The short answer: if the analysis showed value at the time you placed your bet, you have locked that value in. What the model says when you re-run it later does not change the bet you already made at the price you took.
Markets are dynamic: odds move and assessments shift as new data arrives. That is the nature of betting into a live market. Value is judged at the moment you place the bet, against the odds available to you then. A later change in the score does not reach back and remove the value you already secured.
You can choose to run your analysis as close to kick-off as possible, when the market is at its most efficient and you can be most confident the value is real. The trade-off is that a sharper market tends to offer fewer value opportunities. Earlier in the cycle you will find more potential value, but with slightly less certainty that it will hold.
When the Sanity Score itself changes between two runs, it helps to know what is driving it. The score draws on four factors, all shown in the Sanity table on your analysis:
- Model variance
- Data freshness
- Market consistency
- Divergence assessment
Two of these are the ones to watch if the score moves significantly: Model variance and Market consistency. A shift in either suggests something has genuinely changed in the underlying data, and the new assessment is worth taking seriously.
The other two — Data freshness and Divergence assessment — are less of a concern when they move, because they relate to how recently the odds were last checked and how the model and market are aligned, rather than to a real change in the fixture itself. A drop driven by these does not undermine an assessment you have already acted on.
A common example: the first time you run an analysis, the odds have been checked recently and the data is fresh. When you re-run it later, that most recent odds check may simply have failed, lowering the Sanity Score because the data was no longer fresh, even though nothing about the fixture’s value had actually changed. In a case like that, the lower score on the re-run is not a reason to come out of a bet you have already placed: the value you took was banked at the moment you placed it.
So, to the original questions: a Sanity Score dropping on a re-run is not unusual and is often a data-freshness artefact rather than a real change. Running closer to kick-off gives you a more efficient market and more confidence, at the cost of fewer opportunities. And a bet placed on a sound, fresh assessment keeps its value regardless of what a later re-run shows.