Novus Odds
Sweep player lines in the Props Lab
The Props Lab samples fictional player distributions and sweeps a line across them, producing over/under EV curves that show how sharply value changes with half a unit.
Player props look simple — over or under a number — and are the part of betting mathematics where intuition fails hardest. The reason is that the answer depends on the shape of a distribution near a single point, and human intuition about distributions near a point is famously bad.
The Props Lab generates fictional player distributions and sweeps a line across them, so you can watch expected value swing as the line moves. Everything is synthetic: no real players, no real lines, no real prices.
Contents
Two ways to finish
Shape first
Understand the distribution before you argue about the line.
Sweep the line
Watch EV change as the number moves half a unit at a time.
- 1
1. Generate a distribution and look at its shape
Start by generating a player distribution and reading it as a shape rather than as an average. Two players with identical means can behave completely differently around a line if one is consistent and the other is volatile, and props are decided entirely by that behaviour.
Pay attention to skew. Many performance distributions have a floor at zero and a long right tail, which means the mean sits above the median and a line placed at the mean is not the coin flip it appears to be.
- Mean and median diverge whenever a distribution is skewed.
- Variance decides how much a half-unit line move is worth.
- A line at the mean is rarely a 50/50 proposition.
- 2
2. Sweep the line and read the EV curve
Now move the line across the distribution and watch the over and under EV curves. The curves cross somewhere, and the interesting thing is how steeply. A sharp crossing means small line differences matter enormously; a shallow one means the market has room to be sloppy without consequence.
This steepness is the practical lesson of the whole lab. In a high-variance distribution the curves are shallow and a half-unit is close to irrelevant; in a tight one, half a unit can flip a proposition from clearly good to clearly bad.
- 3
3. Add the price and watch value disappear
Probability is not value. Once you attach a price with a margin in it, a lot of propositions that looked mildly favourable stop being favourable at all. Run the same sweep with the overround set realistically and see how much of the curve survives.
This is the same lesson the Odds Lab teaches, but props make it sharper, because prop margins are typically wider than main-market margins. A view that is right slightly more often than not is usually not enough.
- 5
5. Check sample size before believing a curve
A distribution built from too few samples produces an EV curve that looks authoritative and moves substantially when you change the seed. Raise the sample count and rerun before drawing conclusions from any particular shape.
The Statistics pages carry the tools for this — 95% Wilson intervals, a proportion z-test, and a chi-square goodness-of-fit test — and they are worth applying to a prop conclusion before you treat it as one.
- 6
6. Keep the fiction in view
None of this is a real player. The distributions are generated, the lines are invented, and the prices are synthetic. The transferable output is understanding — how distribution shape and line placement interact, and how much margin eats — not a number to take anywhere.
Novus Odds is explicit about that boundary: it does not accept wagers, place bets, operate gambling services, or guarantee outcomes.
The shape decides, not the average
Read the distribution before the line. Sweep the line and note how steeply the EV curves cross — that steepness tells you how much half a unit is worth. Then add a realistic margin and see how much of the apparent value survives, because in props it usually does not.
Frequently asked questions
Quick answers to common questions about this topic.
Are these real players?
No. The Props Lab samples fictional player distributions. Nothing in Novus Odds is a real player, team, fixture, or price.
Why does distribution shape matter so much?
Because a prop is decided by behaviour around a single point. Two players with the same mean but different variance or skew produce very different over/under propositions at the same line.
Why does my apparent edge vanish when I add margin?
Because probability is not value. Prop margins are typically wider than main-market margins, so being right slightly more often than not is usually not enough to overcome the price.
How many samples do I need?
Enough that the EV curve does not change materially when you change the seed. Use the Statistics pages — Wilson intervals, a proportion z-test, and a chi-square test — to check rather than eyeballing it.