Talk:Contest: Haunted house

Tie-breaker

"That resulted in tie with simple vote counts, some kind of weighted calculation was planned for this situation."

"Here we have votes log10(xp) weighted ..."

My 2 cents -

• Is XP a good measure of anything in that respect?

 It might not be best way to do it but currently it is best *usable* general indicator about how well established player is and running XP through log10 also makes XP mostly fair, 2 low XP votes can easily overcome 1 high XP vote.
 However there could also be XP cap to further reduce impact of very high XP votes, for example 100k cap could be used. 100k XP is pretty much well established player account.
 
 --SX (talk) 17:43, 15 November 2020 (UTC)SX

• But when is XP measured? At the time of submission (hardly doable I guess, but that could be the most fair), at the end of the event, or at the end of the voting period after the event ended? Or later, when result are assumed at an undetermined time? Just an example I don't recollect the number, but in that period of time I might have done around 50k XP overall... And I'm not the quickest in that domain ;-)

 Easy to collect when player votes first time or after voting ends. Collecting before voting starts is complicated.
 For current results XP data was collected when contest ended, however with method used (log10 of XP) results would not be different no matter when data is collected.
 Changing position requires either another vote or really huge difference in XP, like some millions, as long as player who voted are not fresh near zero XP accounts.
 
 --SX (talk) 17:43, 15 November 2020 (UTC)SX

• Not sure how the XP-based weighting is done here (without knowing which values were used and when), but does it mean that one with more XP would be favored? Shouldn't we favor the player with less XP = So that a newcomer arriving to a tie with an experienced player would get the benefit for his effort: he probably had less resources and still achieved something comparable to someone with possibly more stock and options?

 Let's go with lua example, XP based weighting was done like this: assuming voters = { gorlock = 98765, swissalps = 9345678 }
 for voter_name, voter_xp in pairs(voters) do score = score + math.log10(voter_xp) end
 
 This does favor newcomers by reducing impact of very high XP players. gorlock vote would be 4.994603068070779 and swissalps vote would be 6.970610813577633.
 Also XP weighting is by voters XP not by contestant XP so 1000000 XP contestant does not have any advantage over 0 XP contestant.
 Voter XP affects final results: 1000000 XP player adds +6 to score, 100000 XP adds +5 to score, 10000 XP adds +4 to score, 1000 XP add +3 to score, 100 XP adds +1 to score.
 If it would be by contestant XP then we probably should favor one with less XP but it is by voter XP and there we should favor established player and possibly drop below 1000 XP votes completely to reduce possible cheating by using multiple sock puppet accounts for voting.
 
 --SX (talk) 17:43, 15 November 2020 (UTC)SX

Gorlock (talk) 16:48, 15 November 2020 (UTC)