Minimum vs. Cumulative vs. Average times for outcome

Debates and discussions on the various race scheduling methods that can be used and their fairness and accuracy in determining the winners.
User avatar
gravityboy
Apprentice
Apprentice
Posts: 6
Joined: Sun Jan 20, 2008 9:55 pm
Location: Warrenton, VA USA

Minimum vs. Cumulative vs. Average times for outcome

Post by gravityboy » Wed Mar 12, 2008 8:31 am

Group, I wasn't able to find a good thread in the archives on the timing methods available on a modern track with good software. I found some reference to 'fast' discussions (viewtopic.php?t=206) but none on deciding a winner. I'm not looking to start a points vs. timing discussion as our Pack has done timing for the last 5 years with no major hangups. I just was curious about timing options as things were a little different at District this year.

As I said before, our Pack has successfully run timed races for many years. We've used the 'cumulative' method as far back as I can tell and it seems to produce results which reward both the fast and reliable cars. On our 3 lane Piantedosi with timer each car runs one run in each lane. At the Districts this year I noticed the sponsoring pack was using the fast minimum as their determining factor on their 4 lane (4 run) races. Checking back to the data from my Pack derby I noticed this would have switched the 1st and 2nd place Pack champions this year. Averaging the times would have kept the places the same.

I come from a race car background and in things like Autocrossing, your fastest single lap is what matters. In Road racing (or NASCAR) obviously it is your cumulative time that matters most. I personally prefer a system that rewards the 'good' car vs. the 'sprint' car, but I'd like to see what others think...

Keith R.
Warrenton, VA USA



User avatar
Stan Pope
Pine Head Legend
Pine Head Legend
Posts: 6888
Joined: Sat Jul 05, 2003 7:01 pm
Location: Morton, Illinois
Contact:

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by Stan Pope » Wed Mar 12, 2008 9:01 am

Average and cummulative time comparisons are, of course, measuring the same aspects. It doesn't matter in the results whether you cite the total time or the average time, so long as each competitor runs the same "number of laps." As you say, they are a measure of consistency as well as speed. Running only three heats per car increases the role of sampling error as compared to running two or three heats per lane. As such, it increases the role of luck in the outcome. Still luck only dictates the outcomes for very closely matched competitors.

"Fastest run", on the other hand, gives full play to sampling error! Still luck does not dominate the outcomes except for very closely matched competitors. But the role of luck is increased as compared to the 3-run average time.

In the end, it is up to the organization to select those aspects that they value most and to structure the competition accordingly.

Good racing .. however you get it! :)


Stan
"If it's not for the boys, it's for the birds!"

User avatar
Randy and Son
Master Pine Head
Master Pine Head
Posts: 272
Joined: Sat Mar 03, 2007 9:02 pm
Location: Elkhorn, Nebraska

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by Randy and Son » Wed Mar 12, 2008 1:15 pm

I can't say that I am a fan of the "fastest run" method.

We have had one experience with it and didn't know about it until after the race was over.

I started to look at the results and couldn't figure out why the winners were the winners. It wouldn't have affected my sons car one way or the other, since we weren't competitive that year. Hadn't gotten edyukated yet.

The sample size was small and "luck" was in full bloom that day IMO. Cars that had better cumulatives didn't even place. The overall results just didn't look right and even the kids were surprised that other cars didn't place better in the standings. Maybe there is some justification in that method but its sure not my first choice.

Randy



User avatar
pwrd by tungsten
Master Pine Head
Master Pine Head
Posts: 723
Joined: Wed May 24, 2006 1:51 am
Location: Seattle, WA

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by pwrd by tungsten » Fri May 08, 2009 12:08 am

Would cumulative score with two passes down each lane be a good way to run a council race?

Any other techniques come to mind?


W Racing!!!!

User avatar
Darin McGrew
Pine Head Legend
Pine Head Legend
Posts: 1735
Joined: Fri Jul 18, 2003 1:23 pm
Location: Mountain View, CA
Contact:

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by Darin McGrew » Fri May 08, 2009 1:14 am

pwrd by tungsten wrote:Would cumulative score with two passes down each lane be a good way to run a council race?
That would be a good way to qualify finalists for a council race. However, the finals should be run with a CPN schedule, to account for differences between lanes. In a council race, you can expect cars to be much closer to each other in performance than is normally the case in a local race.



User avatar
Rod Turnbull
Master Pine Head
Master Pine Head
Posts: 239
Joined: Wed Dec 17, 2008 11:49 pm
Location: Ontario, Canada

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by Rod Turnbull » Fri May 08, 2009 7:18 am

Unless all your lanes are perfect my vote would be for scoring only the fastest ET on any of the 6 runs.

Reason: If you have 3 cars, one is a RR that goes left, one is a RR that goes right, and the last runs on all 4 down the middle of the lane... on a less than perfect track certain lanes will favor each of the cars due to joints or track set-up... if lanes 1,3,5,6 have flaws on the right side of the rail and lane 2&4 have flaws on the left side of the rail, with a total or average time you are giving the advantage to one of the cars... even the car that runs straight could hit the flaw in 5 of the 6 lanes... so an average or combined time shows signs of track condition rather that the cars full potential to go fast.

My professional opinion would be to run for fastest ET and use the times as qualifying for a final event and let the top 6 pick their lane (in order of fast ET time) and have one final race to determine the top 6 positions... that would be the most fair way I can think of to run a race. If you have a large field of cars you might even want to run semi finals and do 3 races with the top 18 cars and take the top 2 finisher from each semi final race to go to the final.



User avatar
Stan Pope
Pine Head Legend
Pine Head Legend
Posts: 6888
Joined: Sat Jul 05, 2003 7:01 pm
Location: Morton, Illinois
Contact:

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by Stan Pope » Fri May 08, 2009 7:54 am

Rod Turnbull wrote:Unless all your lanes are perfect my vote would be for scoring only the fastest ET on any of the 6 runs.
...
My professional opinion would be to run for fastest ET and use the times as qualifying for a final event and let the top 6 pick their lane (in order of fast ET time) and have one final race to determine the top 6 positions... that would be the most fair way I can think of to run a race...
Interesting comments!

This "works" if the track is mostly symmetrical, i.e. each lane penalizes some combined aspects of design least by the same total amount. If this is not the case (and I believe that it is usually not the case), then the the track penalizes some symmetry choices more than others, enhancing the role of PDL (pure dumb luck), or, worse, HTA (home track advantage!) Either the track or the venue (or the combination) can play a role in determining the extent of asymmetry.

Doncha just love TLA's (three letter acronyms?)

My curiosity has overwhelmed me ... what "profession" warrants having a "professional opinion" on competition methods? Only one I can think of is a branch of mathematics called "game theory." But that led to Stearns Method.


Stan
"If it's not for the boys, it's for the birds!"

User avatar
Stan Pope
Pine Head Legend
Pine Head Legend
Posts: 6888
Joined: Sat Jul 05, 2003 7:01 pm
Location: Morton, Illinois
Contact:

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by Stan Pope » Fri May 08, 2009 8:25 am

Darin McGrew wrote:
pwrd by tungsten wrote:Would cumulative score with two passes down each lane be a good way to run a council race?
That would be a good way to qualify finalists for a council race. However, the finals should be run with a CPN schedule, to account for differences between lanes...
I concur with the answer but not with the reasoning. Both have total lane balance! What CPN adds to the process is that observers can see that the times make sense (or not). They can see the finish line separation between each pair of cars and, from that, deduce whether the recorded times are consistent or not.
gravityboy wrote:... our Pack has successfully run timed races for many years. We've used the 'cumulative' method as far back as I can tell and it seems to produce results which reward both the fast and reliable cars.
We think that timer errors are infrequent, but when they occur to critical racers or in critical heats, they can do a real number of trophy awards, not to mention what they do to the credibility of the host organization! It is really embarrassing to have to tell the crowd, "Ummm ... we awarded the first place trophy to the wrong car. The correct winner was _______!"

During the past two years, I am personally aware of two district events in which the announced 1st place trophy winner was the wrong car. In one case, a single bad time (way below the fastest possible time for that track) skewed an also-ran's average time way low. In the second case, analysis of video recordings showed a single heat that was timed approximately 0.1 seconds too low for all three cars in the heat and shuffling the 2nd or 3rd place car up to 1st overall.

Now, these were timing errors where suspicion led to detailed analysis. How many timer errors occur that affect cars that "aren't in the running so the errors don't really matter?"


Stan
"If it's not for the boys, it's for the birds!"

User avatar
Darin McGrew
Pine Head Legend
Pine Head Legend
Posts: 1735
Joined: Fri Jul 18, 2003 1:23 pm
Location: Mountain View, CA
Contact:

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by Darin McGrew » Fri May 08, 2009 10:03 am

pwrd by tungsten wrote:Would cumulative score with two passes down each lane be a good way to run a council race?
Darin McGrew wrote:That would be a good way to qualify finalists for a council race. However, the finals should be run with a CPN schedule, to account for differences between lanes...
Stan Pope wrote:I concur with the answer but not with the reasoning. Both have total lane balance! What CPN adds to the process is that observers can see that the times make sense (or not). They can see the finish line separation between each pair of cars and, from that, deduce whether the recorded times are consistent or not.
Maybe I need clarification on what "cumulative score" means. I assumed that it meant a points-based system of some sort, presumably with a PPN-like schedule. Rereading the subject line of the thread, it seems that it might be timed. If timed, then I agree with Stan.



User avatar
Stan Pope
Pine Head Legend
Pine Head Legend
Posts: 6888
Joined: Sat Jul 05, 2003 7:01 pm
Location: Morton, Illinois
Contact:

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by Stan Pope » Fri May 08, 2009 12:18 pm

Darin McGrew wrote:... Maybe I need clarification on what "cumulative score" means. I assumed that it meant a points-based system of some sort, presumably with a PPN-like schedule. Rereading the subject line of the thread, it seems that it might be timed. If timed, then I agree with Stan.
You're right ... it is ambiguous. You assumed points; I assumed times. That explains what appeared to me as an "uncharacteristically blonde" answer.


Stan
"If it's not for the boys, it's for the birds!"

User avatar
Rod Turnbull
Master Pine Head
Master Pine Head
Posts: 239
Joined: Wed Dec 17, 2008 11:49 pm
Location: Ontario, Canada

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by Rod Turnbull » Fri May 08, 2009 4:14 pm

Stan Pope wrote:This "works" if the track is mostly symmetrical, i.e. each lane penalizes some combined aspects of design least by the same total amount. If this is not the case (and I believe that it is usually not the case), then the the track penalizes some symmetry choices more than others, enhancing the role of PDL (pure dumb luck), or, worse, HTA (home track advantage!) Either the track or the venue (or the combination) can play a role in determining the extent of asymmetry.
This works regardless of PDL, HTA, if your car runs good in all the lanes or only one... Out of 6 lanes you should be able to pull out at least one good run... that's all you need to qualify. Take a photo/video of the cars at the line in the final race and the timers don't even matter... there is no way to hand out a trophy to the wrong person.
Stan Pope wrote:My curiosity has overwhelmed me ... what "profession" warrants having a "professional opinion" on competition methods? Only one I can think of is a branch of mathematics called "game theory." But that led to Stearns Method.
LOL, game theory? How can anything that's only a theory be a profession... (JK) Actually I have spent a great deal of time (25+yrs) in the world of professional motorcycle racing (organizing, not racing)... Setting up race schedules/programs/scoring events in everything from local to national events, spent a few years on Ontario's Board of Directors for the CMA (equivalent to your AMA)... then moved up to represent more that half the professional riders in Canada on the National Board of Directors for the Canadian Motorcycle Association... I have a few awards for my contributions to the sport and feel qualified to advise with some authority on this topic. If anyone is feeling jealous yet... you can stop now, the pay sucks and the word volunteer came up a lot over the years. :)



User avatar
Stan Pope
Pine Head Legend
Pine Head Legend
Posts: 6888
Joined: Sat Jul 05, 2003 7:01 pm
Location: Morton, Illinois
Contact:

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by Stan Pope » Fri May 08, 2009 11:14 pm

Rod Turnbull wrote:This works regardless of PDL, HTA, if your car runs good in all the lanes or only one... Out of 6 lanes you should be able to pull out at least one good run... that's all you need to qualify. Take a photo/video of the cars at the line in the final race and the timers don't even matter... there is no way to hand out a trophy to the wrong person.
Since the question bear on lane equality, consider this track which penalizes RR cars as follows:

Code: Select all

Lane:       1     2     3
L DFW:    -0.03 -0.10 -0.10
R DFW:    -0.10 -0.04 -0.05 
On this track, run two cars equal in all regards except which front wheel is dominant RR, i.e. they run dead even on perfect lanes.

Based on "best run". the left DFW car wins: perfect-0.03 to perfect-0.04 seconds; Based on cumulative time, the right DFW car wins: perfect-0.19 sec to perfect-0.23 sec.
Both methods reward the wrong car! Photos don't help!

Either PDL or HTA plays a dominant role!

Game theory wins! :)

As to the question of whether PDL or HTA should be allowed to play any role ... I think that PDL is part of the game; HTA (or other specific advance knowledge of the track) should be avoided to the extent possible.


Stan
"If it's not for the boys, it's for the birds!"

User avatar
Rod Turnbull
Master Pine Head
Master Pine Head
Posts: 239
Joined: Wed Dec 17, 2008 11:49 pm
Location: Ontario, Canada

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by Rod Turnbull » Sat May 09, 2009 2:12 am

Stan Pope wrote:Either PDL or HTA plays a dominant role!
It's unlike you to make a ludicrous statement like this, it infers that either PDL plays a dominant role so there is no use for this forum and you might as well just hammer the nails through the wheels into the block of wood that came in the kit and let it run like that... or most PWD racers have no moral fiber and abuse the HTA.

PDL, part of the game sure, dominant role... no.

HTA, should be avoided, yes I agree, dominant role... no.

As for methods... even in your example I give the trophy to the car that crosses the finish line first in the final as proved by the photo, the best car for that day under the given conditions won... since you as the person scoring the event has no knowledge of the cars set-up or actual theoretical potential perfect run, and the lanes will not produce the exact same results every run... nether you nor the child nor the parent can dispute the race results... so for an example with no right answer, my method finds one.

If you want to shake this up a bit because you still don't like the ET being used to determine qualifiers starting positions... change the previous method of setting the lane choice for the final race to a random picking of numbers out of a hat to determine which lane cars run in. This random method (used for lane choice not actual lane number) is popular in Europe, in events like Speedway Racing, one of their most popular sports.

An added bonus of running semis and finals is that if you have a few times that came in incorrect due to a technical glitch, it will weed out the slow cars and only the fastest will end up on the podium.



User avatar
Stan Pope
Pine Head Legend
Pine Head Legend
Posts: 6888
Joined: Sat Jul 05, 2003 7:01 pm
Location: Morton, Illinois
Contact:

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by Stan Pope » Sat May 09, 2009 9:09 am

Rod Turnbull wrote:... even in your example I give the trophy to the car that crosses the finish line first in the final as proved by the photo, the best car for that day under the given conditions won...
In one of three runs, if I counted them correctly. It finished 2nd in the other two.

But, remember the context of the discussion ... summarize as "how to minimize/eliminate the influence of bad tracks on the race outcome." The post seems to be ignore that constraint and say "that's racing."

So, in the example that I gave with two equally prepared cars go at it, what was the most important determining factor in who got the first place trophy? I think that it depends on how each builder chose which wheel to make dominant ... inside info? or chance? (HTA or PDL?)


Stan
"If it's not for the boys, it's for the birds!"

User avatar
Stan Pope
Pine Head Legend
Pine Head Legend
Posts: 6888
Joined: Sat Jul 05, 2003 7:01 pm
Location: Morton, Illinois
Contact:

Re: Minimum vs. Cumulative vs. Average times for outcome

Post by Stan Pope » Sat May 09, 2009 9:38 am

Just to get the discussion back on track, I think that there is no one best choice that covers all track situations. My recent posts attempt to demonstrate the reasons for that belief. The best that we can do is pick a method that emphasizes build quality and reduces the influence of track variations to chance rather than to inside information.

I frankly don't know if "best run" or "best cumulative time" produce the best results most of the time. There are unique error risks inherent in each.


Stan
"If it's not for the boys, it's for the birds!"

Post Reply