Overdue is the model’s way of expressing a gap between chance creation and goals scored in the live match.
The raw material is live expected goals. As a match progresses, each chance a team creates carries a probability of being scored, and those probabilities accumulate into a live xG total. In most matches, goals track that total closely enough. Some matches drift away from it: a side racks up meaningful chances and the scoreboard does not move.
Overdue measures that drift. The model adjusts the live xG figure to reflect that sustained chance creation without reward tends to correct, then compares the adjusted figure against goals scored. The wider the gap, the more overdue the match, and the more weight the in-play analysis gives to the next goal arriving sooner than pre-match rates alone would suggest.
It helps to see where this sits in the wider in-play logic. The model does not start with the markets; it first forms a view on the live goal pressure in the match, meaning how much genuine threat is being created relative to what the game state would lead you to expect. Overdue feeds that view. Only once the goal-pressure verdict is established does the model scan the markets: when pressure is high it leans toward the goals being underpriced and projects forward on that basis, and when pressure is low it does not force a goals angle onto a quiet game. A guess some users have made is that the model divides a team’s live xG by their full-time expected xG and calls the result overdue. It does not work that way; the live read is built from several signals of current goal pressure, not from a single ratio of one xG number against another.
Overdue is not the gambler’s fallacy in disguise. The everyday use of the word implies a goal is owed because time has passed; the model does something closer to the opposite. It measures present, live evidence of threat, grounded in the quality of chances already created in this specific match, and asks whether the current price has caught up with that evidence yet. A live recommendation is a statement about mispricing in the moment, not a countdown. Overdue is also never read alone: the analysis weighs it alongside current pressure, game state, and time remaining before it shapes any recommendation.
The exact way these signals are combined, the precise factors involved, and the internal sequence the model follows are product detail held with the technical team. For reading the output, the principle is enough: when you see overdue referenced, the chances already created in this match justify more goals than have been scored, and the model has priced the next-goal markets accordingly.