Two exported prices both display as 2.00. Your filter includes one and excludes the other. Before assuming the filter is broken, inspect the precision of the stored values and whether the comparison uses the displayed or underlying number.
This issue matters only where data is actually available at greater precision than the display. Do not invent hidden digits for a bookmaker quote that was supplied at two decimal places. The examples below use constructed values to illustrate the distinction.
Apply a threshold before rounding the display
Suppose the research rule is “decimal odds at least 2.00”. The original records are 1.996, 2.000 and 2.004. Rounded to two decimals, all three display as 2.00. Applied to the original values, the rule includes the second and third records only. Applied after rounding, it includes all three.
Neither a screenshot nor a two-decimal report tells you which comparison took place. Preserve the supplied precision in the calculation and document the display rounding separately. If the original data contains only 2.00, do not claim you can recover the earlier digits.
Separate inclusive and exclusive bounds
“Between 1.80 and 2.00” is incomplete unless the treatment of the endpoints is known. In an illustrative set containing exactly 1.80, 1.90 and 2.00, inclusive bounds keep three records; strict interior bounds keep only 1.90. Record the operator rather than relying on conversational wording.
For a saved filter, inspect the service's actual criterion definition. If you use an AI assistant, ask it to repeat the bounds and endpoint rule before running the query. Do not assume that every interface uses the same meaning of “between”.
Round a probability at the end
The decimal-odds conversion is 100 divided by the price. At 1.996 the raw implied probability is approximately 50.1002%; at 2.004 it is approximately 49.9002%. Rounding both prices to 2.00 first turns both into exactly 50%. That erases a difference of roughly 0.2000 percentage points.
These are price conversions, not independently measured chances of winning. Bookmaker margin and market definition remain relevant. The precision example cannot establish that either side of the threshold is a profitable selection.
Record the transformation order
A short method note is enough: “Filter on source precision; inclusive lower bound 2.00; compute raw implied probability from the original price; round reported percentages to two decimals.” If another system only exposes rounded values, state that limitation and compare the common available precision honestly.
When two reports disagree by a few rows, check precision, endpoints and missing-value handling before looking for a football explanation. A filter boundary is part of the method, even when the difference is invisible in the displayed table.
Sources and method
Original research examples and prompts, not observed match results or performance claims. Prepared with AI assistance; source statements, examples and calculations were checked.
- Smarkets: converting decimal odds to implied probability — accessed 13 October 2026