Swimming World's top 12 master swimmers!!!!!!!!!!!!!

I know the older age groups are tougher and it looks like you put in a lot of work for the calculator -- but I just don't think it works even for our age group. Here is a look at all the 200 distances in our age group. Your record rating is compared to FINA points. 200 Free -- 1:55.67 = 750 points 200 *** -- 2:25.71 = 691 points 200 Back -- 2:10.88 = 683 points 200 Fly -- 2:07.17 = 717 points 200 IM -- 2:12.20 = 697 points The 200 Free and 200 Back for example are just not equal --- about a 2:07 in the 200 Back should be the same as 1:55.6. Erik, you didn't give quite enough information to tell (course, age), and I'm too lazy to mess around to figure it out...but I'm going to assume that all those times have identical "masters" ratings on the VA LMSC calculator and your objection is that their FINA scores should be much closer. Forgive my long-windedness in the following response; it is an occupational hazard. FINA attempts to compare swims across event and gender; the underlying assumption is that all WRs are equally good. Of course it isn't completely true, but it isn't too bad an assumption. Masters scoring systems, including mine, have to be a little more ambitious: compare swims across event, gender and age. So there will probably be a little more error. My system (and the one that inspired mine, the NEM system) does this by fitting a function to the WR progression for a given gender/event. This tends to average out the fact that some age groups have "softer" WRs than others. I use a robust fitting method, which further means that the fit is resistent to changes if a couple records are extremely good or bad. So this method is going to do a GOOD job of accounting for aging within a given gender-event (and we can even estimate how good a job it does). But to compare across genders and events, just like FINA we have to make an underlying assumption: that the age-dependent "theoretical" WR -- as predicted by the fitted curve -- is equally good for every gender-event combination. Personally I think that is a pretty reasonable assumption, given the averaging effect across age-groups. But if we don't accept that, we can try to make corrections in various ways if we wish; I present one possibility below. (The danger is if the corrections introduce more error than they fix.) To broaden the FINA scoring system, there needs to be some method of accounting for aging. What you seem to be suggesting is that the effect of aging, compared to the "real" WRs, should be the same. In other words, whether it is 1650 free, 400 IM or 50 back, the effect of aging on FINA score should be virtually identical. I would have to verify, but I think this could be easily accomplished with either my or the NEM method: basically just add a conversion factor for the vertical axis to/from the FINA score. But I really question your assumption. Both training and physiology will change pretty much as we age, and I think it stands to reason that they will effect our ability to do certain events in different ways. I know it has for me: I don't remember the FINA score, but based on world rankings and placing at European Championships, my best LCM event as a youngster was the 200 fly, followed closely by the 100 fly. That isn't even close to be true now according to FINA scoring (or even my calculator, for that matter) and it isn't because I don't try. There are other examples for myself -- and I'm sure for others -- but you get the idea. But you have given me an idea that may address your objections a little. The intercept term in the fitting function I use has a physical meaning: it is the theoretical best masters WR time for a given gender/event. Normally I let the date decide what the value is, b/c that gives the best fit. But it would be interesting to force that value to be the current WR and see what happens. The idea would be to improve the inter-event comparisons. The fit for a given gender-event would be worse and it would probably greatly decrease the ratings for the youngest age groups. But maybe that is reasonable, since very often the WRs in the very young age groups are not as good, relatively speaking, as in the older age groups. I'll have more time in the summer to mess around with it on the evenings/weekends and see what happens.

I know the older age groups are tougher and it looks like you put in a lot of work for the calculator -- but I just don't think it works even for our age group. Here is a look at all the 200 distances in our age group. Your record rating is compared to FINA points. 200 Free -- 1:55.67 = 750 points 200 *** -- 2:25.71 = 691 points 200 Back -- 2:10.88 = 683 points 200 Fly -- 2:07.17 = 717 points 200 IM -- 2:12.20 = 697 points The 200 Free and 200 Back for example are just not equal --- about a 2:07 in the 200 Back should be the same as 1:55.6. Erik, you didn't give quite enough information to tell (course, age), and I'm too lazy to mess around to figure it out...but I'm going to assume that all those times have identical "masters" ratings on the VA LMSC calculator and your objection is that their FINA scores should be much closer. Forgive my long-windedness in the following response; it is an occupational hazard. FINA attempts to compare swims across event and gender; the underlying assumption is that all WRs are equally good. Of course it isn't completely true, but it isn't too bad an assumption. Masters scoring systems, including mine, have to be a little more ambitious: compare swims across event, gender and age. So there will probably be a little more error. My system (and the one that inspired mine, the NEM system) does this by fitting a function to the WR progression for a given gender/event. This tends to average out the fact that some age groups have "softer" WRs than others. I use a robust fitting method, which further means that the fit is resistent to changes if a couple records are extremely good or bad. So this method is going to do a GOOD job of accounting for aging within a given gender-event (and we can even estimate how good a job it does). But to compare across genders and events, just like FINA we have to make an underlying assumption: that the age-dependent "theoretical" WR -- as predicted by the fitted curve -- is equally good for every gender-event combination. Personally I think that is a pretty reasonable assumption, given the averaging effect across age-groups. But if we don't accept that, we can try to make corrections in various ways if we wish; I present one possibility below. (The danger is if the corrections introduce more error than they fix.) To broaden the FINA scoring system, there needs to be some method of accounting for aging. What you seem to be suggesting is that the effect of aging, compared to the "real" WRs, should be the same. In other words, whether it is 1650 free, 400 IM or 50 back, the effect of aging on FINA score should be virtually identical. I would have to verify, but I think this could be easily accomplished with either my or the NEM method: basically just add a conversion factor for the vertical axis to/from the FINA score. But I really question your assumption. Both training and physiology will change pretty much as we age, and I think it stands to reason that they will effect our ability to do certain events in different ways. I know it has for me: I don't remember the FINA score, but based on world rankings and placing at European Championships, my best LCM event as a youngster was the 200 fly, followed closely by the 100 fly. That isn't even close to be true now according to FINA scoring (or even my calculator, for that matter) and it isn't because I don't try. There are other examples for myself -- and I'm sure for others -- but you get the idea. But you have given me an idea that may address your objections a little. The intercept term in the fitting function I use has a physical meaning: it is the theoretical best masters WR time for a given gender/event. Normally I let the date decide what the value is, b/c that gives the best fit. But it would be interesting to force that value to be the current WR and see what happens. The idea would be to improve the inter-event comparisons. The fit for a given gender-event would be worse and it would probably greatly decrease the ratings for the youngest age groups. But maybe that is reasonable, since very often the WRs in the very young age groups are not as good, relatively speaking, as in the older age groups. I'll have more time in the summer to mess around with it on the evenings/weekends and see what happens.