Best method for 3 lanes, no timers?

[dknowles67]

Forgive me if this question has already been answered on this forum.
I tried to read all the posts, but I sort of got lost. I'm not familiar with all the scheduling methods, and had a tough time determining which methods required timers.
Our pack uses a 3 lane track, with a double elimination method, and no regard to lane choice. (Assumtion - all lanes are equal, although this is clearly no so).
I'm upset with with the general unfairness (not because my son lost) of this method but rather than complain about it, I would prefer to offer a suggestion for improvement next year. (We raced this past Sunday - for the record, my Wolf son got 1st place in his age group).
Our Pack has approximatly 80-90 scouts. We split the races up by age, and have 4 races, each starting 1 hour apart.
Ideally, I think we should buy an electronic timer, but I'm not sure I can convince the Pack it is worth it, and I don't think I can convince my wife I should buy one and donate it.
I am proficient in computers, and writing software, so I could easily cook up something to determine race/lane assignments at check in time, but I'm not sure what the best method to use would be.
(We typically don't pre-register for the races, its just whoever shows up on race day).

So, without a finish line timer, is there a good way to determing the 3 fastest cars in order, (out of around 20 cars for each age category), taking into account that there may be wide variances in speed in each of the lanes?
Would it take longer than an hour?
{emoticon here of a smily scratching his head - couldn't find one like that}

[Cory]

So, without a finish line timer, is there a good way to determing the 3 fastest cars in order, (out of around 20 cars for each age category), taking into account that there may be wide variances in speed in each of the lanes?
Would it take longer than an hour?

Here are two very good resources on this topic:

Darin McGrew's Derby Race Methods - http://www.rahul.net/mcgrew/derby/methods.html

Stan Pope's Conducting Pinewood Derby Races - http://members.aol.com/StanDCmr/pwraces.html

It's worth taking the time to read through them thoughtfully.

There are lots of build-your-own timer plans out there. Check the PWD Webring. http://www.webring.org/cgi-bin/webring?ring=pwd;home

[Stan Pope]

Well, that is quite a challenge. I can only think of three methods that would do a decent (fair) job with that equipment: Multiple Elimination, PPN, and Stearns.

PPN and Stearns are run using charts which tell which cars to run in which lanes in each heat, and points are given for winning vs. not winning. Charts are based on number of cars running, so you would have to be able produce charts quickly, or use some software which manages the races. PPN will assure lane equity and near opponent equity. Stearns will do nearly as well on both counts. Both produce more accurate results if points are based on finish place than simply winning vs. not winning. Neither method accommodates late arrivals. Software to generate PPN charts is available in javascript form at http://members.aol.com/standcmr/ppngen.html and subroutines for inclusion in race management are licensed free of charge. It has been incorporated into commercial products such as gpraceman's GPRM software.

Multiple Elimination can be run with charts or without. Without charts, the method accommodates late arrivals. Without computers and/or charts, the method is easy to run for quadruple, quintuple, etc. elimination. Lane fairness is determined by drawing for lane just before each heat. Heats can be either two or three racers at a time. For as few as 20 racers, I think that I'd run two at a time, in which case, I'd select the two most evenly matched lanes. You can estimate the number of heats required with a javascript in http://members.aol.com/standcmr/nelim.html From number of heats, you can estimate the time to run at various elimination counts.

[dknowles67]

I reviewed this data, but I'm not sure it answers my question.
It looks like a Partial Perfect N chart would be better than what we do now, but requires prior knowledge of how many cars will race, unless you generate the chart at the races.
Am I correct in assuming that this will accurately determine the 1st, 2nd, and 3rd fastest cars in the correct order regardless of variances in lane speed?
I'm assuming you mark each heat 1st, 2nd, and 3rd, (3-lane track) assign points, and determine the point totals at the end.

[gpraceman]

Ideally, I think we should buy an electronic timer, but I'm not sure I can convince the Pack it is worth it, and I don't think I can convince my wife I should buy one and donate it.

A timing system doesn't have to be an expensive purchase. If someone can do a some soldering and some carpentry, you can build a system for $50 or less. There are several plans on the web for building a parallel port timing system, many of them listed at GrandPrix Race Central's Finish Line page. You can also buy a commercial kit that is simpler to build and has a computer interface, which is theFast Track $40 Cheap Kit.

With these options, I really feel that having a timing system is within the reach of the majority of organizations and you reap the benefits of having an better judge than the human eyeball.

[Stan Pope]

Am I correct in assuming that this will accurately determine the 1st, 2nd, and 3rd fastest cars in the correct order regardless of variances in lane speed?

No head-to-head method can make that claim. Accuracy is a relative term. The real question should be which ones will do so with the most accuracy and, if there is inequity, will distribute inequity fairly.

I'm assuming you mark each heat 1st, 2nd, and 3rd, (3-lane track) assign points, and determine the point totals at the end.

That is the preferred method, and yields the most accurate results. However, most folks without electronics have difficulty calling all three places, so they assign 2 point for 1st (including ties for 1st) and 1 points for other places.

[dknowles67]

A timing system doesn't have to be an expensive purchase. If someone can do a some soldering and some carpentry, you can build a system for $50 or less.

Intriguing.
I have an electrical engineering background, and a soldering gun (a dangerous combination). I also enjoy carpentry, and have an adequate woodshop. I could hit the pack up to pay for supplies, and do the labor myself. Even if I had to spring for the $50, it might not be a bad proposition.

It looks like the $40 cheap kit has a $15 serial interface to hook to a computer.
Supposing I go that route, would you recommned racing in all 3 lanes, and averaging/summing the times?
Now I'm getting excited.

[Cory]

Supposing I go that route, would you recommned racing in all 3 lanes, and averaging/summing the times?

I just started a thread discussing cumulative time as a scoring method. Like any other method, it has its goods and its bads.

As an aside, if you have the resources, it might be worth focusing some of your attention on your track. If you visit http://pack146.nova.org/pinewood/res1999.html you'll see that we were able to make significant improvements in our old three lane track, just by the application of sandpaper and a new coat of paint.

[Stan Pope]

Supposing I go that route, would you recommned racing in all 3 lanes, and averaging/summing the times?
Now I'm getting excited.

If you have timings available, then your pack has to choose either timed or head-to-head racing. If lanes are not too different, then I would choose head-to-head, where everything is open to view by concerned spectators.

If the lanes are so different that lane assignment determines the heat winner, then you have a problem with head-to-head methods. Final standings methods like PPN or Stearns (with points per place in heats) will tend to yield a lot of ties in the final standings. You might have to fall back on some form of timed races.

If you go the timed race route, then you still have decisions to make. If you include all of the runs, then one bad run (caused by poor staging, perhaps) can move a car from first place to last place overall. So, you need to do something to avoid errors by an impartial adult staging the cars. You also need to look at ways timings can be inconsistent during the racing day. Cory has some comments on this in another thread and on his pack's web site.

BTW, watch out with that soldering gun ... today's components like 25 to 30 watt irons much better!

[gpraceman]

A timing system doesn't have to be an expensive purchase. If someone can do a some soldering and some carpentry, you can build a system for $50 or less.

Intriguing.
I have an electrical engineering background, and a soldering gun (a dangerous combination). I also enjoy carpentry, and have an adequate woodshop. I could hit the pack up to pay for supplies, and do the labor myself. Even if I had to spring for the $50, it might not be a bad proposition.

It looks like the $40 cheap kit has a $15 serial interface to hook to a computer.
Supposing I go that route, would you recommned racing in all 3 lanes, and averaging/summing the times?
Now I'm getting excited.

You should definitely use all 3 of your lanes. You can use times scoring (average, cumulative, median, or fastest) or simply use the timing system to determine your finish order and use points scoring. Which scoring method you use should factor in considerations like how reliable is your start gate (or operator), how stable is your track (from being bumped out of kilter) and which scheduling method you use. There are other topics on this forum that you could search for to find out the ins and outs of each.

The "Cheap Kit" does have the serial interface, so you can use it with computer software like HyperTerminal or one of the commercial software packages that support Fast Track timers. For $15 extra bucks, when compared to the "Son of a Cheap Kit", it is well worth the little extra to have that capability. Both of these will work on tracks up to 4 lanes.

For an enclosure, there is an example and photos at http://grandprix-race-central.com/sensor_system.htm, near the bottom of the page, that would do well with a system that you build or just assemble. One of these days I'll give it its own page.

[Darin McGrew]

I am proficient in computers, and writing software, so I could easily cook up something to determine race/lane assignments at check in time, but I'm not sure what the best method to use would be.
(We typically don't pre-register for the races, its just whoever shows up on race day).

In the last couple years, we've moved registration from Saturday (just before the race) to Wednesday (three days before the race). We never had any problem generating race schedules (using a PPN-like system) just before the race. Even on an old 386 laptop, it takes just a few seconds to generate a schedule.

The main advantage of the change is that we can conduct the design judging at our leisure anytime Thursday or Friday. The judges are no longer pressured to hurry up by the throngs of impatient kids pounding on the door.

[MathGuy]

There is the tier method that I am going to try (see other post). It can work for a 3 lane track as well. Since this is for next year you can see how it goes.

The goal is the following:
1. Have every boy race often.
2. Have even the slower boys Win a race.
3. Sort out to the top boys. (note this is listed as 3rd goal)

Given that you have 20 boys, you would split the participants into seven tiers. Since 3 doesn't go into 20 evenly, I would suggest one very slow cub master "turtle" car. If you had 18 or 21 scouts, I would still recomend racing 3 of these "turtles or troll" cars so that the slowest boy don't experience coming in 3rd every time.

You place the boys into 7 tiers by regristration order initially A,B,C,D,E,F,G, and race one round. Then based on the finish you determine the second round race. The winner of A tiers, stays, to race the winners of B and C tiers. The losers move also. I configured this movement chart so that winners move up more than one tier, this accelarates the sorting. The next round scheduling should be done instantly, so there is no delay when the next round comes. It can be done by placing the cars in boxes that are labled, or someone sorting index cards based on the finish of the race. The chart below would be the ordering for the next round.
Tier Racers (Finish goes to what Tier next)

T 1st 2nd 3rd
A: A B C
B: A C D
C: A D E
D: B E F
E: B E G
F: C F G
G: D F G


Thus the turtle car in G stays in the bottom tier for the whole race, A might be king of the hill for a while. The chart can be modified for different number of tiers. You might want to end using a 4 car PPN chart based on the final round Top 2 from A and Winners from B and C. Given 7 tiers, lets be conservative in that you race in round in 10 minutes and 10 minutes for the 4 car PPN. That would mean 5 rounds and 5 races for each kid, with distribution of wins being more evenly distributed among the kids.

Again, my goal is for everyone to have fun, not just for the top 4 cars to figure out who's fastest, and I think these goals are very appropriate at a pack level. Stan's multi-elimination charts are supposed to do a good job on this too.

[dknowles67]

Okay, it's 10 months, and 3 hurricanes later, and still no action on the part of the pack, or mine for that matter. Everytime it got brought up, it got pushed out. Now it's do or die, the race day is set for Jan 9th, and there is no money in the budget, or time to order/assemble a new track/electronic timers.
We still have 3 lanes, no electronics, same as last year.
We still have 4 races, 1 for each age group (Tiger, Wolf, Bear, Webelo), and an overall race for the four 1st place finishers.
We still have an hour per age group, and about 10-25 boys in each age group. 1st, 2nd, and 3rd place awards are given out.
We will have no pre-registration - whoever shows up race day races.
In the past we have always done Double Elimination .
(Don't ask me how they determine 3rd place - that's a sore point)
The emphasis is on what requires the least effort for the race organizers.
(I'm not a race organizer, but have volunteered to help out this year)
I'm willing to do whatever it takes to make the races more fun/more fair for the boys, but suspect I will get stopped if it costs money, or requires extra effort on anyone else's part. From reading past posts it looks like maybe triple elimination may be only way to improve, within the contraints I have, and I don't think that will go over well, since it will take longer than DE. I'm willing to bring a PC, and run the software to do the PPN scheduling, if I can get the Pack to allow it.
Will that be a significant improvement over DE, and still allow us to finish in the 1 hour time period?
If not, I'll have to wait until next year (2006), when there's a chance we might get some electronics.
Any other suggestions?

[Stan Pope]

My first reaction is to run a method that only requires identifying 1st place and can run three cars at a time. Because of time constraint, I'd run something that takes minimum record keeping and minimum time getting the boys and cars lined up to race.

Triple Elim / No Chart runs will on 2 or 3 lanes.
3 lnesa:
25 racers requires 7 rounds, 46 heats plus finals.
10 racers requires 6 rounds, 24 heats plus finals.

Quad elim:
25 racers requires 8 rounds, 61 heats plus finals. <- Too many heats!
10 racers requires 7 rounds, 31 heats plus finals.

These are based on calculator script at http://members.aol.com/standcmr/nelimsim.html

No Chart procedures are at http://members.aol.com/standcmr/nelim.html

You can run with NO Record Keeping by sending the boys to sit/stand with a group leader based on heat result; For large groups (40 or more), I have the boys wear a record card on which their current group is written/stamped by a score keeper near the finish line.