I had suspicions early on. M1 to Saint in 190 hanchan? -- come on, kid. He tried to cover his tracks, but rumors circulated. Jane got him in the end.

I was tipped off by something else. During the ascent, his win rate was 22.6% (74th percentile) and his deal-in rate was 8.7% (1st percentile). The delta of 13.9% is impressive - but to be in the bottom 1% of deal-in rate and the top 26% of win-rate simultaneously is inhuman, since, among good players, the two stats have a strong positive correlation. In other words, low deal-in rate players are more likely to have low win rates, and high deal-in rate players are more likely to have higher win rates (well, assuming they're not a looper).

Seeing the cat get his comeuppance validated my stat-based suspicions. It also made me wonder if mortalbotters could be distinguished via literally just these two stats. Is there a cluster of mortalbotters roaming Jade with extremely low deal-in rates and unusually high win rates (at least, relative to the rest of the Jade Room players)?

Well, let's see. I scraped a few thousand profiles for Jade Hanchan players with >100 games and ended up with >1000 entries for each rank of M2-S3. Mortalbots should show up on an X-Y scatterplot of win rate vs. deal-in rate as a small smear in the bottom right hand corner (high win rate, low deal-in rate).

Hey, look at that. It's not obvious in the M2/3s, but among S1/2/3s, there's a subpopulation centered around (24, 10) distinct from the larger cohort centered around (22, 12). A classifier places ~5% of the Saint players into this subpopulation (represented in blue). Unsurprisingly, these extraordinarily high-win:low-deal-in players tend to have high net win efficiencies (win rate*average win score - deal-in rate*average deal in score) and great placements (high 1st rate, low 4th rate).

Are they bots? In a luck-based game, you can't be certain. Glancing through Amae profiles from a dozen or so suggests most are (based on queue behavior). Maybe some additional investigation is in order.