Predictions Contest - mean, median or mode?

[Born Racer]

crmalcolm requested the following last year:

@BornRacer - I know the rules state that entrants in the contest have their votes removed from the polls for the purposes of this contest. Is there scope for a mean or median calculation to be made as opposed to the mode in future years? If I look at this race (and not exclude the entrants here) the mode is 6, but the mean is actually 4.58 and the median is 5 which seem to be a better reflection on the event's rating?

Thus far, the mode has been applied. For 2022, I'd like to ask you to please vote on whether you'd prefer the mean, median or mode to be used? If the mean was used, it would then be rounded up or down as appropriate for the result. Please see the 2021 regulations to get an idea of how it's worked so far: https://tentenths.com/forum/showthread.php?t=156040

Anyone is welcome to vote, but I will implement the most popular choice from those who participated in the 2021 Predictions Contest.

[BTCC frog]

I like it how it is with the mode. In Abu Dhabi I predicted a rating of 1 and was right, but that kind of thing will almost never happen with the others where it is more likely to be somewhere in the middle. So I would choose the mode as it allows freak predictions to be right more often.

[TrapezeArtist]

BTCC Frog has a good point. On the other hand, use of the 'mode' could make for some strange results. Imagine a race that evokes strong reactions one way or the other. Let's say the poll result is:
10: 3
9: 9
8: 6
7: 2
6: 2
5: 3
4: 3
3: 8
2: 8
1: 5
Using 'mode', the score is 9. A small variation in the voting might have made the score 2 or 3. With 'median' the score would be 4.9, rounded to 5, which would be much more representative of the voting.

If using 'mode', what do you do if two scores are equal? Taking the above example, if 2 gets 9 votes instead of 8.

[crmalcolm]

Originally Posted by TrapezeArtist

If using 'mode', what do you do if two scores are equal? Taking the above example, if 2 gets 9 votes instead of 8.

Reading the previous regulations, if I am understanding correctly, then the points would be awarded to anyone predicting either score that received 9 votes.

[Adam43]

Originally Posted by BTCC frog

I like it how it is with the mode. In Abu Dhabi I predicted a rating of 1 and was right, but that kind of thing will almost never happen with the others where it is more likely to be somewhere in the middle. So I would choose the mode as it allows freak predictions to be right more often.

For this reason I think mode works best.

Originally Posted by crmalcolm

Reading the previous regulations, if I am understanding correctly, then the points would be awarded to anyone predicting either score that received 9 votes.

An option would be to use mode, but if more than one chose the closes to the mean? Or the one closest to the second highest total. Or with the most votes within one?

But I quite like awarding both.

[Born Racer]

Originally Posted by Adam43

For this reason I think mode works best.


An option would be to use mode, but if more than one chose the closes to the mean? Or the one closest to the second highest total. Or with the most votes within one?

But I quite like awarding both.

I'm not sure if this will work well, but we're going to trial it for the duration of this season - we'll use the mean and the mode. For instance, Abu Dhabi got a mode of 1, but (and I just did a quick calculation and didn't exclude the predictions winners as I would as per the rules), but it also got a mean of 5.1. I would round this down to 5.