Smart Parlay Builder

Acca Builder

Build football accumulators with live expected value, Kelly-optimised staking, and leg-by-leg colour-coded value detection.

Quick Fill Presets
Leg
-EV
Leg EV: -2.0%|Fair odds: 1.43
Leg
-EV
Leg EV: -0.5%|Fair odds: 1.54
Total Odds
2.14
Fair Odds
2.20
Fair Probability
45.5%
EV %
-2.5%
Payout
107.10
Full Kelly
0.0%
Half Kelly
0.0%
Bookie Implied
46.7%
Vig / Leg
68.4%

Negative EV. The bookmaker has the edge here. Consider different selections or wait for better odds.

Leg-by-Leg EV Breakdown

SelectionOddsModel ProbFair OddsLeg EV
Man City Win1.4070.0%1.43-2.0%
Arsenal Win1.5365.0%1.54-0.5%

How Acca Math Works

An accumulator (parlay) multiplies the decimal odds of each leg together. If you combine three selections at 1.50, 2.00, and 1.80, the total odds are 1.50 x 2.00 x 1.80 = 5.40. A $50 stake would return $270 if all three win.

The implied probability of the parlay is 1 / 5.40 = 18.5%. That means the bookmaker thinks you have an 18.5% chance of winning. Your model probability is your own estimate — if you believe each leg has a higher chance than the odds suggest, the parlay has positive expected value.

Vig compounds in parlays. If each leg has 5% vig, the total vig on a 5-leg parlay is roughly 1 - (0.95)^5 = 22.6%. This is why bookmakers love accumulators — the edge compounds exponentially.

Kelly staking for parlays uses the formula: f* = (p x O - 1) / (O - 1), where p is your model's parlay probability and O is the bookmaker's total parlay odds. Because parlays have higher variance, most professionals use fractional Kelly (1/4 or 1/2 of the full Kelly amount).

Acca Formula:
  Total Odds = O1 × O2 × ... × On
  Fair Prob  = P1 × P2 × ... × Pn
  EV = Fair_Prob × Total_Odds × Stake - Stake
  Kelly f*  = (Fair_Prob × Total_Odds - 1) / (Total_Odds - 1)

Example (3-leg):
  Leg 1: 1.50 × 70%     Leg 2: 2.00 × 52%     Leg 3: 1.80 × 58%
  Total odds = 5.40      Fair prob = 21.1%
  EV = 0.211 × 5.40 × 50 - 50 = +7.0  (+14.0%)
  Full Kelly = (0.211×5.40-1)/(5.40-1) = 6.5% of bankroll