CrossFit Discussion Board  

Go Back   CrossFit Discussion Board > CrossFit Forum > Competitions
CrossFit Home Forum Site Rules CrossFit FAQ Search Today's Posts Mark Forums Read

Competitions Competitions, contests & challenges

Reply
 
Thread Tools
Old 07-20-2009, 03:47 AM   #101
Cormac O'Connor
Member Cormac O'Connor is offline
 
Cormac O'Connor's Avatar
 
Profile:
Join Date: Jan 2008
Location: Tokyo  Japan
Posts: 253
Re: Games Scoring

Just a quibble - The DL was not just a 1RM event. The ones who finished did 20 lifts in total, all at above 300lbs so I believe it was intended to test endurance as well as maximal strength, since they had a 30 sec time limit for each lift. Nobody could reasonably be expected to hit their true 1RM under those circumstances, unless you factor in the adrenaline rush from competition offsetting the fatigue maybe*.

From Tony Budding's comments on the video, I'm betting that they knew some guys would be able to lift 500lbs (after all, they put together the athlete's profiles, most of which list the DL 1RM). They just didn't expect them to be able to lift that much after a) the hill run and b) 19 or so heavy DLs.

It didn't work out exactly as intended perhaps but it was still a great attempt. We all learned a little about Crossfitters' DL capacity from it I think!

* Jeremy Thiel, for example, has a 1RM DL of 500lbs according to his games profile, and he finished the event. Would love to know if that's an old PR or what. On the other hand, Mikko Salo has a DL PR of just over 500ls (230kg), but he didn't quite hit it despite getting more rest than most thanks to his blistering run.
  Reply With Quote
Old 07-20-2009, 07:14 AM   #102
Joe Cavazos
Affiliate Joe Cavazos is offline
 
Joe Cavazos's Avatar
 
Profile:
Join Date: Jul 2007
Location: Houston  Texas
Posts: 775
Re: Games Scoring

Quote:
Originally Posted by Cormac O'Connor View Post
Just a quibble - The DL was not just a 1RM event. The ones who finished did 20 lifts in total, all at above 300lbs so I believe it was intended to test endurance as well as maximal strength, since they had a 30 sec time limit for each lift.
Obviously, but that doesn't matter. I'm saying the scoring method is flawed because people did not get to demonstrate their true potential in the Deadlift event. So, in his system, Miko suffers less from the event because his max lift is barely 10 pounds under the maximum weight lifted. In the current system, Miko suffers more because he was beaten by 16 or so people.
__________________
My WFS Log || Stats I'm least proud of: Max Muscle-ups: 1 | Max Push-ups: 38 | Max HSPU: 9 | 19" Vertical Leap | Max L-sit: 5 sec.

Last edited by Joe Cavazos : 07-20-2009 at 07:22 AM.
  Reply With Quote
Old 07-20-2009, 08:16 AM   #103
Michael V. Erickson
Member Michael V. Erickson is offline
 
Profile:
Join Date: Oct 2007
Location: Concord  California
Posts: 85
Re: Games Scoring

Quote:
Originally Posted by Joe Cavazos View Post
Obviously, but that doesn't matter. I'm saying the scoring method is flawed because people did not get to demonstrate their true potential in the Deadlift event. So, in his system, Miko suffers less from the event because his max lift is barely 10 pounds under the maximum weight lifted. In the current system, Miko suffers more because he was beaten by 16 or so people.
I agree the "proportional performance scoring" method cannot correct the set-up error of the DL event. In particular it prevented Jason Khalipa from making up as much ground as he should have in this event. Following are Jason's relative rankings at the end of subsequent events w/ both scoring methods:
Games scoring: 72-39-26-18-12-11-7-5
PPS scoring: 72-47-35-18-8-5-1-2
  Reply With Quote
Old 07-20-2009, 12:13 PM   #104
Dimitri Dziabenko
Banned Dimitri Dziabenko is offline
 
Profile:
Join Date: Apr 2007
Location: Toronto  ON
Posts: 1,080
Re: Games Scoring

Quote:
Originally Posted by Michael V. Erickson View Post
After reading the various criticisms about the scoring and giving it a lot of thought I came up with an idea. I've spent a bunch of time applying it to the game data from last weekend and came up with some interesting results. My idea: what if all of the competitors were scored points relative to their performance compared to the top guy in each event. This would work in every type of event, timed events, max weight events, and amrap events. Give a maximum of 1000 points for the winner of an event. Every other competitor gets whatever percentage of 1000 points dependent on their relative performance. The number of competitors in the contest will not effect your points total regardless of how many drop out. The maximum points at the end of the day wins the contest. I applied this method to the men's games data and got this as a final result:

(Final) Points
(x) = games place Total
(1)Mikko Salo 7,187.52
(5)Jason Khalipa 7,141.90
(8)Patrick Burke 6,973.45
(2)Tommy Hackenbruck 6,966.51
(4)Steve Willis 6,879.55
(7)Blair Morrison 6,836.52
(3)Moe Kelsey 6,650.12
(10)DJ Wickham 6,640.27
(9)David Millar 6,472.34
(6)Peter Egyed 6,466.04
(11)Michael FitzGerald 6,422.95

Not saying this is perfect, it's just another way of doing it. I have the complete spreadsheet in google docs which I will share iwith anyone interested. I'm almost finished with analysing the women's data as well. I have no experience in competition myself and am not a trainer. Someone with such experience might glean something interesting form the statistics contained in the intermediate data.
Neat idea, but given the current data it's not very worthwhile since Everett got eliminated and would most likely have won the Snatch, which alters performances relative to whatever benchmark he would have set (alters them in a nonlinear way too). This may be an interesting idea for the future, even though I am not exactly convinced how this relative thing is supposed to work ie. if somebody runs a distance in 20 min, I run it in 30 min and the worst person in 50min, what score do I get relative to the winner? Is this score dependent on just how poorly the worst person does (shouldn't be the case)? Moreover, is this data consistent over all workouts, that is, to what extent are deadlift/run superiorities linearly comparable?
  Reply With Quote
Old 07-20-2009, 08:27 PM   #105
Michael V. Erickson
Member Michael V. Erickson is offline
 
Profile:
Join Date: Oct 2007
Location: Concord  California
Posts: 85
Re: Games Scoring

Quote:
Originally Posted by Dimitri Dziabenko View Post
Neat idea, but given the current data it's not very worthwhile since Everett got eliminated and would most likely have won the Snatch, which alters performances relative to whatever benchmark he would have set (alters them in a nonlinear way too). This may be an interesting idea for the future, even though I am not exactly convinced how this relative thing is supposed to work ie. if somebody runs a distance in 20 min, I run it in 30 min and the worst person in 50min, what score do I get relative to the winner? Is this score dependent on just how poorly the worst person does (shouldn't be the case)? Moreover, is this data consistent over all workouts, that is, to what extent are deadlift/run superiorities linearly comparable?
This is what I think would happen:
The best score is just a reference to judge each competitors relative performance. Wherever the reference is set the competitors still get a score based on the performance compared to each other. If the snatch competition had been early on and Josh had competed the total points scored by the other competitors would be smaller but relative to each other (others besides Josh) would not have changed. If Josh had not made it to the finals the only effect for the other competitors is the snatch points would have slightly less weight than the other events. The final competitors scores would still represent their total work capacity across all of the events relative to each other. If we believe in the crossfit philosophy that being very specialized at one thing will diminish your capacity at other things. Therefore across many events like at the CF games the winner could be a person who scored very well at every event but didn't win any of them. I'm thinking that scoring by relative performance measures total work capacity better than scoring by order of finish.

The worst performer does not effect other performers scores in any way. You're only compared to the best performer.

In your example your score would be calculated by dividing you time into the best time and multiplying by 1000: 20/30 x 1000 = 666.67 points. Lifting events you would divide your best lift by the overall best lift times 1000. (The "times 1000" is arbitrary just because I like big numbers). IE: 250/295 x 1000 = 847.46 points. For AMRAP divide your repetitions / by the best performers reps times 1000, etc.

The data is comparable and fair over widely varied types of events such as in the games because your scores are being compared to the best performance on your live competitors in each event. How do you rank against the best at each? If it sounds like I'm making this up as I go along, I am. I hope this makes sense.
  Reply With Quote
Old 07-20-2009, 09:34 PM   #106
Wade Smith
Member Wade Smith is offline
 
Wade Smith's Avatar
 
Profile:
Join Date: Sep 2007
Location: Savannah  GA
Posts: 1,269
Re: Games Scoring

Disclaimer #1: I loved the Games and have nothing but appreciation for the HQ staff, the Castros, and everyone involved.

Disclaimer #2: Mikko won 1st place so nothing kept him from winning.

Having said those things, I was concerned for Mikko when he was deadlifting the final bar(s). Something occured which could potentially have been a problem. Those of you in attendance there may remember that Dave called the event "over" while Mikko was still on the line. He didn't see Mikko there. Seems everyone else had finished and Mikko had just lifted the 495lbs. Because of the mistake, Mikko seemed to be rushed to the final bar (505lbs) when Dave realized he was there.

No big deal since Mikko won anyway, but I would guess he had that last 505lb lift in him but possibly felt rushed or confused at that point. It might have been good, in retrospect, to give him the full 10 seconds to move to the next bar because of the mistake. The full 10 seconds to move and 20 seconds to lift would have given him full amount of time to mentally prepare and physically give it his 100%. Just a thought.

Ultimately, the Games were awesome and Mikko did an awesome job.
  Reply With Quote
Old 07-20-2009, 10:41 PM   #107
Dimitri Dziabenko
Banned Dimitri Dziabenko is offline
 
Profile:
Join Date: Apr 2007
Location: Toronto  ON
Posts: 1,080
Re: Games Scoring

Quote:
Originally Posted by Michael V. Erickson View Post
This is what I think would happen:
The best score is just a reference to judge each competitors relative performance. Wherever the reference is set the competitors still get a score based on the performance compared to each other. If the snatch competition had been early on and Josh had competed the total points scored by the other competitors would be smaller but relative to each other (others besides Josh) would not have changed. If Josh had not made it to the finals the only effect for the other competitors is the snatch points would have slightly less weight than the other events. The final competitors scores would still represent their total work capacity across all of the events relative to each other. If we believe in the crossfit philosophy that being very specialized at one thing will diminish your capacity at other things. Therefore across many events like at the CF games the winner could be a person who scored very well at every event but didn't win any of them. I'm thinking that scoring by relative performance measures total work capacity better than scoring by order of finish.

The worst performer does not effect other performers scores in any way. You're only compared to the best performer.

In your example your score would be calculated by dividing you time into the best time and multiplying by 1000: 20/30 x 1000 = 666.67 points. Lifting events you would divide your best lift by the overall best lift times 1000. (The "times 1000" is arbitrary just because I like big numbers). IE: 250/295 x 1000 = 847.46 points. For AMRAP divide your repetitions / by the best performers reps times 1000, etc.

The data is comparable and fair over widely varied types of events such as in the games because your scores are being compared to the best performance on your live competitors in each event. How do you rank against the best at each? If it sounds like I'm making this up as I go along, I am. I hope this makes sense.
This is a better system than the ranking one used at this year's games. It doesn't eliminate the problems that come from eliminating participants, especially people who could be #1 in day two events, for the difference in points among people may be more significant in one case than the other and thus over 3 workouts schuffle the competitors around more than would appear. It has the benefits of making a statement about the actual output, vs. ranking.

Potential problems: Not sure if the 20/30 *1000 for times, vs. 250/295 for weights is really consistent across all workouts.
For the hammer/row workout: Best time is roughly 4.5min, worst time: 12min, so 3 times slower (ok, roughly). I don't think there is a way for the best competitor to deadlift 3 times more than the worst. Same for the snatch and the hill run. Basically, some workouts naturally have a high spread, others like the hill run don't (38min vs. 1hr). If you assign scores in a linear fashion, you would fail to distinguish between these types of workouts. I think compared to some of the other flagrant faults, this one is probably more of a theoretical difficulty.

For all of Crossfit's technical talk, we really don't have a practical way to measure fitness objectively and consistently. We get away with it, because people see results (ie. lifts go up, BF down, times down) and of course they are getting fitter. But here a statement is needed on whether a fast run cancels out a lower lift, ie. comparing runs to deadlifts. I fear no reasonable conclusion may be reached, since it's hard to standardize these things objectively.
  Reply With Quote
Old 07-21-2009, 08:52 AM   #108
Michael V. Erickson
Member Michael V. Erickson is offline
 
Profile:
Join Date: Oct 2007
Location: Concord  California
Posts: 85
Re: Games Scoring

Quote:
Originally Posted by Dimitri Dziabenko View Post
This is a better system than the ranking one used at this year's games. It doesn't eliminate the problems that come from eliminating participants, especially people who could be #1 in day two events, for the difference in points among people may be more significant in one case than the other and thus over 3 workouts schuffle the competitors around more than would appear. It has the benefits of making a statement about the actual output, vs. ranking.

Potential problems: Not sure if the 20/30 *1000 for times, vs. 250/295 for weights is really consistent across all workouts.
For the hammer/row workout: Best time is roughly 4.5min, worst time: 12min, so 3 times slower (ok, roughly). I don't think there is a way for the best competitor to deadlift 3 times more than the worst. Same for the snatch and the hill run. Basically, some workouts naturally have a high spread, others like the hill run don't (38min vs. 1hr). If you assign scores in a linear fashion, you would fail to distinguish between these types of workouts. I think compared to some of the other flagrant faults, this one is probably more of a theoretical difficulty.

For all of Crossfit's technical talk, we really don't have a practical way to measure fitness objectively and consistently. We get away with it, because people see results (ie. lifts go up, BF down, times down) and of course they are getting fitter. But here a statement is needed on whether a fast run cancels out a lower lift, ie. comparing runs to deadlifts. I fear no reasonable conclusion may be reached, since it's hard to standardize these things objectively.
The weighting is consistent in that it is set by your peer who does the best in each event. I believe the large differences in best to worst score will show up in the events where there is a special skill. Driving a spike is cake if you've done it a lot, an exercise in frustration if you have not. The biggest differential in best/worst relative scores was in event seven.

I've finished the PPS scoring of the womens event: Here is the final placement and scores:

Clever, Kristan - 7,053.73
Vale, Charity - 6,979.75
Kepler, Carey - 6,492.38
Wagner, Tanya - 6,268.24
Gentry, Jolie - 6,026.23
Olson, Jenny - 5,850.70
Smith, Lindsey - 5,839.27
Dunsmore, Sarah - 5,781.88
Kroon, Stacey - 5,771.16
Phillips, Christy - 5,724.47
Thorisdottir, Annie - 5,698.10
Mcreynolds, Crystal - 5,625.83

Tanya dropped because of a relatively poor performance in the last two events, especially event 7.

Last edited by Michael V. Erickson : 07-21-2009 at 08:57 AM.
  Reply With Quote
Old 07-30-2009, 09:13 PM   #109
Michael V. Erickson
Member Michael V. Erickson is offline
 
Profile:
Join Date: Oct 2007
Location: Concord  California
Posts: 85
Re: Games Scoring

I did some reading about scoring decathlons and found an interesting history. Early on (about 100 years ago) they adopted a points system based on a linear scale of performance. The reference was whatever the world or national record was in the event. Later they went to an exponentially based scoring system to reward the increasing effort required as the limit of human performance was reached in an event. Even later it was determined that this scoring system weighted some of the events much better than others, that is, the specialists (sprinters) had a large advantage over the field guys. They're still debating over how to score using what particular non-linear method.

Here is some interesting reading: [W/F/S]

http://www.coachr.org/decathlon_towa...ustainable.htm

http://swiss2.whosting.ch/mdetting/s...s-history.html

Note on the proportional performance scoring: I averaged Jason Khalipa's performance on the last 7 wods: 92% compared to Mikko's 88%. Khalipa just killed everyone after the first event relative performance wise. The average of all events Mikko has 89.84% compared to Khalipa's 89.27%. Just nipped it at the end. If Jason fixes [Paleao? Zone?!!] whatever the cramping problem caused him all the trouble in the run I think Mikko will have to up his game significantly next year to beat him.
  Reply With Quote
Old 07-30-2009, 10:46 PM   #110
Cormac O'Connor
Member Cormac O'Connor is offline
 
Cormac O'Connor's Avatar
 
Profile:
Join Date: Jan 2008
Location: Tokyo  Japan
Posts: 253
Re: Games Scoring

After reading the Crossfit Journal bit on Khalipa's run, I think it was mental pressure as much as anything else. If they'd done a different event first, he might have screwed that up as well!

Or he might have killed it, we'll never know. Such are the vagaries of sports.
  Reply With Quote
Reply


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is Off
HTML code is Off
Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
CrossFit Games v's Regional/State Games? Scot Heyes Competitions 12 08-19-2009 05:32 AM
'Nicole' Scoring Robert Miller Workout of the Day 5 01-21-2008 04:48 AM
How Fit Are You? scoring Erik Tiner Fitness 19 12-07-2007 10:28 PM
Tabata scoring Jeryd Leuck Fitness 4 09-23-2007 04:41 PM
Tabata Scoring Patrick Johnston Exercises 25 02-26-2004 05:57 PM


All times are GMT -7. The time now is 02:00 PM.


Powered by vBulletin® Version 3.6.8
Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.
CrossFit is a registered trademark of CrossFit Inc.