+EV Betting on MLB Props: Calculating Real Edge, Not Hype

The First Time +EV Made Sense to Me

I spent two years thinking +EV was a marketing word the analytics community used to make their picks sound smarter than mine. Then I sat down on a slow February afternoon, ran the maths on a single strikeout prop I had been backing for months, and realised the prop had been negative-EV every time I had taken it. The hitter angle was right. The pitcher angle was right. The price was wrong. I had been doing the analysis backwards: starting with the bet I liked and working out a story for it, instead of starting with my own probability and asking whether the price actually paid me to take the bet.

That was the day +EV stopped being a slogan and started being a workflow. The shift is small but the consequence is enormous. Once you have a personal probability for a prop, the decision to bet or pass is a single calculation. The decision is not “do I like this play?” – it is “does the price pay me more than my probability says it should?”

What Expected Value Actually Means

Expected value is the average outcome of a bet across infinite trials at the current price. If a prop pays out at decimal 2.20 and your personal probability of winning is 50 per cent, the expected value of a £10 stake is £1 – you win £12 half the time and lose £10 half the time, averaging a £1 return per bet. That £1 is your edge. Across hundreds of bets at the same edge, the £1 per £10 staked compounds into a meaningful long-run return.

The maths is plain. EV equals (probability of winning) × (decimal odds − 1) − (probability of losing) × 1. If the result is positive, the bet is +EV. If negative, it is −EV. For the strikeout prop priced at 1.91 with a personal probability of 55 per cent, the calculation gives 0.55 × 0.91 − 0.45 × 1 = 0.5005 − 0.45 = 0.05, which is +5 per cent EV per unit staked. That is the kind of edge a disciplined prop bettor can find consistently across a 162-game season.

Notice what +EV is not. It is not the same as a hot pick or a confidence rating. A bet can be high-confidence and −EV if the price is short enough. A bet can be low-confidence and +EV if the price is long enough. The market sets the price; the punter’s job is to decide whether the price is generous relative to the underlying probability. Confidence is a feeling, EV is a number.

Devigging Prop Prices

The first practical step in any +EV workflow is removing the bookmaker’s overround from the price to get a fair-market probability. UK prop hold percentages tend to land in the 8 to 15 per cent band, which is wider than the 5 to 8 per cent typical on main markets. That overround is what stands between the listed price and the true implied probability of the event.

The devig is straightforward when you have both sides of a two-way market. Convert each price to its raw implied probability – 1 divided by decimal odds gives the implied probability for that side. The two probabilities will sum to more than 1; the excess is the overround. Divide each raw probability by the sum to get the devigged fair probability. For a strikeout over priced at 1.83 and an under priced at 1.91, the raw implieds are 54.6 per cent and 52.4 per cent, summing to 107 per cent. The devigged probabilities are 51.0 per cent for the over and 49.0 per cent for the under. That 51 per cent is the market’s fair estimate of the over hitting; if your own model says the over is 55 per cent, you have a 4-point edge in probability terms.

The devig is not perfect. Books shade some markets in one direction or the other, and a clean devig assumes both sides are priced honestly. For most player props at established UK books, the devig is a good first approximation of fair probability and is the right starting point for any +EV workflow.

Building Your Own Projection

The harder half of +EV work is building a personal probability for the event. There is no single way to do this well, but the components are usually similar. For a strikeout prop the inputs are the pitcher’s K-rate baseline, the opposing lineup’s K-rate against that pitcher’s handedness, the projected pitch count, the umpire’s strike-zone tendency, and the park environment. For a HR prop the inputs are the hitter’s barrel rate and ISO, the pitcher’s HR rate and pitch profile, the park factor, the wind, and the lineup spot. For a hits prop the inputs are batted-ball quality, BABIP regression, lineup spot for plate appearances, and matchup handedness.

The level of sophistication varies by punter. A spreadsheet model that combines five or six factors with sensible weights produces probabilities that are good enough to identify edges most weeks. A more elaborate Monte Carlo simulation produces tighter estimates but takes considerably more work to build and maintain. For most UK punters the spreadsheet model is the right place to start, and the marginal value of going beyond it is smaller than the time cost suggests.

The honest part of building your own probability is calibration. The probabilities your model produces should be tested against actual outcomes across a meaningful sample. If your model says 55 per cent and the bets at 55 per cent are winning at 48 per cent, the model is over-confident and the EV calculation is reading a phantom edge. Calibration discipline is what separates a working +EV workflow from a story-driven one.

Realistic Edge Thresholds

The realistic +EV percentage on MLB player props for a careful UK punter is in the 2 to 5 per cent range. Anything above 5 per cent should be treated with suspicion – it usually means the price is wrong because the bookmaker has information you do not, not because you have spotted something they missed. The classic case is a late-breaking lineup change or an injury that has not yet hit the public news cycle. Books incorporate that information into prices faster than most public sources.

Inside the 2 to 5 per cent band the bets are usually genuine edges. They are not glamorous. They will not win at 70 per cent or pay 5/1. They will win at the rate the model says and pay at the price the book offers, and the small per-bet edge will compound across a long sample into a meaningful return. The Dimers backtest of 9.6 per cent return across 108 outs-recorded bets is consistent with what a disciplined +EV workflow can produce on a single market when the analysis is sound.

The threshold I personally use to bet is +3 per cent EV. Below that the edge is too easily eroded by line movement, sub-optimal price-shopping, or model error. Above that the bet earns its place in the slip. The threshold should be calibrated to your own confidence in your probabilities; a punter with a less-tested model should set the threshold higher to account for uncertainty in their own inputs.

The Workflow That Actually Travels

The +EV workflow that survives a full season looks the same most days. Pull the slate. For each prop you would consider, write down your probability before looking at the price. Convert the price to a devigged fair probability. Calculate your EV. Bet only the props above your threshold. Track every bet by market, edge, and result. Review monthly to check that probabilities are calibrated.

The discipline is in the order. Probability first, price second. Punters who reverse the order – pricing first, probability fitted to it – produce −EV bets dressed up as confident plays. The probability has to come before the price for the EV calculation to mean anything. Once that order is locked in, the workflow does the work.

Published by the BasePropPro team.