How To Calculate Your Patt Percentages

[old] PaulS · Post by **[old] PaulS** » August 28th, 2005, 1:33 pm

<table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(John Rupp @ Aug 28 2005, 08:27 AM)</b></div></td></tr><tr><td class='quote'>PATT is based on single age performances, across age, weight, and gender.<br /><br />Check the chart at the beginning of this thread, and it will show the smooth curve from one age to the next. <br /> </td></tr></table><br /><br />I've looked at that list, but it has no people associated with the times, this makes your claims relatively confusing.<br /><br />Paces for age 20 to 32 are the same (and the "fastest"), and 18 and 33 are the same?<br />When in fact the 5:37 was done by a 35 year old Matthias Siejkowski.<br /><br />So PATT is based on the estimation of some possible World Record, or actual World Records?<br /><br />All in all, with the exponential nature of the power required to achieve any given split, using a straight percentage is rather meaningless. I.e. Much like comparing Bill Gates Hourly income to the rest of the People, there are just different powers at work.

[old] John Rupp · Post by **[old] John Rupp** » August 28th, 2005, 3:49 pm

The PATT times have been developed from existing World Records and world best times.<br /><br />They are intended for those who are interested in improvement, and for comparing legitimate times across age, weight class, and gender.<br /><br />If you are not interested in improvement or legitimate times, then probably the PATT percentages would not have meaning for you.

[old] PaulS · Post by **[old] PaulS** » August 28th, 2005, 6:33 pm

<table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(John Rupp @ Aug 28 2005, 12:49 PM)</b></div></td></tr><tr><td class='quote'>The PATT times have been developed from existing World Records and world best times.<br /><br />They are intended for those who are interested in improvement, and for comparing legitimate times across age, weight class, and gender.<br /><br />If you are not interested in improvement or legitimate times, then probably the PATT percentages would not have meaning for you. <br /> </td></tr></table><br /><br />180deg out again, Truly amazing!<br /><br />Thank you!

[old] bmoore · Post by **[old] bmoore** » August 29th, 2005, 11:02 am

[/quote]<br /><br />As I understand it, the C2 percentage is a percentage based on the number of entries in a certain class, if you are #25 out of 100 entries then you are at the 25th percentile. PATT is based on a percentage of World Record Time. Two different numbers, take your pick.<br /><br />You could also try the <a href='http://http://www.nonathlon.com/' target='_blank'>Nonathlon</a>. This is similar to PATT in that it assigns points based on percentage of best times, but they use a more complicated method to determine points. There is an explanation on the site. This is fun to use, check it out. <br />[/quote]<br /><br />Thanks for the Nonathlon reference. That's pretty cool. It reminds me of the an old swim meet we used to do that gave points for 5 events: a 100 of each stroke, along with a 200IM. I was a sprinter and usually finished in the top 3 in the league overall.

[old] bmoore · Post by **[old] bmoore** » August 29th, 2005, 11:06 am

<table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(John Rupp @ Aug 26 2005, 08:30 PM)</b></div></td></tr><tr><td class='quote'>I have decided to give the iggie bin a good cleansing, and have removed everyone from that list. Hence it is a clean slate and a starting over point for the list. Hopefully it won't need to be used again, but is there and convenient if need be.<br /><br />The points brought up are interesting, and also have something to do with my initially coming up with the concept of basing the times on world records. This is nothing new, as from high school I did the same with my running times, always aiming to raise the percentages up higher. I didn't do this all the time, but now and then it was entertaining to see. <br /> </td></tr></table><br /><br />John,<br /><br />Thanks for your responses to my inquiry. I appreciate your efforts on this.<br /><br />I must confess that I use the PATT to gage efforts on other distances, such as my recent half marathon, but I'm not so sure of the validity of using it to compare rowers across age, sex, and weight categories. I'll address more on this with your examples.<br /><br />Regards,

[old] bmoore · Post by **[old] bmoore** » August 29th, 2005, 11:36 am

<table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(John Rupp @ Aug 26 2005, 09:13 PM)</b></div></td></tr><tr><td class='quote'>2005 40-49 heavyweight men 2,000 meters (1435 entries)<br /><br />PATT (Dwayne Adams)<br />100% = 5:47.1 . . . . . . current ranking percentiles<br />90% = 6:25.7 . . . . . . 6:45.3 (58 seconds from WR awg)<br />80% = 7:13.9 . . . . . . 6:56.4 (only 11 seconds from 80 to 90 percent!)<br />70% = 8:15.9 . . . . . . 7:05.2 (only 8.8 seconds from 70 to 80 percent!)<br />60% = 9:38.5 . . . . . . 7:16.0 (only another 10.8 seconds!)<br />50% = 11:34.2 . . . . . 7:25.9 (10.9 seconds)<br /><br />The current "ranking percentiles" have the 90th percentile at 58 seconds from the WR awg, then only 40.6 more seconds to the 60th percentile! I understand that on a bell curve there are more in the middle (having gotten straight A's in statistics), however the question is how "important" it is to show the percentages that way, and whether it is encouraging and useful to participants. <br /><br />For one thing the percentiles are not accurate, as most people would normally be at 50% in a bell curve, not 84 or 87 or 90. Secondly, that the majority of rowers are between 84 and 87% of the world record time shows that the sample currently participating in the rankings is in a very small range, and affirms that it is not a good sample size or range of abilities to use the bell curve percentile concept for the rankings. <br /> </td></tr></table><br /><br />Actually, the 50% on the curve shows the point where 50% are above and 50% are below. There can only be one person at this exactly in the middle. Most of the people would fall in between the average plus/minus 1 standard deviation.<br /><br />Please show me how most rowers are between 84 & 87%. I don't follow you on this one.<br /><br />Let me use a personal example to compare my current 2k time to some slight improvements.<br /><br />7:14.9 - 62% ranking percentile (182/482) or 77.6% PATT.<br />7:10.0 - 67% ranking percentile (160/482) or 78.3% PATT.<br />7:00.0 - 76% ranking percentile (114/482) or 80.3% PATT.<br /><br />I think this highlights the differences between systems for improvement in conditioning and effort. When I finally get the 7:00 done (not if), I will have increased my PATT by 2.7. My ranking will have increased 68 places, and the percentile will have risen by 14. Although my numbers are smaller, the increases seem to motivate me more on the absolute ranking. I guess PATT suffers from the law of large numbers, where a larger number can't grow (as a percentage) as a smaller number can.<br /><br />Anyway, that's the highlights of my observations so far.<br /><br />Also, John, why the two times for your PATT results in your signature? Is that a personal best and a season best?

[old] bmoore · Post by **[old] bmoore** » August 29th, 2005, 12:47 pm

<table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(bobamiller @ Aug 27 2005, 04:13 AM)</b></div></td></tr><tr><td class='quote'><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(bmoore @ Aug 26 2005, 11:55 AM)</b></div></td></tr><tr><td class='quote'><br />Finally, if I can see where I rank across events at the click of a button on C2, why would I bother with PATT?<br /> <br /> </td></tr></table><br /><br />Consider this. When you're just starting out it can be hard set reasonable targets for yourself. For example, based on my personal best to date for 5000 meters, what pace can I set for myself at the beginning of 10000 meter row and have a reasonable expectation of being able to carry it to the end? How about for 60 minutes or even a half marathon? PATT provides useful reference information that helps me avoid both short-changing myself and dying midway through a hard effort.<br /><br />Bob Miller<br />Six weeks into a lifelong commitment, one row at a time. <br /> </td></tr></table><br /><br />Thanks Bob.<br /><br />I think The PATT's usefulness is limited to just this example. I also use it for gaging my time in other events, but I don't use it to set expectations between repeated attempts at the same distance. Once you establish a time for each event, and improve conditioning to a sufficient level, from then on you'd be competing against your personal best in that event and not any other distance. I also don't think it can be accurately used across age, weight, and sex groups.<br /><br />Best of luck to you in your training. It took me 7 years to hit 1 million meters, but I did the second million in 7 months. Now I'm addicted, and a heck of a lot fitter.<br /><br />Regards,

[old] bmoore · Post by **[old] bmoore** » August 29th, 2005, 12:50 pm

<table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(John Rupp @ Aug 27 2005, 11:38 AM)</b></div></td></tr><tr><td class='quote'><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(FrancoisA @ Aug 26 2005, 04:58 PM)</b></div></td></tr><tr><td class='quote'>1) Jari Lampi has the 40-49 lwt world record for the 10K with a time of 34:28.4; His PATT score would be 96.4%. In order to get a PATT score of 100% he would need to do 33:15![right] </td></tr></table><br /><br />2k PATT times are calculated based on existing 2k world records and world bests. All the 2k PATT times have been calculated this way.<br /><br />As the other 9 events are not as competitive, overall, as the 2k, they have been determined from the 2k PATT times, by the means of a "conversion factor". There are 9 conversion factors, one for each event. These conversation factors have been determined by the FASTEST performance in each of the 9 events, across all ages, weight classes, and gender. Then these 9 conversion factors are used for the 9 events.<br /><br />For example, lightweight Rod Freed did the 10000 meters in 34:16.2 at age 52.<br /><br />This gives him a 100% PATT and also sets the standard for all the other WR awg classes. Lampi, being much younger and also having a slower time than Freed, would of course not have 100%. His 96.4%, however, is still very good.<br /><br />Remember also that PATT is based on comparision with all other rowers! Other rowers would need to better 96.4% to be ahead of Lampi. However such performances are quite rare, at least comparatively speaking. Thus he doesn't lose anything by having a 96.4 percentage. If he had all time fastest performance then he would have 100%. <br /> </td></tr></table><br /><br />These conversion factors seem to be working out for me, but isn't there a problem with applying the same set of factors to everyone? I would think relative endurance would vary across age groups, and not have the same pace ratio between a 2k pace versus a 10k pace. What was the variability in your formula for longer distances?

[old] Porkchop · Post by **[old] Porkchop** » August 29th, 2005, 1:36 pm

<table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(bmoore @ Aug 29 2005, 10:36 AM)</b></div></td></tr><tr><td class='quote'><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(John Rupp @ Aug 26 2005, 09:13 PM)</b></div></td></tr><tr><td class='quote'>2005 40-49 heavyweight men 2,000 meters (1435 entries)<br /><br />PATT (Dwayne Adams)<br />100% = 5:47.1 . . . . . . current ranking percentiles<br />90% = 6:25.7 . . . . . . 6:45.3 (58 seconds from WR awg)<br />80% = 7:13.9 . . . . . . 6:56.4 (only 11 seconds from 80 to 90 percent!)<br />70% = 8:15.9 . . . . . . 7:05.2 (only 8.8 seconds from 70 to 80 percent!)<br />60% = 9:38.5 . . . . . . 7:16.0 (only another 10.8 seconds!)<br />50% = 11:34.2 . . . . . 7:25.9 (10.9 seconds)<br /><br />The current "ranking percentiles" have the 90th percentile at 58 seconds from the WR awg, then only 40.6 more seconds to the 60th percentile! I understand that on a bell curve there are more in the middle (having gotten straight A's in statistics), however the question is how "important" it is to show the percentages that way, and whether it is encouraging and useful to participants. <br /><br />For one thing the percentiles are not accurate, as most people would normally be at 50% in a bell curve, not 84 or 87 or 90. Secondly, that the majority of rowers are between 84 and 87% of the world record time shows that the sample currently participating in the rankings is in a very small range, and affirms that it is not a good sample size or range of abilities to use the bell curve percentile concept for the rankings. <br /> </td></tr></table><br /><br />Actually, the 50% on the curve shows the point where 50% are above and 50% are below. There can only be one person at this exactly in the middle. Most of the people would fall in between the average plus/minus 1 standard deviation.<br /><br />Please show me how most rowers are between 84 & 87%. I don't follow you on this one.<br /><br />Let me use a personal example to compare my current 2k time to some slight improvements.<br /><br />7:14.9 - 62% ranking percentile (182/482) or 77.6% PATT.<br />7:10.0 - 67% ranking percentile (160/482) or 78.3% PATT.<br />7:00.0 - 76% ranking percentile (114/482) or 80.3% PATT.<br /><br />I think this highlights the differences between systems for improvement in conditioning and effort. When I finally get the 7:00 done (not if), I will have increased my PATT by 2.7. My ranking will have increased 68 places, and the percentile will have risen by 14. Although my numbers are smaller, the increases seem to motivate me more on the absolute ranking. I guess PATT suffers from the law of large numbers, where a larger number can't grow (as a percentage) as a smaller number can.<br /><br />Anyway, that's the highlights of my observations so far.<br /><br />Also, John, why the two times for your PATT results in your signature? Is that a personal best and a season best? <br /> </td></tr></table><br />I never really thought much about this before, but as I read through this discussion, it occurs to me that there is a fatal flaw in trying to compare "PATT" percentages and percentile rankings. Further, there is a flaw in the theory of "PATT" percentages that limits its utility as a comparative tool.<br /><br />Using John's example, a 50% PATT score for 40-49 heavyweight men rowing 2000 meters would be 11:34.2; carrying the calculations further, a 25% PATT score would be 23:08.4, and a 10% PATT score would be 57:51. A 5% PATT score would be 1:55:42.0. A 2% PATT score would be 4:49:15.<br /><br />In sum, the time associated with the PATT score increases geometrically, not linearly. With PATT scores close to a world record, the difference is fairly insignificant, and they may be of some use as a comparative tool, but trying to correlate it to a bell curve or to compare across broad ranges of rowers is simply comparing apples to steamships. <br /><br />There may be 40-49 heavyweight male rowers pulling 11:34.2 as a 2K time, but they are not at the top of the bell curve (i.e., at the 50th percentile). Similarly, there may be 40-49 heavyweight male rowers pulling 23:08.4 as a 2k time, but they are not halfway between the top and bottom of the left side of the bell curve -- I think they would be on the far end of the left tail of the curve. The flaw is in making the world record time the numerator of the comparative fraction, and using an increasingly large denominator. The associated times increase toward infinity as a limit as the "percentage" approaches zero.<br /><br />In short, if there is a calculation that might be useful to rowers in comparing themselves to the world record, it is probably more along the lines of: "[personal best time minus world record time] divided by personal best time." [(PBT-WRT)/PBT] This calculation would tell a rower what percentage improvement is necessary to equal a world record. It won't necessarily tell the rower anything about his time compared to rowers in other age groups or divisions, but that's what percentile rankings are for. If more precise comparison is desired, then one would have to start looking at statistics or other data relating to performance decline as it relates to age.<br /><br />PS: I have also exercised my forum-given iggying rights in a preemptive strike.

[old] John Rupp · Post by **[old] John Rupp** » August 29th, 2005, 1:45 pm

<table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(bmoore @ Aug 29 2005, 08:36 AM)</b></div></td></tr><tr><td class='quote'>Also, John, why the two times for your PATT results in your signature? Is that a personal best and a season best?[right] </td></tr></table><br />Yes, personal bests over the last 3 years, and seasonal bests.<br />

[old] John Rupp · Post by **[old] John Rupp** » August 29th, 2005, 1:47 pm

<table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(Porkchop @ Aug 29 2005, 10:36 AM)</b></div></td></tr><tr><td class='quote'>it occurs to me that there is a fatal flaw in trying to compare "PATT" percentages and percentile rankings. [right] </td></tr></table><br />Right -- they are not comparable.<br /><br />PATT is absolute, i.e. it will always give the same percentage for time as relates to the WR for that AWG.<br /><br />The percentile just depend on who entered their times in the rankings.<br /><br /><br />

[old] bmoore · Post by **[old] bmoore** » August 29th, 2005, 1:54 pm

<table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(John Rupp @ Aug 29 2005, 01:45 PM)</b></div></td></tr><tr><td class='quote'><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(bmoore @ Aug 29 2005, 08:36 AM)</b></div></td></tr><tr><td class='quote'>Also, John, why the two times for your PATT results in your signature? Is that a personal best and a season best?[right] </td></tr></table><br />Yes, personal bests over the last 3 years, and seasonal bests. <br /> </td></tr></table><br /><br />Last 3 years? Slowing down are we? I'll hit 40 next month and am trying to avoid slowing down.<br /><br />Just out of curiosity, how many PBs are from each year? Is this calendar year or the C2 May-April year? If so, I wouldn't expect a lot of PBs from this season yet.

[old] bmoore · Post by **[old] bmoore** » August 29th, 2005, 1:56 pm

<table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(John Rupp @ Aug 29 2005, 01:47 PM)</b></div></td></tr><tr><td class='quote'><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(Porkchop @ Aug 29 2005, 10:36 AM)</b></div></td></tr><tr><td class='quote'>it occurs to me that there is a fatal flaw in trying to compare "PATT" percentages and percentile rankings. [right] </td></tr></table><br />Right -- they are not comparable.<br /><br />PATT is absolute, i.e. it will always give the same percentage for time as relates to the WR for that AWG.<br /><br />The percentile just depend on who entered their times in the rankings. <br /> </td></tr></table><br /><br />I realize that participation in the online rankings skews the rankings, but how bad do you think it is? I think the statistical packages can estimate this.

[old] Porkchop · Post by **[old] Porkchop** » August 29th, 2005, 3:47 pm

<table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(bmoore @ Aug 29 2005, 12:56 PM)</b></div></td></tr><tr><td class='quote'><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(John Rupp @ Aug 29 2005, 01:47 PM)</b></div></td></tr><tr><td class='quote'><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><div class='genmed'><b>QUOTE(Porkchop @ Aug 29 2005, 10:36 AM)</b></div></td></tr><tr><td class='quote'>it occurs to me that there is a fatal flaw in trying to compare "PATT" percentages and percentile rankings. [right] </td></tr></table><br />Right -- they are not comparable.<br /><br />PATT is absolute, i.e. it will always give the same percentage for time as relates to the WR for that AWG.<br /><br />The percentile just depend on who entered their times in the rankings. <br /> </td></tr></table><br /><br />I realize that participation in the online rankings skews the rankings, but how bad do you think it is? I think the statistical packages can estimate this. <br /> </td></tr></table><br />I don't think it skews them at all. This is (again) an apples and oranges question. The idea that self-selection skews the rankings is based on two fallacies: First, that there is utility in comparison to the general population; and second, that there are "rowers" who row at a snail's pace just to populate the left tail of the bell curve. <br /><br />There is no point in comparing oneself to people who aren't competing. Even at my pathetic pace, I can row faster than 99+% of men my age -- because well over 99% of men my age don't row at all. Their speed is zero and their time, therefore (by calculation), is infinite. By the same token, my major league batting average is .000 -- because I've never played major league baseball. I cannot imagine that any professional baseball player would care to compare his performance to my nonperformance. The only comparison that is meaningful is a comparison with those who do row, and the only source of usable information is the online rankings that allow one to compare oneself to those others. <br /><br />Are there rowers who don't report their times? Most probably there are, but my guess is that the ones who don't report them are less likely to attain competitive times than the rowers who do report their times. One cannot discount the possibility that somewhere out there is an individual who is a perfect rower and could blow away the world record in his/her age and weight group, but chooses not to participate in the online rankings. The percentile rankings, in effect, report one's position among those who matter -- fellow competitors. <br /><br />Someone who has never rowed before and is in utterly abysmal physical condition would probably turn in a terrible (comparatively speaking) 2K time. (Nevertheless, I doubt that anyone would, or could, actually row so slow as to be in, say, the 10% or 25% PATT group.) That is inconsequential to anyone who rows "seriously." It is probably inconsequential to the beginning, out-of-shape rower, too. Either he or she will never row again, or he or she will learn something about technique, get into better shape, and row better times in later attempts. In either case, the results of the first 2k row are irrelevant to other rowers (unless, for some odd reason, one needs to puff up one's ego). <br /><br />Someone who is chasing Dwayne Adams' record doesn't have any reason to care that he rows a 2K in half the time of some anonymous guy at the other end of the bell curve. The only thing he needs to know is how much faster he needs to go to catch Dwayne; he may also want to know how many others are ahead of him in that chase. PATT scores are essentially meaningless, because they don't answer either question. Simple subtraction (or, to couch the answer in terms of necessary percentage improvement, the calculation I proposed above) answers the first question, and the percentile rankings answer the second. Further, the "percentage" scores are mathematically meaningless due to the geometric increase in assumed times as the PATT score decreases.

[old] bmoore · Post by **[old] bmoore** » August 29th, 2005, 4:01 pm

Porkchop,<br /><br />I believe that the rankings would reflect a valid sample of the population of rowers, and that the self-selection wouldn't skew the times. I just want to see how John interprets the curve. <br /><br />I'm not sure which is crazier, not reporting a slow performance or chasing Dwayne Adams. (There's a movie title for you..."Chasing Dwayne").

Concept2 Forum

How To Calculate Your Patt Percentages

Training

Training

Training

Training

Training

Training

Training

Training

Training

Training

Training

Training

Training

Training

Training