How to Win at Mahjong: Strategy for New Players

Pick a Direction Early

In the first few turns, look at your hand and choose what it wants to be. A hand with seven or eight tiles of one suit is telling you something. A hand with three pairs is telling you something else. Players who wander through the first half of a deal discarding whatever looks least useful end up with no shape at all, and then they lose to the player who committed at turn three. See also How to Play Mahjong Online: Free in Your Browser.

Committing early turns your tiles into a plan instead of random discards. This is the single biggest difference between a beginner and a player who wins consistently, and it costs nothing to start doing. See also Mahjong Strategy: How to Actually Win More Hands.

"Committing" does not mean locking in. It means having a default direction you are steering toward, one you are willing to abandon when the draws contradict it. The mistake is not changing your mind — it is never having a mind to change. A player who has decided they are building an all-simples hand will discard a lone honour without hesitating; a player who has not decided will keep it "just in case", and by the time they work out what they are doing, the wall is half gone.

A useful habit for the first three turns: name your hand out loud, even silently. "This is a bamboo hand." "This is a pairs hand." If you cannot name it, you do not have one yet, and the correct action is to keep discarding the tiles that help nothing.

Discard Lone Honours First

Wind and dragon tiles are the hardest to pair up. A lone honour does almost nothing for a beginner hand, so the usual answer is to throw it early, while it is still safe to do so.

The exception: if you have a pair or triplet of a useful honour (like a dragon or your seat wind), keep it — it is worth points in most rulesets and calling rules make triplets easy to complete.

The reason honours are different from suit tiles comes down to arithmetic. There are only four copies of any honour, and no way to build a run out of them, so a lone honour needs exactly a pair to become anything at all. By contrast a lone numbered tile can connect upward or downward into a sequence, which means it has roughly twice as many ways to become useful. When you are choosing what to throw, that difference in potential is the whole calculation.

Early game is when discarding an honour is cheapest, because nobody has had time to develop a hand that needs the tile you are throwing. Late game the same discard can be a direct payment, which is why the order of your discards matters as much as their content — honour tiles out first, and let the dangerous suit tiles wait until you know more.

Read the Discard Pile

Every discard is information. If three players are throwing away bamboo, the bamboo tiles you need are probably sitting in their hands or are still buried in the wall — and, more usefully, you have just learned which suit nobody wants. The pile tells you what is safe to throw and which suit is crowded.

Watching the discards also helps you avoid discarding a tile that would complete an opponent’s hand.

The most valuable reads come from what is *not* being discarded. A player who has thrown away every honour but keeps passing on a particular suit is signalling something. If a suit stops appearing on the table after a few turns, that is often because the people who could throw it are now collecting it, and the tiles you were hoping to draw from that suit are no longer coming.

Combine the pile with what you see called. When a player pongs a tile, they have declared three of the four copies exist and two of them are on the table. That makes the fourth copy significantly safer for you to discard against them — and it also tells you their hand has committed to that direction, which is useful when you are deciding whether to push or fold.

Keep a running count of one or two tiles you actually need rather than trying to track everything at once. If you are waiting on a 5 of bamboo and you have watched two of them go past, you know your wait is thin and it may be worth switching to a shape with more live tiles — a wider wait is often worth more than a prettier one.

Do Not Over-Call

Calling tiles makes your hand open and closes off scoring patterns. In most rulesets a concealed hand is worth more, and an open hand loses access to the patterns that depend on being closed.

A good rule for beginners: only call when the call puts you one or two tiles from ready. Otherwise keep drawing and keep your options open.

There is a second cost that beginners rarely notice. Calling reveals your direction. The moment you pong two tiles of a suit, everyone at the table knows what you are building and will hold back the tiles you need — and will also stop throwing the tiles that would let you finish. A concealed hand gives nothing away and can change direction mid-deal; an open hand has publicly announced what it is.

So weigh the call properly: it buys you speed, and it pays for that speed with information and with scoring potential. If the call takes you from three tiles away to one, it is usually worth it. If it takes you from four to three, you have given away your plan for almost nothing, and the correct play is to keep drawing.

Reach Ready as Fast as You Can

The player who gets ready first usually wins. Trade tiles you cannot use for tiles you might use, and keep the shape flexible until the wall forces your hand.

Our tables include a readiness hint that shows exactly how far you are from a complete hand, which trains you to see the distance without having to count it yourself. Watch what changes when you consider different discards — that is the fastest way to internalise hand-reading.

Speed compounds. The earlier you are ready, the more draws you get at the tile you need, and the more chances you have to win before someone else does. A hand that becomes ready on turn eight has most of the wall left to work with; the same hand becoming ready on turn fourteen is relying on luck.

There is a tension here worth understanding rather than resolving. Chasing the fastest possible readiness can push you toward a hand worth almost nothing, while chasing a big hand can mean never becoming ready at all. Most of the practical skill in mahjong lives in that trade-off: how much hand value you are willing to give up for how many turns of speed.

A workable default for beginners is to aim for readiness first and let the points follow. Once you can reliably reach ready while the wall still has tiles in it, you have a foundation you can trade against; until then, a beautiful hand you never complete is worth exactly nothing.

Know the Common Winning Patterns

These patterns appear constantly and are worth memorising:

• All-simples (斷幺九): no 1s, 9s or honours. Reachable from almost any hand.

• Flush (清一色): all tiles from one suit. Worth a lot of points.

• All-triplets (碰碰胡): every set is a triplet.

• Seven pairs (七對子): the classic alternative winning shape.

• Dragon sets (三元牌): three dragons score in most rulesets.

Learn these five and you will recognise most of the hands you actually see at a beginner table. They are not the full catalogue — Chinese Official alone has eighty-one patterns — but they are the ones that decide ordinary games, and recognising a pattern three turns early is what lets you steer toward it while everyone else is still discarding at random.

中 · Red Dragon發 · Green Dragon白 · White Dragon七万 · 7 Crak七万 · 7 Crak

Play Your Ruleset, Not a Generic Game

Hong Kong, Riichi and Chinese Official reward different things. Hong Kong values a quick, cheap three-faan hand; Riichi demands a yaku before you can win at all; Chinese Official sets an eight-point floor that rules out most beginners’ instincts. A strategy that wins in one is often wrong in another.

Pick one ruleset and learn its patterns — switching between them too early is the fastest way to confuse yourself and to start applying the wrong rules mid-game.

Concretely: the "discard lone honours first" advice above is straightforwardly right in Hong Kong and Chinese Official, but in Riichi a dragon pair is a route to a yaku and is worth keeping far longer than a beginner would guess. The advice is not wrong, it is conditional, and knowing which conditions apply is the difference between rules and strategy.

Pick the style you can actually play with people — that usually means Hong Kong for a beginner — and stay with it until its scoring stops being something you have to think about. Fluency in one ruleset beats vague familiarity with three, and it makes learning the second one much faster later.

If you are playing our tables, the practical version of all of this is: pick a direction by turn three, throw your lone honours, check the readiness hint before every discard from turn six onward, and call only when it takes you to one tile away. Do those four things consistently and you will beat bots that are individually better than you, because consistency is the whole game at this level.