One position a day, the same one for every visitor, with no account required and no mode to choose. It resets each day, so there is one puzzle to solve and then nothing more to do.
The daily puzzle is not a training programme and will not move your rating on its own. Its value is that it is small enough to actually happen. Most people who decide to train tactics plan long sessions, do two of them, and stop. One puzzle is short enough to survive a bad week, and a habit that survives bad weeks is worth more than a routine that only works when you have an hour free.
It is also a low-stakes way to keep pattern recognition warm between longer sessions. Recognition fades faster than understanding — you do not forget what a pin is, but you do get slower at spotting one.
Because there is only one, it is worth solving properly rather than quickly. Look for what is loose or exposed before calculating: an undefended piece, a king short of escape squares, two pieces sharing a line or diagonal. Then check forcing moves only — checks, captures, and direct threats. Calculate the opponent's best reply before you play, not after.
If you cannot find it, sitting with the position for a few minutes is more useful than guessing. And if you get it wrong, work out which reply you overlooked. That is usually the actual lesson.
For a real session, rated puzzle training matches puzzles to your level and adjusts as you solve. Puzzle Storm is a timed run that trains recognition speed instead of depth, and Woodpecker training repeats a fixed set until the patterns become automatic.
To train the mistakes from your own games rather than general tactics, see mistake training or review a game move by move with Game Report.