, and perhaps you can indeed do better than existing age-correction methods. On the other hand, I don't think there is a perfect method, as each method needs to find a compromise between transparancy and fairness. For instance, the anomaly you highlight between the age groups 40-49 and 50-59 is explainable given that a fair fraction of the 50+ population can make more time available for training (children out of the house, careers plateauing,etc). Such effects make the creation of a level playing field across ages rather subjective.

So wouldn't focus too much on fairness, and rather optimise for transparancy. I think with the C2 rowing demographics being biased towards the 40+ age group, a simple and transparant system that allows for competition across wide age brackets would probably be an attractive option to many. One simple idea is for each individual in a certain event (distance, male/female, HW/LW) to show in the rankingtool 1) how many younger individuals he/she is beating, and 2) how many older individuals beat him/her. And of course one should than also have the option to create ranked llist based on these scores.

[sjors]

I see. Yes, first deriving a score that compensates for age and then aggregating the scores for all events within the all-round category makes sense.

With your tool you can slice and dice the data without limit, and perhaps you can indeed do better than existing age-correction methods. On the other hand, I don't think there is a perfect method, as each method needs to find a compromise between transparancy and fairness. For instance, the anomaly you highlight between the age groups 40-49 and 50-59 is explainable given that a fair fraction of the 50+ population can make more time available for training (children out of the house, careers plateauing,etc). Such effects make the creation of a level playing field across ages rather subjective.

So wouldn't focus too much on fairness, and rather optimise for transparancy. I think with the C2 rowing demographics being biased towards the 40+ age group, a simple and transparant system that allows for competition across wide age brackets would probably be an attractive option to many. One simple idea is for each individual in a certain event (distance, male/female, HW/LW) to show in the rankingtool 1) how many younger individuals he/she is beating, and 2) how many older individuals beat him/her. And of course one should than also have the option to create ranked llist based on these scores.

Transparency is good point. I will bare that in mind.

Another theory about the 40-49 group: it is the group after the group that professionally rowed. So the amount of hours that that group is doing is incomparable.

Some other interesting points you made, made me thinking: I.e. what about making age groups of 5 years. And just compare within that groups without compensating for age. Makes the whole thing very transparent and easy to code.

Your last suggestion I don’t follow. Can you explain a bit more?

[Dino]

The site is looking great now

As for the multi-event scoring... its a difficult one. I'd be inclined just to stick to the current C2 age groups for now and not try to age weight it - at least to start with. I am sure you have seen how the C2 Nonathlon is scored to allow everyone to compete together?

Something else that might of interest is the way the Decathlon events in athletics are converted into points. Of particular note is how you get more points the closer you get to the threshold. So in rowing terms for example, a 5 sec decrease in your 2k e.g. 7:10 down to 7:05 you would get for example 100 extra points, but a 5 sec decrease from 6:45 to 6:40 you would get 200 extra points. I am just making these figures up but some similar formula could be used to convert each event into points tables. There are lots of articles online explaining how the points are calculated for the decathlon. Just a thought

[sjors]

The site is looking great now

As for the multi-event scoring... its a difficult one. I'd be inclined just to stick to the current C2 age groups for now and not try to age weight it - at least to start with. I am sure you have seen how the C2 Nonathlon is scored to allow everyone to compete together?

Something else that might of interest is the way the Decathlon events in athletics are converted into points. Of particular note is how you get more points the closer you get to the threshold. So in rowing terms for example, a 5 sec decrease in your 2k e.g. 7:10 down to 7:05 you would get for example 100 extra points, but a 5 sec decrease from 6:45 to 6:40 you would get 200 extra points. I am just making these figures up but some similar formula could be used to convert each event into points tables. There are lots of articles online explaining how the points are calculated for the decathlon. Just a thought

Thanks Dino.
I have taken noticed of how Nonathlon works. I assume they calculate with watts as well. And maybe also therefore as faster you get the more each second increases your points???
The problem I am/was? trying to overcome is the fact that when there are relatively less performances in a specific class the higher your score tends to be. It is seldom you see a Heavy weight younger than 60 in the top 10 for years. (Probably some little personal frustration is coming up here. )