Tanyao vs Yakuhai?

After a certain tournament in Europe, I stayed behind and played a few games with some other people who also stayed behind. I asked them if they wanted to play with open tanyao, since I hadn’t had a chance to do so for the entire day (given that the tournament was an EMA one), and I would like to play something I’m more familiar with. Two of them said okay, but when the third player joined she exclaimed, “Oh no, not open tanyao!” Or something to that effect. This made me curious and prompted me to ask why so many disliked open tanyao. I got a lot of answers, but most said the same thing: “It’s too fast” or “it’s too cheap” or “it’s too easy!”

Given that we have an amazing tool for data mining available, I figured I might as well use it. So, with the help of Umai I got my hands on some data, sadly not presented in graphs (because graphs are awesome that way). The rest of this post will detail the specific numbers regarding Open Tanyao (and open hands in general).

The first argument against Open Tanyao is: “European players aren’t good enough to handle Open Tanyao”. In my mind, however, practice makes perfect. You can’t get better at using Open Tanyao if you’re not allowed to use it.

For the second argument I’ll quote GRDavies (Dutch champion, if I’m not mistaken): 1) It’s relative easy and thus less skilled players tend to play that hand every time (from turn 1). Comparing to another easy Yaku, the Yaku pung (Dragon/Seat-/Roundwind), which you – most of the time – must be lucky to be paired up in the early stage of the game since the tendency is that those tiles are discarded early in the games, you have 84 of the 136 tiles (~61,8 %) with Tanyo.

I’m gonna bring maths into this, because he started it! What do we have? Our data mining comes from 19,515 game logs analyzed, where we compare Open Tanyao and Yakuhai (he did bring it up). We also look at open hands in general.

Out of 19,515 games, we have a total of 98,734 hands played. The average value of these hands is 4784.399. 77,029 (78.016%) contained at least one dora and 49,012 of the total amount (98,734) were open, or 49.640%.

Of the 98,734 hands, 23,265 (23.563%) contained tanyao (either closed or open).
18,656 (80.189%) had at least one dora.
14,477 (62.226%) of the total amount (23,265) were open.

Of the 98,734 hands, 33,775 (34.208%) contained yakuhai (either closed or open).
20,884 (61.832%) had at least one dora.
30,108 (89.142%) of the total amount (33,775) were open.

Now, since you don’t get any han for open tanyao in EMA (rather, open tanyao isn’t a valid yaku), we looked at the average hand values for open hands. This means that any hand scoring for open tanyao was discarded, and we looked only at with or without dora.

34,535 hands were open, average value was 3758.823 points
14,119 hands had no dora, average value was 2310.014 points
20,416 hands had one or more dora, average value was 4760.769

Since we’re comparing tanyao and yakuhai, we also need to look at all hands that include tanyao (open and closed).

Closed hand, no dora: 2651.660 points
Closed hand, dora: 6189.749 points
Open hand, no dora: 2071.485 points
Open hand, dora: 4533.364 points

Now we’ll look at the specific values for tanyao and yakuhai. The two hands are extremely different, according to GRDavies, so we should see rather different numbers here.

Tanyao only:
Closed, no dora: 2606.336
Open, no dora: 1171.246
Closed, dora: 7566.990
Open, dora: 4100.922

Yakuhai only:
Closed, no dora: 3441.277
Open, no dora: 2176.830
Closed, dora: 6972.059
Open, dora: 4570.784

Oh my. I’ll repeat what a lot of people say about open tanyao: it’s a skill-less, cheap hand. A lot of people are also irritated by it since people go for it all the time. Well, suffice to say, people go for an open yakuhai hand a lot more often. There are a few curious numbers in our results, so I’ll touch on them:

Tanyao will more often than not have a dora – remember that you have three static dora (red fives) as well as a dora that will be within the tanyao range more often than not (around 62% of the time).

I’m really quite tired of the open tanyao debate. I find that the people against open tanyao have no real arguments to support their position, and are only afraid of change itself. But personally I’ve moved on. Thanks to Umai for digging up the data though, I do love when numbers prove my points.

6 thoughts on “Tanyao vs Yakuhai?

  1. This is a very interesting post, but it’s hard to see the significance of the data since it’s unclear what it actually describes. Quantitative analysis is all very well and good, but without context it raises more questions than it answers. Here are some:

    Where did these games come from? Are the total numbers of hands winning hands or all hands? What happened to the hands that didn’t win? Are there any games played by Europeans in the data? Etc.

    “Well, suffice to say, people *go for* an open yakuhai hand a lot more often [than open tanyao].” (my emphasis)

    Where is the proof of this? It seems that it would be difficult to quantitatively prove how often people *go for* a certain yaku. Moreover, this just doesn’t make sense. In order to aim for a yakuhai hand, you actually have to have a pair of them in your hand. Most starting hands will have a good amount of simples in them, which will beginners might think makes them good candidates for open tanyao.

    Incidentally, I’m not arguing against EMA’s adoption of the rule. I just wish your arguments were a little more clear.

  2. You haven’t actually dealt with the issue head-on at any point though. All you’ve indicated is that Yakuhai are as good (if not better) a target for reducing quick, cheap hands (though actually theres no indication in your statistics how quick open tanyao vs yakuhai is).

    You haven’t indicated in any way that the game won’t improve by closing Tanyao (assuming one’s standard for “improvement” is the reduction of quick, cheap hands), even if protesters of Open Tanyao are hypocrites for not also changing Yakuhai.

  3. Benjamin, I started out trying to deal with the tanyao issue, but quickly got bored of it. As it was, I decided to at least release the numbers to the public. As for source, http://arcturus.su/tenhou/gamelogs/ would be where you should look – it’s logs of the phoenix tables on tenhou (afaik). Europeans I’ve no idea about, since there’s no list on where players are from on tenhou.

    Total number of hands are winning hands only, I doubt you could write a program that searched through logs and could tell what a user was aiming for on its own.

    Krill, what you’re saying is absolutely true. I basically wanted to point out the bullshit argument I kept hearing that open tanyao was somehow cheaper or easier or more often played than yakuhai. It could, obviously, be that open tanyao fails more often than yakuhai (and I wouldn’t be entirely surprised if that is actually true, since it’s a lot easier to defend against). Anyway, I’m bored of banging my head against the wall and wanted to get the data we had out there.

    And now I’m late.

  4. Thanks for putting that data out there. It’s always interesting to look at these sorts of things.

    In truth, I don’t think you even need data to argue against the “too fast”, “too cheap” and “too easy” arguments, myself. If you’re an experienced person playing against someone who plays tanyao every time, for a thousand or two every time they win, you still shouldn’t lose in a full hanchan most of the time. And if you are, you’re doing something horribly wrong.

  5. >In truth, I don’t think you even need data to argue against the “too fast”, “too cheap” and “too easy” arguments, myself. If you’re an experienced person playing against someone who plays tanyao every time, for a thousand or two every time they win, you still shouldn’t lose in a full hanchan most of the time. And if you are, you’re doing something horribly wrong.

    Which I find to be exactly the case. I can’t find a better word to describe losing your semi-concealed dai san gen to a kuitan than just “cruel,” buuut, however, a player going for a kuitan every single hand shouldn’t be an obstacle impossible to overcome.

    The problem more often than not, would be assembling your comeback hand in the last few rounds, and have the player in fourth place kill the game with a no-dora nakitan. But it’s just the same as if the guy in first just killed it with a yakuhai.

    So I don’t really think open tan yao cheapens the game either; only cheap players cheapen it.

Leave a Reply

Your email address will not be published.