When a betting line moves opposite to the majority of public money on the game.
Sportsbooks adjust lines based on the composition of incoming money, not just its volume. If a small number of large, historically-sharp accounts bet one side, the book respects that information and moves the line — even if 80% of the ticket count is on the other side. The result is a line moving in the opposite direction from the public's clear preference. That's reverse line movement.
The information value is subtle. Public-money percentages — what fraction of bets are on each side — are widely published. If the public is 75% on the Yankees and the line moves from -130 to -140, that's normal: ordinary money is driving the line. But if the public is 75% on the Yankees and the line moves from -130 to -120, the book is telling you something. Sharp money is on the Mariners with enough conviction to move the line against the ticket count.
Reverse line movement is not a guaranteed winner. Sharp accounts are right more often than the public, but they're far from infallible. The metric is best used as a confirmation rather than a primary signal — if a model already likes a side and the market is showing RLM in agreement, the conviction grows. If the model likes a side and the market is moving against it through ordinary action, that's a yellow flag worth investigating before publishing.
Reverse line movement has become harder to detect cleanly in modern markets. Many books no longer publish granular ticket vs. handle splits. Aggregator sites estimate them, but the estimates carry noise. The cleanest version of the signal is still the comparison between the line at open and the line at close: if it moved one direction while public action ran the other way, sharp money was the cause.
We treat reverse line movement as a tiebreaker, not a primary input. When our model edge is marginal and RLM confirms the same side, we'll ship the pick at a higher tier. When RLM contradicts the model, we restudy before publishing.