On a chain board you drag through a run of matching tiles instead of swapping two. The input is easier, which hides the fact that the scoring rewards something quite different.
The rule that makes these boards different
On a swap board, a match is three or more in a line and that is the end of it. On a chain board you draw a path through adjacent matching tiles, in any direction the game allows, and the length of that path is limited only by what is connected.
The important part is what happens at the end of the path. In most chain games the tiles you passed through do not simply vanish; the tile at the end of the chain grows into a larger version of itself, and the larger version clears a wider area when it is used later. So a chain is not just a clear, it is an investment placed on a specific square.
Longer is not automatically better
Because the growth lands on the final tile, where you finish matters more than how far you travelled. A twelve-tile chain that ends in a corner has put its payoff in the least useful square on the board. An eight-tile chain that ends in the middle of a crowded cluster has set up something that will clear half the screen two moves later.
The habit to build is to decide where you want to finish first, then work backwards to find a path that gets there. That is the reverse of what the board invites you to do, which is to start dragging from wherever you happen to see three of a colour.
Stacking growth in one place
Because each chain grows the tile it ends on, finishing several chains on or next to the same square compounds the effect. Two or three chains ending in the same neighbourhood produce a cluster of grown tiles, and setting off one of them tends to set off the rest.
This is the closest thing these boards have to a combo, and it is entirely under your control. It costs nothing except the discipline to end chains where you meant to rather than where the path happened to run out.
A useful default on an unfamiliar chain level: aim every chain at the centre two columns. Growth in the middle affects more of the board than growth at the edge, and it is a decision you can make without reading the level.
When the board has no chain left
Chain boards deadlock differently from swap boards. On a swap board a deadlock means no two adjacent tiles can be exchanged into a line, which is rare. On a chain board you only need two adjacent matching tiles to exist, which is far more common, so genuine deadlocks are unusual and most games reshuffle automatically when one occurs.
The more frequent problem is a board where every available chain is short and badly placed. That is not a deadlock and the game will not rescue you from it. The way out is to take the least bad short chain in the most useful position, accept a small turn, and rebuild. Two wasted moves spent repositioning are cheaper than five moves of increasingly desperate long chains ending in corners.
What carries over from swap boards
Gravity still works the same way. Tiles still fall vertically, refills still enter from the top, and the loop described in the refill guide still runs after every chain. So clearing low on the board still moves more of it than clearing high, and that principle survives the change of input method intact.
What does not carry over is the value of size for its own sake. On a swap board, a bigger match is nearly always better. On a chain board, size is a means to reach a good finishing square, and a player who optimises for length alone will consistently lose to one who optimises for position.