Originally posted by MacGuffin
View Post
Announcement
Collapse
No announcement yet.
FGCOTY 2017  Results!
Collapse
X

I have a lot of data on all of you now, and I plan on abusing it
I was thinking about how the votes of different people correlate between each other, and I made a quick script* to run through the votes and do some matchmaking. It is Valentine's Day, after all!
Here's the full table of vote correlations (names are in alphabetic order), as well as the general "antisocialness" factor, which is just a sum of all correlations.
The person representing the average FG is SirEdward, with 0% total correlation over all voters! Congrats!
The worst is tcdoe at 147%, Lars at 145% and DFence at 136%. All of them submitted just 3 cars for the vote, which affects this result **
The least compatible pairs are DFence  samulis, Dr Grip  leviathan and IceBone  tcdoe.
And the best matches are:
Galantti  IceBone at 10.4 %
gt1750  TobiasG at 10.8 %
DanRoM  Matt2000 at 11.2 %
* A slightly modified Kendall Tau correlation. Basically going through every possible pair of cars, and comparing the order in which they appear in each person's vote ranking. If order is the same the score goes up, if order is reverse the score goes down, if order is unknown (both cars in the pair is missing from one of the voter) the score stays the same. Basically simulates a quiz where you're forced to pick a favorite between two cars for all (14choose2) possible combinations, and comparing matching results.
** The method emphasizes difference between the amount of cars submitted. Which makes sense, because votes generally have a trend, and by omitting part of the vote, you're missing out on gaining correlation points from certain "obvious" choises. So every row and every column is consistent within itself, but not necessarily comparable to other rows and columns.
 2 likes
Comment

Originally posted by MXM View Postgt1750  TobiasG at 10.8 %
I don't mean to sound bitter, cold, or cruel, but I am, so that's how it comes out.
I reject your reality and substitute my own!
Why is Denis Leary a star while Bill Hicks is unknown? Because there's no cure for cancer.
Racing is the only real sport. Everything else is just a game.
In engineering terms, you really can’t have "brilliant, apart from ..." I mean, the Titanic was brilliant, apart from its crappy rudder
Comment

Originally posted by TobiasG View PostThat was a fantastic resault! Much better than expected and so close to the no 1 spot! Should have bougth a Q
Originally posted by DaBoom View Post
Same with Audi bros.Last edited by MacGuffin; February 15th, 2018, 6:26 AM.“If Liberty means anything at all, it means the right to tell people what they do not want to hear" (George Orwell)
Comment

This is how I implemented it, say I'm comparing your vote with Matt2000's vote:
1. Make a list of all car pairs. Car a vs car b. Like: Lexus vs Disco, Lexus vs 156, Lexus vs Tesla... then Disco vs 156, Disco vs Tesla... etc. That's what "14 choose 2" operator does. You get 91 combinations of car pairs.
2. For each car pair, check whether you prefer a or b. So get a list like: a a a b a b b, except it's 91 letters long.
3. Do the same for Matt's vote, say the list for him is b a a b b b b.
4. If you both have the same letter for a pair of cars, you get a score +1 (concordant pair), if you have a different letter you get 1 (discordant pair). For the 7 pair example above the score is 52=3. The percentage is score/amount of pairs, so 3/7 = 43%.
5. Repeat steps 24 for every voter pair to fill the table above.
Why do it like this? Because it simulates the decisionmaking when ordering a list of items. It shows similarity between two people's decisions. It does not care about absolute position of the car in the list. Say you have Disco and Lexus placed 5 and 11, but Matt has them 1 and 3. That's still a positive correlation.
Could be done better? Yes, I could've assigned weights to the distance between pairs, or make it "top heavy".
https://en.wikipedia.org/wiki/Kendal...on_coefficient
My "modifications" to the algorithm are about dealing with incomplete lists, where I assume that every missing car has lower placement than cars in the list. Except if both cars in a pair are missing, then I can't make a decision and the score remains the same.
The score in the last column is just a sum of all scores on one line, but it's not very representative, as it favours people who sent a more complete list.Last edited by MXM; February 15th, 2018, 6:38 AM.
 2 likes
Comment

I hadn't read the calculation until now, but my only response is:
I was, and still am, terrible at statistical maths.Last edited by Matt2000; February 16th, 2018, 8:40 AM.
Comment

Originally posted by Matt2000 View PostI hadn't read the calculation until now, but my only response is:
I was, and still am, terrible at statistical maths.
Before I got my Msc. I was useless, now I am merely lousy, and I appreciate people able to do that.?
"I'm sure Burger King is doing great in Germany, but they're not exactly stopping making schnitzels."
Comment

Sorry to grave dig this thread, but thank you guys! I had no idea this vote was on again, but I'm amazed I've won it for a second time. I must be more active around here, this completely slipped me by.
Thank you allPhotobucket can sod themselves
'04 Fiesta '06 & '03 Mondeos '65 Anglia '29 Austin 7'36 Standard 12'86 Fiesta '69 Corsair
'31 Rover Two Litre
Comment

Also gravedigging, I didn't even know this was a thing! And I got 5th place, woot!
Bikes:Honda VFR 800 Honda 919
Ducati Multistrada 1200 S Sport
Cars:Hyundai Genesis 3.8 Jaguar XJR 4.0
Broken:Ford Bronco 4.9 4x4
Comment
Comment