by Elynnaad

With the announcement of Doman Mahjong (identical to the real-world equivalent of Japanese Mahjong, or Riichi Mahjong), if you’re intrigued by the game in the slightest and saw the threads when it was announced, you may have come across this video. This is a comprehensive run-down of all the rules of the game, including scoring and real-world play – you don’t have to worry about penalties and things thankfully as almost no digital version of the game will let you, but everything is covered regardless. It has been recommended to play the video on 1.5x or double speed, but there’s no real guide out there much better than this one on how to play.

Beyond that though, how do you win, and potentially more importantly, climb ranking? I’m going to go a little bit into ways that can help you, as well as get a feel for if this is something the community wants to see. I’m going to assume you’ve watched the guide or otherwise know the rules.

If you want any background, I’ve been playing the game for a few years, including some time spent talking with and playing against members of the European Mahjong Association (who are all a better player than I am). I’m no expert and I’m still at an “intermediate” stage of sorts, but what I’ve learned should help you get started.

First Things First: Just Play

If this is your first time ever playing the game, then just play it. Have some fun figuring out the different aspects of the game, throwing tiles around and finding the luck for that one big hand you’ll never see again. End up in furitentoo, have that big win denied and learn how not to make that mistake again. Just have fun.

Once you’ve had a few games (maybe 10-20, with friends or against random people) come back here and have a read further.

The Grand Strategy: Losing By Less

Before I get into individual strategies, I’m going to cover the aspect that will hold over long strings of games, and will eventually (hopefully) result in a slow-but-steady increase in rank:

Don’t aim to win, aim to minimise losses.

Sure, you want to win hands and you want to win games, or you’ll never gain rating. Bear in mind, there are 4 people playing this game and it depends moderately on luck; even most pro players out there have a game win rate of 30~35%, with greater than that extremely rare.

The key overall is to win a hand when you have a very good chance of doing so, but otherwise know when and howto cut your losses; the eventual winner winning by tsumo will cost you a lot less, a draw will cost you practically nothing and someone else dealing into their ron will not only cost you nothing but also put you ahead of the person who dealt in.

I’m not sure how rating is going to change in FFXIV’s version of the game, but if the developers were clever about it, they’ll do it based primarily on placement and not much on actual score – a 2nd place will hopefully gain rating whether they lost by 100 points or by 40000 to a lucky yakuman, or even if your score ends up negative. That should hopefully mean that, despite being way, way behind 1st place, getting to or holding 2nd place should be your top priority. This will hopefully get you rating and you can continue your climb.

The lodestone specifically says in its mahjong play guide that your rating and rank points will be decided only on your placement and the other players’ ratings.

I will edit this section (more) later as the playerbase figures out how (much) rating changes on a game. For the record, I’ve played 5 games – two wins, a 2nd and two 3rd – and I’m now on R1549. It’ll have depended on my opponents’ R too, but it seems to follow similar systems I’ve seen (namely Tenhou’s system, though Japanese players are brutal so climbing is a lot harder there).

Reading Your Hand

Since I’m going to start getting into the nitty-gritty side of things, I’ll start putting up full hands, and I’ll be formatting them with code formatting, like this: 223m45677s999pEE Wh

This denotes a hand with three tiles of characters (man as I’ll use) of two 2 and a 3; 5 bamboo tiles (sou) of 4, 5, 6 and two 7; three dot tiles (pin), all 9; two East wind tiles and a White Dragon tile that’s just been drawn (note the space). Hopefully it should be easy to read this hand; I’ll use Wh for White Dragon and We for West wind, but nothing else shares an initial so it should hopefully make sense from there.

But this section is supposed to be not only about reading this syntax, but also how to read what to do with a hand, so let’s use that example. What would you discard?

Hopefully this one is simple – your hand is fairly close to ready and the White Dragon you just pulled can’t go with anything you have, so you just discard that Wh. Many of your draws will be like this, so knowing the tile is all-but-useless is the very first step. Of course, you will get the odd time you discard a seemingly-useless tile before you draw tiles later that, if you’d have kept that tile, would’ve made your hand ready faster or score a lot higher. That’s the nature of the game, you didn’t discard incorrectly but the tiles just fell that way.

Let’s have another example: 678m1345678s678p 2s

This one is a lot trickier and will depend on more factors than you can see immediately by these tiles, but if we assume it’s the first draw on the first hand (and count yourself lucky if that ever happens because you don’t need much to win a decent-scoring hand from here).

There are two obvious ways you can go from here: you can either go for a full straight (ittsu) of 1-9s since you only need a 9s to finish it, or you can go for a triple run (sanshoku) with the 678 in all three suits. Both score 2 han when closed, and both will give the same fu unless something drastically changes, so score will generally be identical either way.

The key here is that the straight option requires you to get a 9s, and then also to make a pair with what’s left (discarding from either the 678m or 678p twice and making a pair with the last tile). For the triple run option however, the hand is already ready – you actually only needed another 1s to win the hand with the triple run, making the 1s the pair.

Now that you’ve drawn a 2s though, what you can now do is discard the 1s and keep the 2s, hoping to make a pair out of that instead. This will give us the same triple run as before, but also “all simples” (tanyao) for another han, boosting the hand’s score. Since this is being counted as turn 1, you can discard the 1s and call riichi since your hand will be closed; this will add double riichi if you win, one-shot (ippatsu) if you win before or on your next draw, and with any dora, the score can start going crazy, easily into mangan or haneman territory.

That’s one example of how to increase a hand’s score. However, if it takes time to build that hand, someone else may win before you and that effort goes to waste. Let’s go into the next section doing the opposite: making a hand quickly.

Tile Efficiency

This is a concept I learned from the EMA guys I knew a while back, which is all about “how easy your hand is to complete”. Let’s look at another example, but this time we’ll look at a few groups instead of a whole hand. Take a look at each group and see what’s needed to complete a meld.

123m 46m 89m 23s SS R

• Firstly, the 123m is done – you have a full meld here, so it’s not waiting on anything.
• The second group is 46m – this is waiting on a 5m only. There are four 5m tiles elsewhere in the wall, in players’ hands or in the dead wall (or in discards if this isn’t turn 1, but we’ll assume it is). This is a centre wait.
• The third group is 89m – this can become a run with a 7m only. Same as the previous group, there are only fourtiles of 7m to complete this run. This is an edge wait.
• The fourth group is 23s, which can use either a 1s or a 4s to complete it. This allows you eight tiles to pick from to complete it (four of each 1s and 4s). This is a side wait.
• The fifth group is SS. This can either be your pair as is, requiring no extra tiles to complete, or it can become a set of 3. To become a set, it needs another S; however, you already have two, so there are only two tiles available to complete it. It may be better to leave this is a pair in some cases. This is a set wait.
• The sixth group is a lone R. This has little chance by itself of making anything, so if this is the last thing to do to win your hand, you are looking at three remaining R tiles to find one of to complete your pair. This is a pair wait.

So, in order of highest-to-lowest tile efficiency – the chance of actually drawing or someone discarding a tile you need to complete the group – you want groups of:

Side waits > Edge/Closed waits > Pair waits > Set waits

Of course, if you don’t care about calling and opening up your hand, set waits are fairly easy to call pon on since they can be from any player, while pair waits cannot be called at all unless it is for ron to win the hand.

There is one thing that puts closed waits slightly ahead of edge waits though. Take these examples: 12s 24s

• The 12s edge wait can only be completed with a 3s. There is little you can do about this in 1 tile apart from hope you get a 3s.
• The 24s closed wait can only win on a 3s as well, but what if you happen to pull a 5s instead? You can then discard the 2s and you have a 45s side wait instead. You can now win on either 3s or 6s, increasing your tile efficiency from before. It’s purely advantageous over the side wait as, in order to make the 12s more efficient, you need to draw a 4s to get 24s, which just puts you in the same position as this but 1 tile later.

Of course, if you have 123m and 123p then generally you want to keep 12s instead of going for efficiency as that boosts score (for triple run, sanshoku), but giving yourself more tiles to win on in the general case is a good practice to get into.

There are some groups that give more tile efficiency than the base 8 tiles of a side wait, or similar groups. These will be more fluid than standard, as you can adjust them to suit your needs. Here are some more example groups:

667p 6789m 1123s 246m 5588p

• This first group, 667p, means you can either grab another 6p (two left elsewhere) and discard the 7p, or a 5p or 8p to complete a run and discard a 6p (another eight as per a standard side wait). This gives ten tiles to complete the group on.
• 6789m lets you complete a pair with either the 6m or 9m, and leave the other three as a run. Each pair has three tiles elsewhere, so this gives six other tiles you can pick to complete. (You can also pick up a 5m which will allow you to find a 4m or 7m to complete two runs, which advances tile efficiency further to seven – you already have a 7m – and in a different way, useful if you already have a pair elsewhere).
• This group can allow you to use the 11s as a pair and look for another 1s or a 4s to complete the 23 run. That’s six tiles available. You can also use a 1s to complete a set, but then you’re still looking for a 1s or 4s to complete a run, leaving only five tiles.
• 246m allows you to find either centre wait of 3m or 5m, discarding the 6m or 2m respectively if you find one. This turns a normal centre wait of four into eight tiles, same as a side wait.
• The last group, 5588p, counts for any two pairs, suits or honours, but the principle is the same. You’re waiting on a tile to make a set of three, and leaving the other as the hand’s pair. In this case, you’re waiting on a 5p or 8p, and you have two of each already. This leaves two of each tile, or four total tiles available for you to wait on.

Tile efficiency goes into much greater depth than this, and there are many more groups than what I’ve listed here, but these are relatively common and should give you an idea of where to start.

Riichi: The Name Of The Game

This game is more interesting in my opinion than other versions, purely because of this mechanic. If you watched the video (or already know) then you should understand the rules behind this and what the rewards are for taking the risk.

So when do you call riichi, when the opportunity arises? The answer, for a simple strategy primer, is also quite simple:

Pretty much always.

There are very few instances where you should not call riichi, like “if you’re less than 1000 points in front of someone and the game is about to end”. There are times you will need to know when not to, but for a general guide, you should pretty much always riichi when available.

The only time you absolutely mustn’t riichi is when you’re in furiten. If you have a hand ready waiting with 67p and you have a 5p in your discards, you are in furiten until you change your waits to not include anything in your discards. If you riichi, you cannot change your waits, which means you cannot leave furiten for the rest of the hand. You can only win by tsumo from that point, which reduces your chances of winning to about 1/4 of normal. (I have done this. Recently. Like, yesterday. I’m still kicking myself.)

Easy stuff, eh?

Defence: The Art Of Not Losing (By Much)

So you have a basic understanding of how to build a hand, and how to build it a little more quickly. Let’s set up a short scenario.

You’ve just been holding onto 2nd place, but you’re now dealer and you’ve just drawn a tile that will allow you to riichion turn 5. It’s a cheap hand but a win means you’ll get another turn as dealer (and another round to catch up in points).

Then disaster – someone ron‘d on your discard and you have to pay them 8000 points, pushing you into last place. As it turns out they’d called riichi the turn before and were ready, faster than you, on a better hand. So what can you do when you can’t see the opponents’ hands? How do you prevent yourself dealing into them?

There are a few things to consider: when to defend, and how to defend. Before either of those, though, there’s one thing that will be a good general rule to follow:

In most cases, you either push or you fold entirely – do not half-ass folding.

Of course, there is plenty you can do to fold while keeping options open, or push for a hand while minimising ron risk. As this is a basic guide for beginners, I’ll just recommend for now that if you try too hard to defend while still building a hand, you’ll be unable to do either effectively.

So, how about when to defend? It can vary heavily, but again, for a general beginners’ guide to strategy, we can leave it something like:

• If your hand is not ready (you couldn’t riichi even if it were fully closed).
• Even if it is ready, it’s not going to score too highly.
• Even if it is ready, it’s got a bad wait (pretty much anything other than a side wait is “bad”, but we can cover that in more detail another time).

Based purely on your own hand, these are the things to think about when deciding whether to push or defend. If your hand is ready and will score a lot of points if you do win, then go ahead and push on. You’re about as likely to complete your hand as they are theirs; the reward outweighs the risk.

If your hand might score really well, but you’re waiting on a pair or are not in a ready state, it might be better to defend. In most cases, this will be defending entirely by breaking apart your hand and discarding the tiles that are safest to discard, even if it’s breaking up your highest-scoring parts of your hand.

Now we move onto how to defend. The first step is to go over some of the rules again, particularly those surrounding the state of furiten. If you suspect someone is ready (which when under riichi, it always will be when playing digitally…unless the game is life-like, but don’t count of FFXIV’s version being that way), then they have to obey the rules of furiten. You can use this to your advantage when defending against that. Let’s go over the basics of the rule again:

• You cannot win on ron if one of your potential waits are in your discard pile (even if one of the other tiles comes up).
• If you pass up a winning tile, you cannot win on ron on that same tile in the same turn from someone else. This expires on your next turn or when someone makes a chii, pon or kan call.
• If you have called riichi and pass up a winning tile, you cannot win by ron at all as you can no longer change your waits.
• All instances of furiten (while they last) can only be won by tsumo.

So how do we use these rules to our advantage? There are some perfectly safe, and somewhat safe options.

The first rule means that anything in an opponent’s discard pile is perfectly safe from them. If an opponent has a 5p in their discard pile, you can safely discard one as they are either

• not waiting on 5p, or
• waiting on 5p, but in furiten, meaning they cannot claim your discard.

This does not make a 5p discard safe from the other two players (unless they also have a 5p in their discards), but it’s a safer choice than something you can’t see any copies of. These tiles are known as a player’s genbutsu tiles, and they can never be won by that player on ron that round. If you’re just starting out learning defence, this is your simplest and most effective defensive tool.

Another tile that is perfectly safe is the same tile the player to your left just discarded. Since a player cannot win by ron on a tile if they passed up that same tile within the same turn, it goes without saying that if two other players have passed up on the left player’s tile, they can’t win on yours if you discard the same one. Less common but common enough that it’s a good one to remember.

The last perfectly safe option is simply based on the discard piles and not on furiten. If you see three of an honour tile (a wind or a dragon) out on the board, you can almost 100% safely discard the fourth. I say “almost” because someone could be waiting on it to win by Thirteen Orphans (kokushi musou, one of every honour tile plus any one more), a very rare yakuman hand. By this, if you see four of another honour tile on the board or in your hand, nobody will be able to obtain such a rare hand and it becomes a guaranteed safe tile.

You won’t always have the option of perfectly safe tiles though – sometimes you will have to put down risky tiles, but there is still some strategy you can use. Remember the first rule: “You cannot win on ron if one of your potential waits are in your discard pile (even if one of the other tiles comes up).” The main idea behind the strategy I’ll cover here is one known as suji.

Let’s say the opponent you’re looking at has a 6m in their discards. You look at your hand and you decide to fold, but you have no 6m (nor any other tile in his discard pile). You do instead have a 3m and a 9m. Why have I brought these up?

The rule states that you cannot call ron on a tile if one of your waits is in your discard pile, even if it’s not that tile that appears. If your opponent has a 6m in their discards, they cannot be waiting on a 45m or a 78m side wait, or they’d be in furiten. This means that a 3m or 9m is safer to discard than most other tiles. It expands on the tiles that you can discard to minimise risk when defending (and thus minimise point loss, meaning high placements and more rating over time). Of course, they could be waiting with 99m or 24m, but good players will tend towards good tile efficiency so these waits will be a little less common.

There is one situation where this suji tile is actually more dangerous to discard than normal, and that’s on the tile that was used to discard as riichi. If someone discards a 5m and calls riichi on it, it’s often more dangerous to discard a 2m or 8m. This is because it’s not uncommon for someone to have an almost-ready hand, but have a 135m pattern or something. If they call riichi with this pattern the only remaining group to complete, they can discard the 5m to wait on a 2m to win. See how this makes what would normally be a safer tile the exact opposite? This is known as a suji-trap, and can sometimes bait people into discarding the tile they need. Of course, this is something you can use to set up your own traps.

This is as far as I’ll cover today. I’ve put down a lot of words, but I’ve also barely scratched the surface of mahjong strategy. If this gains any traction I can either continue with a more in-depth guide to some aspects, or perhaps put up a discord channel that people can organise games in and discuss the game. I’m still learning myself, so any corrections or modifications to this guide (as well as any more advanced tactics from players better than me) are more than welcome.

Hopefully, if you give the game a go and it’s caught your interest, this short guide will help. On any other account, just have fun with it; it’s a deep, interesting game that I enjoy, and I hope I can help others enjoy it as much as I do.

And hopefully I’ll see you at the table.