A follow-up is
- the next alternating play in a sequence; or
- after tenuki, the next play in a local area of the board; or
- for a ko-threat, the next move if the threat goes unanswered.
In the second case, it may mean a move that follows another move of the same color after some lapse of time (because the opponent played tenuki). When one speaks of a follow-up play to a joseki, however, it means just that the joseki comes to a natural end; and later either player may wish to return to add a play that wasn't urgent earlier.
Related terms are follower, threat, target, aim, fukumi and nerai.
This is one of the most common concepts in informed discussion and analysis of particular positions, and is much used on SL.
There are slightly different usages to be found here.
Follow-up to set patterns
For example 'cross-cut follow-up' would mean how to play next as soon as a cross-cut happens (which turns out to be a complex discussion - see Cross-cut Workshop).
There are pages here such as
These discuss the same matters in other common shapes. These are (all three) volatile situations in which it is clear enough to good players that it is important to continue somehow.
There are also well-known follow-up plays in many standard joseki; here there is the difference that joseki which end do so, commonly, at relatively stable positions rather than volatile ones. That wouldn't be true of all tenuki joseki, though.
These pages deal with openings:
Follow-ups for ko threats
A ko threat is normally a play that would be sente assuming no active ko fight. Therefore in the usual run of play you don't expect to play a follow-up to it. If however your ko threat is ignored, there is scope for unexpected types of follow-ups, which may destroy territory, or lead to the capture of or severe attacks against substantial groups of stones.
In fact, the same can be said for the player winning the ko: the next local play after that may be big. One kind of picnic ko is seen when the follow-up to capturing the ko for one side is 'too hard to handle'.
Follow-ups implied by plays presumed gote
For example, a checking extension is supposed to be distinguished by some attack or invasion that can take place, if the opponent doesn't answer it locally and directly.
An almost sente play is distinguished as a play that is essentially gote but has a large or attractive follow-up sequence implied.
While a small play with a large enough follow-up is probably sente, and a large play with a small follow-up probably gote, there is a cross-over region where the play and its follow-up are of similar size. This leads to the ambiguous move theoretical concept, at the precise point where there is a difficulty resolving the distinction.
According to the tsume discussion, that's incorrect. Charles