I’m hoping to do an emitter-vs-color-temperature heat map of all votes with a matching name on both polls, to get more info about which specific emitter+color combinations are most popular. And since I can’t map anon votes between polls, they won’t be on the graph.
However, I have a tool to run the Condorcet algorithm locally, and here are the results so far with anon votes removed:
Color temperature:
4500K 4750K 5000K 4250K 4000K 5500K 3500K 3000K 6000K 6500K
4500K - 104 113 126 128 149 157 160 168 173
4750K 78 - 102 108 119 160 155 163 179 183
5000K 86 89 - 106 114 186 157 160 189 192
4250K 51 71 91 - 85 124 130 135 139 144
4000K 57 74 88 70 - 125 139 144 141 145
5500K 53 29 12 73 75 - 104 119 153 158
3500K 33 42 48 42 30 82 - 122 107 116
3000K 38 38 44 49 39 68 39 - 101 112
6000K 32 18 9 52 53 6 67 66 - 122
6500K 25 12 7 39 40 4 48 46 7 -
Emitter type:
Samsung LH351D Cree XP-L HI Nichia 219c Cree XP-G3 Cree XP-G2
Samsung LH351D - 126 153 181 188
Cree XP-L HI 120 - 152 233 236
Nichia 219c 77 98 - 154 162
Cree XP-G3 62 18 84 - 127
Cree XP-G2 53 14 71 76 -
Removing anons doesn’t change the color temperature rankings. It does flip the top two for emitter type though. In either case, the margin isn’t big. In one result, XP-L HI wins by 5 votes. In the other result, LH351D wins by 6 votes.
The main trends I see in the votes are:
- The top two emitter types are pretty evenly split, so we should probably try to get Lumintop to make both.
- Tint for each emitter type should be 4000K to 5000K, but the details need a closer look.
BTW, it’s worth mentioning that the polling site does at least try to prevent ballot stuffing. It’s not very hard to get around that, with a bit of internet trickery, but it at least doesn’t let people do it easily. And at a glance, it doesn’t look like that has happened, even among anon votes. Like, for color temperature, there are currently 119 anon votes, of which there are 106 different unique ballots. And half of the non-unique ones just have “10” in every cell, no preference expressed, so they don’t even affect the result.