The modern Tetris randomizer that guarantees each tetromino once in seven pieces is the bag system.
The seven-bag system places one copy of each of Tetris’s seven tetrominoes— I, J, L, O, S, T, and Z—into a group, shuffles that group, and deals its pieces before reshuffling a new bag. Consequently, every consecutive seven-piece group contains all seven shapes exactly once. This greatly reduces long droughts of a particular piece compared with unrestricted random selection.
The system does not mean every seven-piece window across a boundary is unique: the end of one shuffled bag and the beginning of the next can still contain repeated shapes close together. It guarantees one of each piece per bag, not perfect spacing or equal visual patterns. Players therefore use the queue for planning, but cannot assume the next seven pieces form a predictable order.
Older Tetris games used other random-generation methods, including systems that could produce streaks or droughts. The seven-bag approach became strongly associated with modern Tetris implementations and the Tetris Guideline, helping make play more consistent for advanced techniques such as T-spins and structured openings.
“Chain,” “stack,” and “roulette” are not the standard name for this seven-piece randomizer. The canonical term is bag system, also called the random bag or seven-bag system.