Doesn't work

Betting on favourites, underdogs and draws: which pays better

“Only back favourites” does not work: they lose less than the rest (−4.1%), but they still lose. Bookmakers earn the most on underdogs (−10.1%).

  • −4.1%backing the favourite in every match
  • −10.1%backing the underdog
  • −21.7%on outcomes at odds of 10+

How the margin is spread across the odds

Betting results by odds1.00–1.30: −1.9%; 1.30–1.50: −0.4%; 1.50–1.80: −2.6%; 1.80–2.20: −4.7%; 2.20–3.00: −4.9%; 3.00–4.00: −5.9%; 4.00–6.00: −12.3%; 6.00–10.00: −19.0%; 10+: −21.7%1.00–1.30: −1.9%1.00–1.30−1.9%1.30–1.50: −0.4%1.30–1.50−0.4%1.50–1.80: −2.6%1.50–1.80−2.6%1.80–2.20: −4.7%1.80–2.20−4.7%2.20–3.00: −4.9%2.20–3.00−4.9%3.00–4.00: −5.9%3.00–4.00−5.9%4.00–6.00: −12.3%4.00–6.00−12.3%6.00–10.00: −19.0%6.00–10.00−19.0%10+: −21.7%10+−21.7%
Every outcome of every match at an average bookmaker's closing odds, a stake of 1 on each.

The higher the odds, the more they are cut. Outcomes at 1.30–1.50 come out almost level (−0.4%), while those at 10 and above end at −21.7%. It is a long-known effect, the favourite-longshot bias: people are more willing to overpay for a small chance of a big win, and bookmakers build more margin into long odds.

OddsBetsWonImplied by the oddsResult
1.00–1.302,19381.5%83.1%−1.9%
1.30–1.504,16371.3%71.6%−0.4%
1.50–1.808,77659.0%60.6%−2.6%
1.80–2.2013,36847.9%50.2%−4.7%
2.20–3.0025,67136.8%38.8%−4.9%
3.00–4.0046,43627.5%29.3%−5.9%
4.00–6.0020,24718.8%21.4%−12.3%
6.00–10.006,85311.2%13.8%−19.0%
10+2,1575.9%7.6%−21.7%

“Implied by the odds”: the probability built into the odds, margin included (1 / odds).

So should you just back favourites?

No. Backing the favourite in every match wins 51.0% of the time, but the result is still −4.1%. Even in the best range (1.30–1.50) the margin eats everything the market's rare mistakes give. Favourites lose less than underdogs — but they still lose.

We also tested a bet on our model's pick in every match — that is how the model runs the “flat stake” live in Premium. Its pick came in 49.9% of matches, but the result at the average price is −4.7%, and at the best price on the market −0.6%. The model is often right, but the odds on its picks already take the same things into account.

Favourites in hockey and basketball

The picture is the same: in hockey backing the favourite gave −4.3% and the underdog −10.5% while in basketball −4.9% and −8.1%. Here too the bookmaker earns most on underdogs.

SportFavourite: wonFavourite: resultUnderdog: wonUnderdog: result
Football51.0%−4.1%21.9%−10.1%
Hockey65.0%−4.3%35.0%−10.5%
Basketball68.8%−4.9%31.2%−8.1%

The bottom line

  • The margin is smaller on favourites and biggest on underdogs and long shots.
  • The simple rule “back the favourite” still loses: there is no edge over the market in it.
  • Betting at long odds needs an especially accurate estimate of the chance: that is where the margin is biggest.

This is an analysis of history, not advice to bet. Betting means risking money; 18+. The numbers: 43,288 completed matches in 27 football leagues between 15 July 2017 and 12 October 2026; bets on the match result (home win, draw, away win) after 90 minutes. In hockey and basketball the winner includes overtime; a hockey draw is the one after regulation time.

Other strategies