How to Track Other Players in American Mah Jongg

Tracking other players means using visible information to narrow their possible hands and improve your own decisions. Exposures give you the strongest evidence. Discards and Charleston passes add context, but they rarely prove which hand someone is building.

You cannot see another player's concealed draws or know every tile they choose to keep. Good tracking starts by separating what happened publicly from what you are guessing. Use the NMJL card being played at your table to test those guesses.

Start with facts you can observe

You can see a player's exposures, hear the tiles they discard, and know the tiles you personally send and receive during the Charleston. You may remember who made a particular discard, but the shared discard area alone does not preserve that history for you.

You usually cannot tell which numbered tiles a player is collecting just by watching them move tiles around a rack. A pause, smile, or hesitation is not reliable evidence of a particular hand.

Use the clues with different levels of confidence:

This distinction keeps you from treating an attractive theory as a fact.

Filter the card using the first exposure

Suppose a player exposes a kong of 6 bams. Search the card for hands that can legally contain that group. Include any permitted choices of numbers and suits, rather than searching only for a line that literally prints a six.

Remove candidates requiring a pung or pair in that position instead of a kong. Remove concealed hands, because they cannot retain the exposure. Check each line's other requirements and notes. The same exposure may fit several candidates. ^[2]

Do not assume that the first exposed group is the first group printed on the card. Exposures need not be made in card order. Use their identity and size to search the whole hand. ^[2]

If a second exposure appears, repeat the search for both groups together. A candidate must accommodate both under one consistent suit and number assignment. This is often where a broad guess becomes a short list.

Use the remaining requirements to judge a discard

Imagine that, after checking the actual card, two hands remain compatible with an opponent's exposures. One requires flowers in its concealed portion; the other requires a particular dragon group.

If your possible discards include a flower and an unrelated numbered tile, the numbered tile may be the better defensive choice against that opponent. You still need to assess the other players and your own hand before discarding it. A tile that looks harmless against one exposure can help someone else.

You do not need to identify an opponent's exact waiting tile for tracking to be useful. Recognizing that one discard fits several remaining candidates while another fits none of them can improve the decision, provided your candidate list is complete and current.

Avoid describing the latter tile as guaranteed safe. You may have missed a permitted interpretation of the card or misunderstood an exposure, and the other players still have their own possibilities.

Count natural tiles inside exposures

An exposure of three natural 6 bams and one joker represents a kong of 6 bams, but only three natural 6 bams are physically used. That distinction matters when you count unseen copies or look for a possible joker exchange.

If you hold the fourth natural 6 bam, you may have an exchange opportunity on your turn under the normal rules. Do not assume the joker will still be available then; another player may exchange for it first. If you are counting tiles for your own required pair, count actual naturals rather than the identity represented by the joker. ^{[1], [2]}

Exposures show completed groups, not the contents of the concealed rack. Even two exposures do not reveal exactly how many tiles the player still needs.

Treat discards as changing clues

Several discards from one suit may suggest that the suit is less important to the player. They do not rule it out. The player may have completed a required group, discarded excess copies, or changed plans.

A tile someone discarded earlier can still help that person later. For example, a player may switch to a hand using that tile or need another copy after changing a suit assignment. An earlier discard is not a permanent promise about future needs.

Likewise, a player who does not call a tile has not proved that they do not want it. They may lack the tiles needed for an exposure, be protecting a concealed hand, or choose to preserve flexibility.

Use discards to refine a candidate list already supported by the card and exposures. Avoid building the entire list around one remembered throw.

Keep Charleston inferences modest

During the Charleston, you know what you pass and what you receive. You do not know every tile retained by each player. A tile that reaches you may have traveled through earlier passes, and an allowed blind pass can move a tile without the passer inspecting it. ^[2]

Even when someone deliberately passes a tile, their plan may change after the next pass. Receiving several tiles from one suit is a useful observation about that moment; it does not prove the sender will avoid that suit throughout the hand.

The most useful habit is to update your view when stronger information appears. Give a new exposure more weight than an early Charleston guess.

Keep your own decisions in focus

Use other players' turns to inspect exposures and compare them with the card while continuing to follow play. On your turn, combine that information with the actual needs of your rack.

Early in a promising hand, keeping your own progress may justify some discard risk. Late in a weak hand, protecting against another player's likely needs can matter more. Neither choice requires certainty about every opponent.

Track observable groups, retain more than one possible explanation when appropriate, and revise your assumptions as the table changes. That makes opponent tracking useful without pretending the concealed tiles are visible.