🏅 Top traders

Someone explain to me what I've misunderstood about the D1 retention metric. How on earth can it be higher on the 9th than on the 8th (with the 8th being well below normal)? That makes no sense at all.

I calculated it myself and got 53%, hence the confident bets as it looked like that's where it was headed. Always a chance I coded a bug of course (happens often), but the official numbers are not what I would expect either.

Might be getting ahead of myself but @jacksonpolack if there's literally a bug in Manifold's code calculating the D1 metric, would you consider waiting for a fix to resolve?

Either way I'll go take a look at Manifold's code, either a find a bug or realise why the metric isn't what I thought it was.

@chrisjbillington There was a bug and I've submitted a pull request:

https://github.com/manifoldmarkets/manifold/pull/2062

Fix incorrect denominator on d1 stat by chrisjbillington · Pull Request #2062 · manifoldmarkets/manifold

I notice the d1 stat on https://manifold.markets/stats makes no sense in the context of recent traffic. It dips on the 8th when the NYT article was released, and then spikes back up the day after. Clearly we expect it to be normal-ish on the 8th and dip on the 9th. The problem seems to simply be using the wrong denominator. We want retained DAUs (number active yesterday who were also active today) as a fraction of yesterday's DAUs, not today's. (I haven't tested this, I wouldn't know how)

yeah i'll wait for that

@jacksonpolack Sweet, they've merged it, I'm guessing the next stats update after midnight PST should have the fix.

@jacksonpolack Fix is live at https://manifold.markets/stats, D1 retention was 52%.