The acca that changed my thinking about parlay strategy was not a winner. It was a four-leg ATTS accumulator that landed three out of four on a Sunday afternoon in November, paid nothing, and cost me a week’s worth of carefully researched singles profit. I sat there watching the fourth player – a running back who had scored in four straight games – get pulled for a third-down back on the only red zone possession where his team needed a touchdown. That was the moment I realised I had been approaching parlays backwards. I was building them because the payouts looked exciting, not because the maths supported them.
This guide is about the maths. Parlays and accumulators are the most popular way UK bettors combine touchdown scorer picks, and they are also the fastest way to drain a bankroll if you do not understand what you are buying. Every parlay introduces a structural cost – the bookmaker’s margin compounds with each leg you add, and that compounding erodes your expected value faster than most punters appreciate. But parlays are not inherently bad bets. Used selectively, with the right number of legs and the right type of selections, they can serve a legitimate purpose in a disciplined approach.
I am going to walk through how parlays actually work mathematically, when they make sense, when they destroy value, and how to size them properly within a bankroll framework. If you have been stacking five-leg accas every Sunday because the potential payout looks impressive, prepare to reconsider.
How NFL Touchdown Parlays and Accumulators Work
A mate at the pub once told me he “understood” parlays because he had been placing them for years. When I asked him to explain exactly how a three-leg acca’s odds were calculated, he went quiet. Understanding the mechanics is not optional – it is the foundation for every strategic decision that follows.
A parlay – called an accumulator or “acca” in UK betting – combines two or more individual selections into a single bet. All legs must win for the bet to pay out. The combined odds are calculated by multiplying the decimal odds of each leg together. If you back three ATTS picks at 2.50, 3.00, and 2.25, the parlay odds are 2.50 x 3.00 x 2.25 = 16.875. A 10-pound stake returns 168.75 pounds if all three players score.
The multiplication is where the trouble starts. Each individual set of odds already includes the bookmaker’s margin – typically 5% to 15% on ATTS markets. When you multiply three sets of odds that each contain a margin, you are not just paying one margin; you are paying the margin three times over, and the compounding effect means the total margin grows faster than linearly. A three-leg parlay with legs that each carry a 10% margin does not cost you 30% – it costs you closer to 27% when you calculate the true expected return against fair odds. By the time you reach five legs, the effective margin can exceed 40%.
Cross-game accumulators – combining ATTS picks from different games on the same slate – work on straightforward multiplication because the outcomes are genuinely independent. What happens in the early London game has no bearing on the late kick-off in San Francisco. This independence means the odds multiply cleanly, and the only cost is the compounding margin. Same-game parlays and bet builders involve additional adjustments for correlation, which I will cover in a later section.
The key distinction to internalise is that parlays do not create value. They rearrange the payout structure by concentrating a smaller number of large wins against a larger number of complete losses. If each individual leg has a negative expected value, the parlay has a worse expected value. If each leg has a positive expected value, the parlay also has a positive expected value – but typically a smaller one than betting each leg as a single. Parlays are a packaging choice, not an edge generator.
The Expected Value Maths Behind TD Parlays
Expected value is the number that tells you whether a bet is worth placing, and most punters never calculate it. Here is how it works, stripped of complexity.
Expected value (EV) equals the probability of winning multiplied by the net profit, minus the probability of losing multiplied by the stake. For a single ATTS bet at decimal odds of 3.00 on a player you estimate has a 40% chance of scoring: EV = (0.40 x 2.00) – (0.60 x 1.00) = 0.80 – 0.60 = +0.20. That is positive EV of 20p per pound staked. Over time, this bet makes money.
Now take two such bets and combine them in a parlay. Assuming independence, the combined probability is 0.40 x 0.40 = 0.16, or 16%. The parlay odds are 3.00 x 3.00 = 9.00. EV = (0.16 x 8.00) – (0.84 x 1.00) = 1.28 – 0.84 = +0.44. The parlay is still +EV, and the absolute EV per unit staked is actually higher than a single bet. So why am I cautious about parlays?
Because that calculation assumed each leg is +EV, and it assumed the bookmaker margin is already baked into the 3.00 odds. In reality, if the true fair odds on each leg are 2.50 (40% probability) and the bookmaker offers 2.25, each leg is -EV. The parlay with two -EV legs is even more -EV: you lose more per pound staked, not less. The margin compounds.
The 73.9% of NFL touchdowns scored inside the red zone concentrates scoring among a relatively small pool of players, which means true ATTS probabilities are often higher than implied odds suggest for the right players. This is where +EV singles exist, and it is where +EV parlays can exist too – but only if every single leg in the parlay is individually +EV. One -EV leg in a three-leg parlay can turn the entire combination negative even if the other two legs are strongly positive.
The practical rule I follow: never include a leg in a parlay that I would not bet as a single. If the pick does not pass my standalone analysis – red zone share, snap count trend, matchup, price assessment – it has no business being in an accumulator. The parlay is a vehicle, not a filter. The filter is the analysis; the parlay is just how you choose to combine the results.
Same Game Parlays and Bet Builders for TD Scorers
Same-game parlays and bet builders have exploded in popularity since UK bookmakers began offering them around 2019. The appeal is obvious – you can combine an ATTS pick with a yardage prop, a game total, and a team result all within a single match, creating a customised bet with eye-catching odds. The problem is equally obvious once you look at the pricing.
Bet builders apply a correlation adjustment to the combined odds, and that adjustment almost always works against the bettor. When you combine an ATTS pick with the game going over the total, the bookmaker recognises the positive correlation – more touchdowns naturally push the total higher – and reduces the combined payout accordingly. Fair enough; the correlation is real. But the size of the reduction is opaque. Unlike a standard accumulator where you can see each leg’s individual price and verify the multiplication, bet builder pricing is a black box. The bookmaker inputs the selections and outputs a combined price, and you have no way to audit the correlation adjustment or the margin embedded within it.
I tested this over a 12-week period by calculating what a two-leg bet builder should pay based on the individual leg odds and a reasonable correlation estimate, then comparing it to the actual bet builder price offered. The bet builder price was consistently 15% to 25% lower than my fair-value estimate. That is a staggering margin, and it explains why bookmakers promote bet builders so aggressively – they are among the most profitable products in the industry.
This does not mean bet builders are never worth using. For entertainment bets with small stakes where the process matters more than the outcome, they are fine. For occasional promotional offers where the bookmaker boosts the bet builder price above fair value, they can actually be +EV. But as a regular component of a serious ATTS strategy, bet builders drain value faster than almost any other bet type. I use them sparingly – two or three times per season at most, and only when a promoted boost or insurance offer shifts the expected value into positive territory.
For a deeper look at which specific bet builder combinations offer the best structural value and which to avoid, the bet builder strategy guide breaks down the maths for the most common ATTS combinations.
How Many Legs? Optimal Parlay Sizing
How many legs should a touchdown parlay have? I have been asked this question more than any other at NFL betting meetups, and the answer is simpler than people expect: as few as possible.
Every additional leg in a parlay reduces the probability of the bet winning. A two-leg parlay with each leg at 50% probability has a 25% chance of winning. Three legs drops it to 12.5%. Four legs: 6.25%. Five legs: 3.125%. At the same time, each additional leg multiplies the bookmaker’s margin into the combined price, eroding expected value even if every individual leg is +EV. The maths is straightforward: more legs equals lower win probability, higher margin drag, and worse expected value per unit staked.
The NFL averaged 18.7 million viewers per game during the 2025 season, and the explosive growth of parlay betting is one reason why. Five-leg accas at 50/1 are exciting television. They are also terrible investments. I cap my ATTS parlays at three legs, and the majority of my parlays are doubles – two-leg combinations. At two legs, the margin compounding is manageable, the win probability remains reasonable, and the payout uplift over singles is meaningful enough to justify the format.
There is a specific scenario where I extend to three legs: when I have three strongly +EV singles on the same slate that I am already betting individually, and I want a small additional exposure to the possibility of all three hitting. In that case, I place the three singles at my normal stake and add a three-leg parlay at a reduced stake – typically 20% of my single-bet stake. This approach captures most of the expected value through singles while adding a lottery-ticket-style upside through the parlay without meaningfully increasing risk.
Four legs and above? Never. The margin drag is too severe, and the win probability drops below the threshold where variance can be managed within a seasonal timeframe. A bettor placing one four-leg acca per week would need over a year of consistent +EV selections just to have a reasonable chance of their results reflecting their edge. Life is too short, and bankrolls are too finite.
Bankroll Allocation for Parlay Betting
Roughly 10% of the UK population bets on sport online, and based on my conversations with dozens of recreational NFL bettors over the years, the vast majority of them have no bankroll management system at all. They deposit 50 pounds when they feel like betting, scatter it across a few accas, and either win big or lose the lot. That is not a strategy; it is entertainment with a variable admission fee.
A proper bankroll for parlay betting starts with a total seasonal allocation. I set aside a fixed amount at the start of the NFL season – call it my “prop bankroll” – and that amount does not change regardless of results during the year. Within that bankroll, I divide my exposure between singles and parlays at a ratio of roughly 85% to 15%. The overwhelming majority of my prop bets are singles, and the parlay allocation is a deliberate minority position.
For singles, I stake 1% to 2% of the prop bankroll per bet. For parlays, I stake 0.25% to 0.5% per bet. The lower parlay stake reflects the higher variance – a two-leg parlay has roughly double the variance of a single, and a three-leg parlay has roughly four times the variance. Sizing the stakes proportionally to the variance keeps the impact of any individual loss consistent across bet types, which is the whole point of bankroll management.
The UKGC conducted 9,700 compliance actions during the 2024/25 period, and the new financial vulnerability checks triggered at 150 pounds net spend in 30 days mean that operators are paying closer attention to betting patterns than ever before. These checks are not a hindrance to disciplined bettors – they are a safety net. If your parlay staking is proportional and controlled, the checks will not affect your activity. If they do flag your account, it is worth examining whether your staking has drifted from the plan.
One final note on parlay bankroll management: keep your parlay results tracked separately from your singles results. It is tempting to lump everything together, but the variance profiles are so different that combining them obscures your performance in both areas. I maintain separate columns in my tracking spreadsheet for singles and parlays, and I evaluate the parlay component only after a full season of data. Trying to judge your parlay approach after eight weeks is like judging a cricket batsman’s average after two innings – the sample is far too small for the variance involved.
