Shanten, Tenpai, and Effective Draws: Are More Waits Better?

Key takeaways

• Shanten answers “how many effective improvements remain before tenpai,” not how many turns a win is guaranteed to take.
• Zero-shanten is tenpai; winning then requires a valid completion tile and the rules’ winning eligibility.
• Define effective draws before counting: only shanten-reducing tiles, or also same-shanten shape improvements?
• More wait types do not guarantee more live copies. Subtract tiles visible in the hand, discards, and calls.
• Multi-sided waits are usually wider, but a single wait can be better through live tiles, yaku, value, furiten, or flexibility.

What is shanten?

Shanten is a structural measure of a hand’s distance from tenpai.

• Zero-shanten: the hand is at tenpai and needs a winning tile.
• One-shanten: normally one effective improvement is needed to reach tenpai.
• Two-shanten: normally two effective improvements are needed to reach tenpai.

“Normally” matters. Shanten measures distance, not the certainty of drawing an improvement next turn. Two one-shanten hands can differ sharply: one may have more than twenty live effective tiles while another has only a narrow route.

Standard hands, Seven Pairs, and Thirteen Orphans use different calculations. Thirteen-tile and sixteen-tile games also target four and five melds respectively, so a calculator must support the actual rules.

What is tenpai?

Tenpai is zero-shanten: one valid completion tile can finish a winning structure. Those completion tiles form the wait.

Structural completion and permission to win are not always identical:

• Riichi also requires yaku and a furiten check.
• Cantonese rules may require minimum fan.
• Taiwanese tables may require minimum tai.

A calculator reporting tenpai does not mean the hand can declare a win under every ruleset.

What counts as an effective draw?

The strict definition is a draw that lowers shanten. It moves one-shanten into tenpai or tenpai into a completed hand.

In practice, “improvement” may also describe tiles that:

• Keep shanten but turn an inside wait into a two-sided shape.
• Keep shanten but increase next-step acceptance.
• Preserve speed while adding a yaku or value route.

Before comparing an article, explanation, or tool, establish whether its effective-draw count includes improvements. Different definitions can produce different correct totals.

Wait types versus tile copies

The block 45 Characters accepts 3 or 6 Characters:

• It has two effective tile types.
• It has eight theoretical copies because four of each exist.

If one 3 Characters is in the hand and two 6 Characters are visible in discards, only 3 + 2 = 5 unknown copies remain. This is live acceptance.

A useful comparison records:

1. Which tile types advance the hand?
2. How many unknown copies remain?
3. Do multiple blocks count the same tile?
4. What wait appears after each draw?

Common wait shapes

Compound shapes can create three or more wait types. The name describes geometry; actual width still depends on every completion tile and its live copies.

Is a multi-sided wait always better?

No, although it is usually favorable when all other conditions match.

The wide wait may be dead

Two listed tile types can have many visible copies. A single wait with nearly all four copies unknown can have more live tiles.

Furiten may affect the multi-sided wait

Riichi furiten applies to the entire wait. If any tile in a multi-sided wait appears in the player’s own discards, the player cannot ron the other tiles. A non-furiten single wait may preserve discard-win eligibility.

A single wait may retain value

The single wait may keep a Value Honor, dora, flush route, or another scoring feature that the wider choice breaks.

A single wait may be easy to change

Some single waits can switch on the next draw to a live, safer, or more valuable tile. A fixed multi-sided wait may lack that flexibility.

Game state can demand another objective

Winning any small hand, finding comeback value, or retaining a safe tile gives wait width different weights. The number of wait types alone cannot settle the table decision.

How should two waits be compared?

Use six checks:

1. Winning eligibility: yaku, fan, tai, and furiten.
2. Complete wait set: do not inspect only the preferred tile.
3. Visible copies: count unknown copies of every wait.
4. Overlap: never add one physical tile twice.
5. Value: do self-draw, discard win, or different waits change the score?
6. Game state: turn, attack and defense, placement, and opponent speed.

Pure efficiency can stop after step four. A table decision needs the final two.

What matters from one-shanten?

Before tenpai, adding every shanten-reducing tile is still incomplete analysis. Inspect the wait created by each improvement.

Candidate A may accept 12 immediate tiles that mostly lead to inside waits. Candidate B may accept 10 but usually reach two-sided or multi-sided waits. Over several draws, B can complete more often. This is why the largest immediate count is not always best.

Common misunderstandings

• One-shanten means one draw from winning: it normally means one improvement from tenpai, followed by a winning tile.
• A two-sided block always has eight tiles: only when all copies are unknown.
• Multi-sided totals can be added blindly: compound waits and overlapping shapes can duplicate tiles.
• Tenpai always permits a win: rules eligibility still applies.

Scope and limitations

This article supplies shared language for standard-hand efficiency. It does not calculate complete scoring or attack-defense value for a specific ruleset. Unknown tiles are not guaranteed to be uniformly accessible, and wall reading is inference rather than certainty. Formal play follows the current rules’ definitions of winning and furiten.

Building the sequence with Tingpai Lab

Tingpai Lab places these concepts in complete-hand candidate comparisons. Practice by naming each candidate’s shanten, effective tile types, live copies, and resulting shapes before checking the explanation. Memorizing only the displayed answer does not build a reusable process.

Continue reading:

• Why more effective draws are not always better
• 13-tile vs 16-tile efficiency
• What to do with too many blocks