Percentage of World Records

[JimR]

Why do you choose time as your yardstick when power output in Watts would likely be a better comparision?

Jim Pisano asked this too and I have been meaning to answer.

The primary reason is that time is easier to comprehend and to use.

Were watts to be used, then every time would need to be converted to watts, each of which would be more "time consuming" and also much more complicated than any of the calculations done now.

A second reason is that watts are not obviously connected with time. For example, if you row 230.1 watts for a 5k, no one would know what pace or time you did, unless they did a calculation to find out. However if you averaged 1:55 pace, then many would know that you did 19:10 for the 5k and visa versa.

Time is what is universally used. Cars measure speed in miles or kilometers per hour, not power output. The same for bicycle cyclometers, swimming, skiing, running. All of these show performance in time and not power.

So I have done the PERathlon the same way for these reasons.

So the answer you gave is best summarized as "using watts is too hard and nobody understands watts anyway"? If you want to be a thought leader on this topic then you would take times, convert to watts (to improve the accuracy), do the rankings (in watts) and convert back to time.

By the way ... as an owner of a BAII calculator you can do this math. You could however do it much easier in Excel.

JimR

[johnlvs2run]

JimR,

If you think using time is too easy, then what are your times for the 10 events, and what are your PERathlon scores?

I'd like to honestly see your results and, based on those, your suggestions.

I am open to improvements and refinements.

[johnlvs2run]

[johnlvs2run]

Jim Pisano,

You can convert your PERathon score to a WATTathon score by cubing it.

Let's say you have a PERathlon score of 92.5 percent, or 925 points.

You can find your Wattathon percentage by calculation of the cube of this score.

.9250 cubed = .7910 = 7910 points

[fish]

I hope no one minds if I join this discussion, I am intrigued by this topic of mathematical models. I also hope that I don't offend anyone because it is not my intent.

In theory, as we look at what someone is capable of rowing over a lifetime, we would initially expect improvement until that person reaches a peak and then performance would start to decline. In indoor rowing, there is not a long enough history to produce that sort of curve with the best times. A few exceptional individuals have produced some outstanding times but there are still a few "holes" in some of the age categories. For instance, for Heavyweight Men 500 meters, a 41 year old has a better time than the 40 year old (from Paul's? spreadsheet). In theory, the 40 year old should be faster.

The nonathlon (from my understanding) uses depth to overcome these issues. Yet even with that depth, fitting a polynomial with degree 4 to the data still produces a curve that goes up and down which isn't the type of curve we expect.

I don't know that much about the TI-BAII but a power curve is just a polynomial with one term. While this a power curve does not go up and down, it is symmetrical. In theory, we would expect a young person to improve rapidly, reach a peak, and decend slowly (thankfully) into old age.

I was thinking some kind of logistic curve would model the data better or perhaps a polynomial (maybe a third degree) that is forced to have only one turning point and minimizes the distance bewteen the curve and the data points.

Anyway, these are just thoughts. I think that the nonathlon is great and I didn't intend these comments to criticize it in any way.

Arlene

[PaulH]

Fish - no offense taken at all, you make some very good points. In fact, if I ever get around to it, I'll be updating the spreadsheet that's available online because I agreed with your reasoning, and switched to using a simple 2nd order polynomial a couple of years ago. It loses some theoretical precision, but smooths out the inflections you noted.

One thing I would quibble over is the idea of a smooth decline - it seems entirely possible to me that we might see plateaus in the decline (and even more so in the initial ramp-up). I don't yet have a specific model for this, but I'm thinking for example, that between 40 and 45 the average might decline very little, that the pace would then pick up until perhaps 70, and then slow down again once people approach the point where they've lost as much power as they're going to! Time and the accumulation of data will hopefully show if something like this does happen.

Cheers, Paul

[Porkchop]

One thing I would quibble over is the idea of a smooth decline - it seems entirely possible to me that we might see plateaus in the decline (and even more so in the initial ramp-up). I don't yet have a specific model for this, but I'm thinking for example, that between 40 and 45 the average might decline very little, that the pace would then pick up until perhaps 70, and then slow down again once people approach the point where they've lost as much power as they're going to! Time and the accumulation of data will hopefully show if something like this does happen.

Cheers, Paul

Not to be morbid, or anything, Paul, but "the point where they've lost as much power as they're going to" is when they're dead, isn't it?

Interesting questions. Let me throw in another variable that is likely to have an effect -- injury. On "use it or lose it" approach to performance, the guy who exercises regularly and pushes himself to his limits should see only a gradual decline in ability. But even that guy is susceptible to injury, and as one ages, the ability to recover from injury is compromised. The loss of conditioning during recovery and rehabilitation of an injury will be all the harder to make up as the athlete ages, but the loss of performance would not be attributable to aging per se.

[PaulH]

Porkchop,

I almost qualified my remark with "as much power as they're going to lose and still make it to the erg"

The point you make is true for an individual, but I think there's an assumption that the 'average' athlete doesn't suffer from this, because we're not really talking about plain averages, but the average of the best athletes in the world. So the fastest 60 year old of the next 200 years will presumably not have been slowed overly by injury, even though the average 60 year old may well have been.

Cheers, Paul

[fish]

I wonder if the plateau that occurs at age 40 is actually a renewed commitment for athletes entering a new age group. Let's face it, it's hard to keep up with those 30 year olds when your 39.

If an athlete could be at an optimal fitness level, year after year, with no injuries or illness, and could have perfect conditions for that one good race, I would guess that the decline would be smooth. Of course, real life isn't like that. Especially as we age, illness, injury and motivation dictates how we perform and for an individual, the curve becomes jagged. But for a large group, those ups and downs would smooth out the curve (except for those 39 year olds who waiting to turn 40).

I didn't mean to go on and on. Some kind of logistic like curve would have a horizontal asymptote to deal with the issue of reaching the ultimate decline. And perhaps this is great in theory but impractical for the Nonathlon's purpose.

Paul, have you considered a piecewise function to generate scores?

Arlene

[PaulH]

Regarding the 40 year olds, I wasn't suggesting that as a specific age, but just an example. So I wouldn't expect the percentage drop between 42-47 to be the same as for 52-57, for example, but I also wouldn't expect the rate of change of that decline to be the same. I think that there are times in life where fitness stays relatively constant, and times where it declines. Many of these are circumstantial, such as retirement giving us more time to train, but I suspect that some is physical, i.e. we don't just smoothly decline, but we decline relatively in fits and starts. For men this is pure hypothesis, but for women menopause would clearly be a possible break point between rates of change.

Nonathlon - I did consider a number of ways of calculating scores, and had several more suggested (including, I think, your idea). In the end I went with quadratic curves because it is, relatively, easy to understand - it's high school math, but no more.

[johnlvs2run]

I don't know that much about the TI-BAII but a power curve is just a polynomial with one term. While this a power curve does not go up and down, it is symmetrical. In theory, we would expect a young person to improve rapidly, reach a peak, and decend slowly (thankfully) into old age.

I was thinking some kind of logistic curve would model the data better or perhaps a polynomial (maybe a third degree) that is forced to have only one turning point and minimizes the distance bewteen the curve and the data points.

Hi Arlene,

The power regression analysis (learning curve) is a method for analyzing two-variable statistics. Other regression analysis methods are linear, logarithmic, and exponential. I have been using the power curve for quite a few decades, originally with running, and find it quite useful. I have tried the other three methods from time to time but they have not worked as well for my purposes.

The formula is: Y = a*X^b

[JimR]

JimR,

I noticed you don't have your times in your signature file.

If you think using time is too easy, then what are your times for the 10 events, and what are your PERathlon scores?

Let's calculate them out, and see how easy they are for you to accomplish.

John -

I have no need to rank my times, calculate my PAT, PATT, PERathlon scores or maintain my times in a signature file.

I would be one of the odd people on this forum who actually erg, look to improve by times, hope to get info to acheive my personal goals and, if it would help someone else, share what others have taught me for the greater good.

Every bit as odd, I find your postings on this forum amusing at best but never useful to any of the ideas listed above.

I am however very greatful that you participate as I have learned a lot as others provide valuable information in the process of trying to get you to understand how little you actually know about erging and how utterly stupid some of the things you post really are.

JimR

[johnlvs2run]

[Citroen]

I am however very greatful that you participate as I have learned a lot as others provide valuable information in the process of trying to get you to understand how little you actually know about erging and how utterly stupid some of the things you post really are.

JimR

ROTFLMAO. That's a diamond. Excellent comment!

[JimR]

I would be one of the odd people on this forum who actually erg, look to improve by times, hope to get info to acheive my personal goals and, if it would help someone else, share what others have taught me for the greater good.

Jim,

Good luck to you with that, since you don't know what your times are, don't know what you're doing, and don't care.

If you ever get some times and figure out your PERathlon scores, let me know.

... and John, if you ever decide you want to improve your erg times let everyone on the forum know ... I'm sure they would be willing to share good information to acheive your goals, even to you ...

A starting point would be when you free your mind and open up to the methods that actually have a proven track record of results. I'm one who thinks that not all new ideas are good ones ... but hopefully some are. As an optimist I hope that one day you will have a new idea that is also a good idea.

Your times at bat are greater than any ... if only you could hit the ball!

JimR