Value notes
Which drop percentage is worth your attention
There is no universal right drop threshold. How to calibrate your percentage from your own alert log, by sport, and when to change it.
Ask ten bettors what drop percentage matters and you will get ten numbers. The honest answer is that the right threshold is the one your own log justifies: low enough to catch real moves early, high enough that you can act on what survives. Here is how to find it from your own data instead of copying someone else's number.
What the percentage actually measures
A drop is the fall divided by the old price: 2.37 to 2.25 is 5.06%, the same arithmetic the dropping odds calculator applies, checked 2026-10-01. The percentage form exists so a rule treats a 1.50 favorite and a 5.00 outsider consistently, but the same percentage on those two prices is not the same event. The probability shift, the points of implied chance the move represents, is the truer measure of how much the market's view changed.
So before choosing a number, know what you are choosing it in. A 5% rule means: tell me when the market repriced this side by five percent of its old price. No more than that.
Calibrate from your own alert log
Run the feed wider than you think you need for a week. Set the threshold low, near the documented floor of 1, and log every alert with its fields: sport, market, the two prices, the fair price and the time.
After a week, replay the log against candidate thresholds. For each candidate, count how many alerts would have fired, and for each of those, check whether the price still on offer at your book beat the fair price when the alert arrived. The EV calculator does that comparison per bet; here you are doing it in bulk over history.
You are looking for the knee: the threshold where raising it further stops cutting noise and starts cutting value. Below the knee you drown in moves you would never bet. Above it, you miss ones you would.
Expect the number to differ by sport
A threshold is a property of a market's rhythm. Soccer prematch prices drift slowly, and a 3% move can be news. A live tennis price can swing that much on a single game. If your log mixes sports, calibrate per sport, and run live and prematch separately: the volume and cadence differ too much for one number to govern both.
Change it on evidence, not on a bad day
Pick a review rhythm, weekly or monthly, and change the threshold only when the log says so. One noisy afternoon is not evidence. Neither is one missed winner.
When you do change it, change it in one place and note the date. A threshold with a history you can read is worth more than a perfect number you cannot explain.
The feed delivers every move past your line. The log tells you where the line belongs.