TagPro Region Wars X map vote

The results of the experimental Region Wars X map vote are in! Carbon has told me 9 maps will be played, so that is the number of maps picked. Attached are the 25 raw votes and the list of voters.

After playing around with preliminary results I found one weakness in my proposed system: the proportional representation. When electing n maps, 1/n votes is enough for a map to be picked even if all other people intensely hate it (e.g. Arena). Puzzling with variations of the system, I came up with a good solution that maintains the other good properties of the Single Transferable Vote (STV) system:

  1. Find the single best map with Instant-Runoff Vote (equivalent to STV with 1 seat).
  2. Remove this map from everyone's map ranking.
  3. Go back to step 1 to determine the next map.

This effectively lets everyone vote on every map and transfers votes until consensus for every map. It's like electing n representative presidents instead of one representative parliament with n members.

A side benefit is that it is easier to vary the number of maps, because the voting result is simply one overall ranking of all maps. Thus it is easy e.g. for ENLTP to have the same maps as ELTP but one map more. Also, because maps are picked one by one, any other constraints like a minimum number of new maps are a bit easier to apply. Finally, having an overall ranking of all maps allows fun comparisons between seasons.

Below are the full results for Approval Vote (AV, normal ELTP system), Meek's Single Transferable Vote (STV), and the new sequence of Instant-Runoff Votes (IRVs) which will form the official results (Carbon agreed). Note the sequence of IRVs is continued so you can see what the next maps would be if more than 9 would be elected. I also added the percentages for how often each map was a first choice or a (shared) last choice. Note that all methods involved some random tie-breaking; you can view detailed results by clicking the appropriate links. For the AV it is assumed that all non-shared-last-place maps are approved.

RankMapELTP statusFirst choiceLast choiceIRVs electedSTV electedAV elected
1stMarketcurrent20%40%
2ndPilotcurrent8%44%
3rdTransiliocurrent0%32%
4thQiocurrent4%60%
5thEMERALDcurrent4%40%
6thGeoKoalaold12%68%
7thAtomicold0%60%
8thConstrictionold12%64%
9thScorpiocurrent0%48%
10thWormyold0%52%
11thRushold4%60%
12thIRONold8%56%
13thTehuitzingocurrent4%68%
14thLong Islandnew8%76%
15thCedarnew0%72%
16thMild Highnew0%64%
17thSmirkold4%72%
18thCurbold0%72%
19thApparitionnew0%76%
20thArenanew12%84%
21stTrebuchetnew0%80%
22ndFloral Mappenew0%76%
23rdNirvananew0%88%
24thGoldienew0%84%

Overall, the new system does improve the voting incentives (rank as much as you can), gives us more details about in what order people like maps, and allows us to build an overall map ranking for the community with a more solid theoretical basis (runoffs for each position). New maps still face difficulty surfacing (even with proportional representation in the STV results), but less common maps like Qio, GeoKoala and Constriction do reach high positions in the ranking. For the ELTP map vote we are probably still going to need constraints on the number of returning and new maps.

Meanwhile, your maps for tomorrow are: Market, Pilot, Transilio, Qio, EMERALD, GeoKoala, Atomic, Constriction and Scorpio.

0 replies

Add a reply