Yes, I realize I'm comparing apples and oranges, or at least tangerines and clementines.glaforge wrote:Perhaps the close tolerances you had are attributed to a couple of factors...
The groups were not independent. You intermingled the groups, so that track/starter variations (if any) were applied somewhat consistently to all four groups. Thus I might expect the variances to be about the same. Your data hints at this, the aggregated by heat average times go up and down somewhat in sync between the groups.glaforge wrote:With your analysis, I'm now comfortable with 30 ms...especially when you factor in 4 different independant groups.
Had you raced the groups separately and gotten the same variances, then I'd be more convinced.
These two cars deserve a CPN chart, along with 156. My choice would be a 2-round chart (twice in each lane) so they'd race 4 to 8 times depending on how many cars you let into the finals.glaforge wrote:How many more times do you want them to go at each other?
It might. CPN will produce ties if the track is more uneven than the cars.glaforge wrote: Question: Would a perfect N chart fail to pick winner in this instance?
Breaking two-way ties is not hard, though, if you have two decent lanes. I have never gotten into the dreaded race-forever situation.
And within the Pack, I do not break ties for first, and I do not break ties for third. I just give out extra ribbons, medals, trophies, whatever.
The only ties I break are ties for second, because that decides who advances from the Den to the Rank final, or from the Rank finals to the District.
Did 156 get to go to the District derby also? It certainly seems possible to me that he was the fastest.
Very interesting study. I'm becoming a believer that time is much better at selecting finalists than points. And while I still like the CPN finals approach, but I understand the appeal of your approach also.
Thanks! I'm ready for this year's data!