Analysis - Taatsu Otoshi

While reading zeRo's book, I came across a chapter talking about taatsu otoshi (when people drop a two tile shape from their hand) and how it indicates the speed of their hand.

I am generally quite skeptical of anything related to discard reading. So, I decided to do a quick analysis on them to see if the data gathered matched zeRo's rankings, rather than blindly believe it.

There are a few caveats here.

• This is Houou room data. It may be more consistent in Tokujou, and is probably less consistent in Joukyuu or lower rooms.
• I only looked at quiet games. If anyone called riichi or a tile, I stopped looking at that game. This likely doesn't affect the early data much, but could affect the later data quite a bit. The Row 3 data is quite thin, with many of the rows only having several thousand cases.
• This data was gathered continuously, so if someone cut a run from their hand, like 2 3 4, then it would be counted as both 2->3 and 3->4.
• I didn't separate out the yakuhai winds from the guest winds, so the "wind pair" row is a combination of both. Guest winds should be safer, and yakuhai winds probably similar to dragons.
• Karagiri is counted as tsumogiri in the code, but looks like tedashi in reality.

Anyway, onto the data. The first section of the spreadsheets will be discards from different suits (eg 2/8->middle means they discarded a 2/8 from one suit, then a 3~7 of another suit) as a sort of control sample.

Row One

This early on, people could be cutting bad shapes or shifting towards yaku. Cutting middling kanchans or ryanmen could be aiming towards honitsu or chanta or similar.

Row Two

Here's where I would expect them to have the most weight. It's starting to be too late to mess around with your hand too much, and people will be consistently nearing tenpai.

I think the outside kanchans are a bit interesting. When discarded from the outside, like 1->3, they are more likely to have a further developed hand. But, when discarded from the inside, like 3->1, they are more likely to have a wider wait. Discarding 3->1 could also just be to keep the safer tile around. I think I'll just ignore the suggestion about reverse discards being dangerous.

Pair drops also don't seem to be too dangerous here, but maybe it changes when open hands are included. Maybe I'll do another analysis that only tracks tedashis after a call.

Row Three

Here, people will be warier of others being in tenpai. They might play safer, or they might risk getting tenpai before the round ends. Since all this data is from quiet games, with no calls or riichis, it is probably a bit niche, but included for completion.

The dragon pair drops become quite safe in the third row according to this. That's probably a folding move. Also, the colours are relative to other discards in their row. Even though the dragon pair is very green here, it's still more dangerous than anything from the Row 1 sheet.

Bonus Row 2 Data

Here's the rest of the options. The only particularly interesting thing here to me is that discarding consecutive terminals from two different suits is a sign of slowness. That's probably intuitive, though.

You can see all the data in this spreadsheet.