Talk:270edo

TE logflat badness

Sequence of decreasing 13-limit TE logflat badness for GPVs from 5 to 196608:

• 9, 58, 72, 270, 96478

I feel this demonstrates perfectly why TE logflat badness isn't a very relevant metric for equal temps.

—FloraC (talk) 13:01, 12 April 2026 (UTC)

Why not? I genuinely don't see your point. --Eufalesio (talk) 17:13, 12 April 2026 (UTC)

It's simply not a good sequence. The sparsity. The presence of 9. The leap from 9 to 58. None of them makes sense. —FloraC (talk) 19:56, 12 April 2026 (UTC)

Surprised you out of all people think it doesn't make sense... Based on the little understanding I have about logflat badness, it makes perfect sense to me. The way I see it: TE logflat badness punishes complexity exponentially, and since higher ranks get comparatively less badness, you get the gravelly and innacurate edo scoring way higher than many other better edos because it's extremely simple and does a "relatively good" job of crudely representing the 13-limit with minimal complexity albeit very high error. Barring the coarse edos, TE logflat badness is a metric that feels stronger than relative error alone. That jump from 270 to 96478 to me says that there is nothing as GOOD for the 13-limit as 270edo, up until insane edos. If it is denoted in absurdly fine edos like 4296, it should be denoted in much coarser - and useful edos. --Eufalesio (talk) 20:19, 12 April 2026 (UTC)

I know how it works in theory. I still don't think it's a good sequence. Mainly it's hardly informative when you have only five or so entires in the sequence. —FloraC (talk) 20:58, 12 April 2026 (UTC)