Efficiency is not everything: speed vs value at the table

1. Efficiency measures speed, not return

The standard efficiency question is “which discard reaches tenpai most often?” It assumes every win is worth the same and that reaching tenpai means winning. Neither assumption survives a real hand.

The same tenpai may be worth 1,000 or 8,000. The same useful draw may add two safe tiles — or consume the last genbutsu in your hand. Efficiency cannot see this because it counts accepts, not returns and costs.

None of that makes efficiency useless. It is the foundation: without speed, value never gets paid. But judging the whole hand by the height of the foundation is how players become fast in the wrong spots and slow in the right ones.

2. Dora and yaku: three value tiers

It helps to sort hands into three rough tiers. One to two han is 1,000 to 2,600 points, usually enough only to keep a flat game moving. Three to four han is 3,900 to 7,700, where a win starts to move placement. Mangan and above starts at 8,000 and is the main target when you trail.

Dora tells you which tier is closest. One dora plus riichi is usually two han; two dora can push you into the three-to-four range; a dora pair or triplet raises the value again. Yaku work the same way: tanyao, pinfu and iipeikou are “incidental” yaku that turn two han into three.

So when a discard “reaches tenpai”, add two questions: roughly how many han is this tenpai, and is that enough for the current score? If not, fast tenpai may be converting a hand with options into a guaranteed 1,000 points.

These are relative judgements, not formulas. Many dora do not oblige you to play big. If an opponent has already riichi, holding a dora may simply deal in first. Value and danger are read together.

3. When a slower hand is correct

A slower line usually needs three conditions: the value crosses a tier, the delay is controlled, and the score actually needs that value. Remove one condition and “slow” becomes “stuck”.

Example: you hold 234m 456m 678p 77s 34s and draw 9m, with 9m as dora. Efficiency says discard 9m, wait on 2s/5s, and take riichi plus tanyao for about two han. At flat scores that is correct: bank the roughly 2,000-point certain gain.

But in South 4, trailing by 8,000, two han is not enough. You can consider discarding 3s and keeping 9m. The hand drops to one shanten; if you draw a second 9m, the pair turns the shape into a 7s/9m shanpon, and riichi plus dora reaches three or four han. The price is one extra turn, and if the 9m never comes you may end with nothing.

That is what a slower hand really looks like. It is not “slow is better”; it is buying a specific value tier with a controlled delay, only when the score needs it. Do not do it at flat scores. Do it when the alternative is a cheap win that changes nothing.

4. The defensive cost of chasing value

Chasing value is not only “keep one more dora”. Every attacking tile you keep displaces a safe tile. Every suit you redirect toward honitsu may be the suit an opponent is waiting on.

Example: you hold a 77s dora pair, and two each of 6s and 8s are already visible. You keep the 7s for value. But if an opponent is waiting on 6s/9s or on a 7s pair, that same tile is both your points and their winning tile. When you chase value, ask what the tile is worth to each side of the table.

The second cost is optionality. With several safe tiles you can defend for two turns and then decide. Spend them all chasing value and you lose the right to change your mind. Before committing, count: after this discard, how many safe tiles remain? Fewer than two is a warning, not a detail.

5. Two worked hands and a drill

Hand one: 234m 456m 678p 77s 34s, draw 9m, dora 9m. At flat scores, discard 9m for a 2s/5s wait, riichi plus tanyao, about two han. In South 4 down 8,000, discard 3s, keep 9m, accept one slower turn and aim for the three-to-four-han shape. Same tiles, opposite answers, because the score changed.

Hand two: 234m 456m 678p 77s 99s, draw 5s. Discard 5s for a 7s/9s shanpon; riichi gives at least two han, and if 9s is dora the 9s win is three. That is already enough. Do not break the 99s to chase a hypothetical extra han. Fewer winning tiles, but certain value, and speed matters more.

Drill one: for each hand, compare only two candidates — fastest tenpai, and the slower line that keeps dora or a yaku. Use a score calculator to write down the winning value of each, then count the safe tiles each line preserves. Do not say “this one feels bigger”; write the number.

Drill two: collect ten positions from the South round where you trail. Decide only whether you need to cross a value tier. If yes, find a slower route; if no, play for efficiency. Then review how often you paid two extra turns of risk for one han you did not need.