Tiebreaker Scenario

[gpraceman]

I've been trying to think of ways to break a tie without having to do any additional racing.

Here's a hierarchical set of tiebreaker rules which I came up with. I would only employ these rules if a trophy was at stake.

Times
1. Head to head results
2. Most victories
3. Best results against common competitors
4. Best average finish order
5. Best cumulative results (if not scoring by cumulative time already)
6. Single fastest time (if not scoring by single fastest time already)

Points
1. Head to head results
2. Most victories
3. Best results against common competitors
4. Best cumulative results (if a timer used)
5. Single fastest time (if a timer used)

Doing a tiebreaker round is still an option though. There is still a chance that a tie would exist after applying these tiebreaker rules and some people would prefer to do the head-to-head runoff.

Let me know what you think. Did I miss anything?

[Stan Pope]

Times
...
7. Most consistent time
8. Coin toss

Points
...
6. Most consistent place
7. Coin toss

[gpraceman]

Times
...
7. Most consistent time
8. Coin toss

Points
...
6. Most consistent place
7. Coin toss

Coin toss - I don't know that a coin toss will sit well with anyone. I'd rather do a tiebreaker round if these rules fail to break the tie.

Most consistent time - Pretty much the same as average time, since a racer is not likely to have more than one of the same time.

Most consistent place - Possibly. This may likely be the same as using a count of the most victories since my scenario is that a trophy is at stake.

[Stan Pope]

Coin toss - I don't know that a coin toss will sit well with anyone. I'd rather do a tiebreaker round if these rules fail to break the tie.

I'd put it at the very end of the list of criteria ... if they come out even in all other respects, but you have to make a choice, then ...

Most consistent time - Pretty much the same as average time, since a racer is not likely to have more than one of the same time.

I'm thinking more like take standard deviation of each tied racer's run times, reward the smallest s.d.

Most consistent place - Possibly. This may likely be the same as using a count of the most victories since my scenario is that a trophy is at stake.

Maybe, maybe not. I've mostly had to break ties like 3rd-5th.

Of course, I have settled this trophy problem by buying another 1st place trophy ... the track just wasn't good enough to differentiate the cars. And, I like that solution better than a coin toss.

BTW, I like a two way "race-off" as follows:
Racers alternate lanes. First to win 2 heats in a row wins the tie break.
Set an arbitrary limit on the total number of heats to run... maybe 8 or 10. Then call it a draw or try next paragraph!

Here is a 2-way tie break that I have analyzed but never tried, applicable when no timer is available and cars are too equal for the method in the last paragraph:
Racers run twice, alternating lanes.
Judges (3 at least) mark a paper with the winning margin and the winner's ID. ("Mark" could be a horizontal line of length equal to winning margin or two vertical lines separated by winning margin.)
Divide the papers into stacks by ID.
Add up (or lay "end to end") the winning margins as drawn by the judges.
Biggest (longest) wins.
(Might be better to use 5 judges, collect papers after each heat and discard the high and low.)

[Cory]

I don't want to sound harsh, but I wouldn't use any criteria where the results are potentially skewed by lane draws or opponent draws. Unfortunately, this covers a lot of the items in your lists.

Even with CPN there is potential skewing that is more likely than some people realize. And the kind of skewing that can occur with anything less than a CPN chart would be noticeable by many people.

I always looked at race-offs as a good, exciting thing, not something to be avoided. Why are you thinking about eliminating them?

[Mr. Slick]

My solution for ties where a trophy is involved is to recognize both and order another trophy!

Ties do exist and can be legit. The crowds are always excited when the FastTrack displays a tie during the racing, 4 times so far this season. . . this is about .05% of the races.

Once again, recognize both and order a second trophy FOR THE KIDS!

[gpraceman]

I always looked at race-offs as a good, exciting thing, not something to be avoided. Why are you thinking about eliminating them?

It is not that I want to eliminate doing a tiebreaker round, but many GPRM users requested having the ability to easily break a tie. This can be done with a tiebreaker round or possibly by applying a set of tiebreaker rules, which is common in many sports.

I don't want to sound harsh, but I wouldn't use any criteria where the results are potentially skewed by lane draws or opponent draws. Unfortunately, this covers a lot of the items in your lists.

There are pros and cons definitely, as is true with other topics discussed here like points vs. time, different scheduling methods, etc. I thought it would be good to explore the subject and see if, amongst the members of this board, there was an agreeable set of tiebreaker rules to apply without requiring further racing.

With times scoring, it isn't so much of an issue but it is with points scoring.

[gpraceman]

My solution for ties where a trophy is involved is to recognize both and order another trophy!

I like this option myself, but not all organizations want to deal with the cost and trouble to get extra trophies.

Ties do exist and can be legit. The crowds are always excited when the FastTrack displays a tie during the racing, 4 times so far this season. . . this is about .05% of the races.

I am not at all worried about a tie in any one heat, since it is even less likely to contribute to a tie in the overall standings. I am looking at the final overall standings which determine the trophy order.

[Cory]

It is not that I want to eliminate doing a tiebreaker round, but many GPRM users requested having the ability to easily break a tie. This can be done with a tiebreaker round or possibly by applying a set of tiebreaker rules, which is common in many sports.

It is a reasonable thing to ask for, I guess.

If it were totally up to me, though, I would eliminate items 1 thru 4 from your first list, and items 1 thru 3 from your second list.

One measure that could be used for points scoring is strength-of-schedule, provided you are using a PPN-or-worse chart. For example, suppose car 1 and car 2 both scored the same number of points, but car 1's opponents scored more points than car 2's opponents. Then car 1 deserves to advance to the trophy round, because it scored an equal number of points against tougher competition.

[gpraceman]

If it were totally up to me, though, I would eliminate items 1 thru 4 from your first list, and items 1 thru 3 from your second list.

Why would you eliminate "Best results against common competitors"? I don't see where there could be a bias in those results. It seems like a compliment to strength-of-schedule.

One measure that could be used for points scoring is strength-of-schedule, provided you are using a PPN-or-worse chart. For example, suppose car 1 and car 2 both scored the same number of points, but car 1's opponents scored more points than car 2's opponents. Then car 1 deserves to advance to the trophy round, because it scored an equal number of points against tougher competition.

I do like that option. It may be adequate in of itself to break the majority of ties whether scoring by times or points.

One tiebreaker option that I forgot for times scoring would be "Total points". Even when scoring by times, points can be awarded based on finish order. GPRM does this already in case a user wants to switch back and forth between times and points scoring (like if the timer goes belly up).

[Stan Pope]

One measure that could be used for points scoring is strength-of-schedule, provided you are using a PPN-or-worse chart.

I prefer not to trophy-rank by points when using a PPN-or-worse chart. Strength of schedule can invert some standings.

Instead of using a tie-break for those rankings, I'd rather pull the top approx 2XTrophy places into a PN or CPN final.

Or was your point that tie break be used in that selection of finalists, since point ties grow increasingly likely that far down the chart?

[gpraceman]

One measure that could be used for points scoring is strength-of-schedule, provided you are using a PPN-or-worse chart.

...

Or was your point that tie break be used in that selection of finalists, since point ties grow increasingly likely that far down the chart?

Personally, I would rather advance tied racers within the top N to the trophy round and only worry about a tiebreaker to decide the trophies. This may mean not being able to get a CPN chart for the finals though.

[Cory]

Why would you eliminate "Best results against common competitors"?

As a simple example, suppose Car 1 and Car 2 had only one common competitor, Car 3. Car 1 might have had the best lane when it raced Car 3. Car 2 might have had the worst lane when it raced Car 3. If those three cars were more evenly matched than the lanes, then you have the sort of anomoly that astute parents notice.

I am a little suprised that you would not axe "Single fastest time" since there could be lane draw issues related to that.

If everyone races the same number of times in each lane, then I wouldn't axe it. If not, then, yes, you're right, it probably gets chopped also.

[Cory]

Or was your point that tie break be used in that selection of finalists, since point ties grow increasingly likely that far down the chart?

Yes. I like the excitement and accuracy of race-offs, so I'm not in favor of automatic tiebreakers for any purpose. But given a choice between using them for picking finalists versus using them for awarding trophies, I'd use them for picking finalists.

[gpraceman]

As a simple example, suppose Car 1 and Car 2 had only one common competitor, Car 3. Car 1 might have had the best lane when it raced Car 3. Car 2 might have had the worst lane when it raced Car 3. If those three cars were more evenly matched than the lanes, then you have the sort of anomoly that astute parents notice.

Yes, that would be the worst case scenario. This problem could be lessened if there was a minimum number of common opponents in order to apply this rule. Maybe common opponents >= number of lanes.

If everyone races the same number of times in each lane, then I wouldn't axe it. If not, then, yes, you're right, it probably gets chopped also.

Looks like I wasn't fast enough deleting this from my post. I did eventually realize that with everyone racing in each lane, then lane draw is not an issue.