How many advance to the pack final?

[Buckeyefan]

Hi all,
Brand new to the board and in my second year as PWD chair. Pack 35 now has over 140 scouts in 12 dens. We run individual den races and award 1st through 3rd place trophies. In years past, we advanced each den champ to the pack final. Unfortunately, in highly competitive dens, one or more cubs who could have placed in the pack final did not get there. Last year, with 9 dens, we took the top 2 from each den, for 18 cars in the final. This worked well on a 3 lane track with 6 runs for each car with the slowest time thrown out and winner determined by average of best 5 times. Unfortunately, the third fastest car in the pack on average after the first 6 runs came from our den also (it really wasn't close). So again, a fast car that would have placed 3rd in the pack didn't make the final. This year, I'm thinking every den champ gets in the pack final, and the next 6 fastest cars who didn't win their den make the final also. We'll keep a running list of who is on the bubble, kind of like a "Race for the Chase" PWD style. That will qualify 18 cars for the Pack final on our new 6-lane Best-Track. This way every den is represented in the final, and all the fastest cars earn their way into the final. What do you all think?
Buckeyefan

[Stan Pope]

If your timing setup is "solid" (e.g. spring start gate, consistent timer operations, etc.), then you plan works okay. I think that it improves over last year.

Now, I assume that you are running the den races and the finals by time. Do you "wipe the slate" and start new average times for the finals competitors? If so, what in your operations makes the finals times more meaningful and/or accurate than the times from the den races? What car attributes are you thereby emphasizing? Lubrication longevity? Racing skill improvement (if the boys stage their own cars)? ...

If den and finals times are combined, then the results may gain accuracy by averaging the times over more runs.

[Buckeyefan]

Stan,

Thanks for the quick reply! First off, we are running the finals by time, and with a brand new 6-lane BestTrack with spring gate and Champ timer, I plan to have a rock solid timing setup. I plan to run a series of test races after initial setup to compare times on each lane and across all lanes to check for any significant variation. We should have the track in 2-3 weeks, so I may be asking for opinions on a test plan when we are up and running.

Next, I guess I never really though out the reason for wiping the slate clean and starting over in the finals. It may be a relic of derbies past and the way that the Grand Prix software was set up. Since we have such a big pack, we do a weigh-in on Friday night and impound the cars until after the race. We start the races on Saturday with the Tigers at 9:00, Wolves at 10:00, Bears at 11:00 and Web I & II at 12:00 followed by the Pack final at 1:00 and the parent/sibling/open races at 2:00. Many of the Tigers and Wolves leave early and come back for the finals so everyone doesn't see all the races. (Even though we serve donuts early and pizza for lunch and encourage them to stay and cheer for the other dens.) Because of this, we treat the finals as a completely different race and start everyone on an equal footing. It would seem that the well built and prepared cars would retain their speed through 12 races, but I may be wrong. I may go back and pull the times for the top 10 from last year and do some analysis to see if using the best 10 or 11 times over 12 races (throw out the 1 or 2 slowest) would make a difference in placing.

Thanks for the ideas! Any more suggestions would be GREATLY appreciated!
Buckeyefan

[Stan Pope]

Since your racing may span nearly 5 hours, including den and finals times requires that the track characteristics are stable through the entire 5 hours. By the same token, even if the den and finals times are separated, you still need stable characteristics through about an hour. So, what might change track characteristics and how might you prove that the the track is stable?

Sources:
1. Track adjustments: Any physical change to the track that changes the location of a track segment will alter the times. Even small subtle changes will show up in the times.
2. Electronics stability: Most well designed electronics will be adequately stable over the duration and range of temperatures you are likely to experience. However, you can't just assume that your specific timer is that stable.

Proofs:
1. A small group of well prepared test cars should show consistent times over a series of 12 runs. For a six lane track, you could probably demonstrate consistency with six runs prior to the start and six more runs prior to the finals. Cars are three left front wheel dominant rail-riders (cars A, C, and E) and three right front wheel dominant rail-riders (cars B, D and F). Then run a set of six timed heats as follows (A,B,C,D,E,F), (A,B,C,D,E,F), (A,B,C,D,E,F), (B,A,D,C,F,E), (B,A,D,C,F,E), (B,A,D,C,F,E). Compute average time and standard deviation for each test car for each lane it uses. In preparing these cars, consistency is more important than speed. Make DFW toe-in adjustments as needed to get good consistency.

Calibration: Prior to race day (or just after setting up the track), run two sets in succession to show effects of graphite wear between the sets. Then relubricate and run two more sets in succession to show insensitivity to relube. Set to set variations for each car in each lane it uses during these test runs tells how much variance you should expect in system validation runs.When staging the cars use something like Lastufka's staging guide to assure consistent staging.

Validation: Run a set before the first den races and again before the finals.
A. If the "validation timings" show the same pattern and magnitudes of variance as they did in the "calibration runs", then combining den and finals times is indicated.
B. If the "validation timings" show less individual variations than the difference between the pairs of top six cars, then combining den and finals times is okay, since such track changes as may have occurred will have a low probability of affecting the placement results. (You could look to racing history to make this call.)

A statistics "purist" may say that you need more than three runs in each lane by each car in each set, but I think you can do a very satisfactory job with the plan I've laid out.

Intuitively, some might criticize the plan for not running each car in each of the six lanes. The reason that it is not necessary to do so is that we are trying to demonstrate individual lane consistency.

[Buckeyefan]

Stan,
Thanks for the detailed reply! I am still working on a single rail rider, let alone a set of six! This definitely gives me something to shoot for though. We will be setting the track up in a neighbor's basement on a concrete floor for the initial testing, so as long as nobody trips over it, the track should be very stable. I plan to use cars from previous years for the initial runs to make sure everything works: smooth joints, starting gate, stop section, finish line, computer, and software etc. Then we will switch to the test cars and run the experiment. I guess I better get started on wheels and axles.
Thanks again,
Buckeyefan

[Stan Pope]

... track should be very stable ...

True. But "should" and "is" are sometimes miles apart. And "your should" and "another's should" may also differ. The solution is a demonstration that brackets the racing environment. The tests in your basement are there to prove the cars and to quantify the lube losses. The verification happens on race day, before and after the racing.

I'm digging through the math and logic underlying my "intuition based plan" given above. Partly the goal is to ground it with better logic and partly to detail the actual before/after comparison procedure. A spreadsheet would be nice, I think, but what that would be depends on what the underlying theory allows.

[Stan Pope]

I plan to use cars from previous years for the initial runs to make sure everything works: smooth joints, starting gate, stop section, finish line, computer, and software etc.

Good tools for the joint evaluation include a sharp-corner wood block that you can slide down-track against each side of each lane guide. If it "catches" on either the guide or the track floor, it is a problem.

A 3' to 4' "straight edge" (I have a 4' long aluminum "yard stick") can show that the joints mesh without the track changing direction. You should be able to hold the center of the straight edge against the rail joint with both ends of the stick in contact with the rail. Pushing each end of the straight edge toward the rail should cause no lateral motion.

[Stan Pope]

Follow up on combining prelim and finals times ...

If you conclude that you can not satisfy the vast majority of families that the track and timing characteristics are unchanged for the duration of the event, then you can fall back on what most people do ... start afresh in the finals. Since the cars in the finals are faster than the cars in the prelims (all the slow cars have been excluded), the probability of inaccurate ranking increases unless the number of heats is increased. But running 12 finals heats for each car might be carrying a good thing too far! However, if your prerace testing shows one or more lanes to be more erratic than the others and your aren't able to fix 'em, then you could drop one or more lanes from the finals and run 8 heats (4 lanes) or 10 heats (5 lanes).

[Buckeyefan]

Stan,

I have been thinking about altering the derby setup to average all runs (Den races and pack finals) to crown the champion, but in our pack it comes down to expectations. We really only have three true "pineheads", several who are curious and on the fence, and a lot of parents and cubs who just build a "cool" car for the different judging categories. In my informal poll, we pineheads and most of the others all agreed to wipe the slate clean and start fresh in the finals so anyone not staying for all the den races, but showing up again for the pack finals, could follow the results and see who has the upper hand. A quick review of the top 5 from last year showed that the margins narrowed between first and second in the finals, but the finish order would have been the same if we had used the results from the den races to crown the pack champion. But where's the fun in that? Let the kids race as many times as they can!

Thanks for all the great advice, I will be putting it to use when the track arrives!

Regards,
Buckeyefan

[Stan Pope]

"Works for me", Bef... and it sounds like it will work for you! You did your homework!