Expected loss is the most quoted number in slot analysis and the one most often misread. Multiplying stake by spins by (1 minus RTP) gives you a true long-run average, but players treat it as a budget forecast for a single session. It is not. On high variance maths models the mean and the median tell very different stories, and the gap is where bankrolls go missing.

The formula and what each term actually measures

Start with the arithmetic. Expected loss equals total stake multiplied by (1 minus RTP). Spin 1,000 times at £1 per spin on a slot with 96% RTP and the calculation is 1,000 x 1 x 0.04, which gives £40. That £40 is not a fee, a charge or a prediction. It is the average amount retained by the operator per £1,000 wagered if you played an infinite number of spins.

The RTP figure comes from the game’s paytable and is tested across millions of simulated spins. It describes the whole distribution, not your slice of it. A 96% RTP slot returns 96p of every £1 staked over that enormous sample, and the remaining 4p is the house edge. Nothing about a single session has to land anywhere near that 4%.

Three inputs matter and each one is easy to get wrong. Stake is the total amount wagered, not your deposit. Spin count is the number of paid spins, not the number of minutes you played. RTP is the theoretical return for the specific game version, and versions differ: check the paytable in the game client rather than a review written two years ago.

Working the numbers in order

  1. Find the RTP in the game’s information screen and convert it to a decimal (96.04% becomes 0.9604).
  2. Subtract it from 1 to get the house edge (0.0396 in this case).
  3. Multiply your stake per spin by your planned spin count to get total wager.
  4. Multiply total wager by the house edge to get expected loss.
  5. Compare the result against your bankroll, not against your deposit, and ask whether you could absorb three times that figure without changing behaviour.

Run the same steps on the six 2026 Nolimit City releases, which average an RTP of 96.04%, and the edge sits at 0.0396. At £2 per spin across 500 spins the wager is £1,000 and the expected loss is £39.60. That is the honest long-run cost. The problem is what happens across those 500 spins rather than after them.

Why the mean badly understates a real session

Expected loss is an average, and averages are dominated by rare large outcomes. Picture a simplified slot: 999 sessions out of 1,000 lose £100 and one session out of 1,000 wins £50,000. The mean session loss is £50, but the median session loss is £100. Half of all players do better than the median and half do worse, and on high variance titles the median sits well below the mean in player terms, meaning you lose more than the formula suggests in the typical case.

This is not a quirk of a badly designed game. It is the direct consequence of a high variance maths model. To fund a 40,000x maximum win potential, the machine must produce long barren stretches where the RTP is not being returned to anyone. Those stretches are the cost of the tail. Duck Hunters 2 has the highest max win of the 2026 batch at 40,000x, and that ceiling is paid for by thousands of spins that return nothing.

The practical result is that a session’s realised return clusters around a number meaningfully lower than 96%. You see the RTP only if you play long enough to hit the rare events that pull the average up. Most players stop before then, which means the typical session experience is closer to 60% or 70% of stake returned than 96%. The formula is correct and still misleading, because it answers a question nobody is actually asking.

What the volatility rating changes

Volatility describes the spread of outcomes, and the rating on a game tells you how wide that spread runs. AFK Airport Security is rated high rather than very high volatility, which means its outcomes cluster more tightly around the mean. A high variance title like Duck Hunters 2 scatters outcomes far more widely, so the median drifts further from the mean and the expected loss figure becomes an even weaker guide to a single session.

Two slots with identical 96% RTP and identical expected loss can feel completely different to play. The low variance one grinds you down steadily and predictably. The high variance one does almost nothing for 400 spins and then pays 200x on the 401st. Your bankroll needs to survive the first 400 spins to reach the payoff, and the expected loss calculation tells you nothing about that requirement.

This is why Nolimit City slots pair long base-game barren stretches with complex bonus purchase menus. The purchase option is a shortcut past the barren stretch, and it is priced at the expected cost of reaching the bonus organically plus the house edge. Buying in does not reduce your expected loss, as explained at https://nolimit-city-casino.com/. It converts a variable wait into a fixed cost, which is a bankroll management decision rather than a mathematical improvement.

Running the numbers across the 2026 releases

Nolimit City released six slots between June and September 2026, and the mechanical variety makes them a useful test set for the formula. Each one needs its own RTP and structure entered into the calculation, because ways-based and pay-anywhere games distribute their returns differently across the spin count.

Ding Dong Death has a 729-ways structure and a 14,280x max win. Ways games pay for adjacent symbols across reels rather than along fixed paylines, which changes hit frequency and therefore the shape of the loss distribution. Gator Hunters 2 uses pay-anywhere mechanics, where winning combinations form anywhere on the reels. Pay-anywhere maths tends to produce more frequent small wins and rarer huge ones, which shifts the median relative to the mean in ways the simple formula cannot capture.

Outsourced 2 offers 625 ways and an 11,111x cap, while Soaked by Seamen features 225 ways and a 30,000x ceiling. The lower ways count on Soaked by Seamen concentrates the payout into fewer, larger combinations, which raises variance and pushes the median session further below the expected loss figure. Cluster pays mechanics, where you need a minimum group of connected symbols, behave similarly: fewer but chunkier hits.

Worked example on Soaked by Seamen at a stated RTP near the 96.04% average: £1 spins, 800 spins, total wager £800, expected loss £31.68. Now consider the realistic session. With a 30,000x ceiling and 225 ways, the distribution is heavily skewed. The median 800-spin session might return £560 to £640 of the £800 staked, a realised loss of £160 to £240. The formula said £31.68. Both numbers are correct; they describe different things.

That gap is the single most useful thing to understand before you open a Nolimit City review or a slot comparison tool. If a review quotes expected loss as your session budget, it is quoting the mean and calling it a forecast. Treat it as a floor on long-run cost and an underestimate of typical session cost, and your planning gets considerably more honest.

Turning the calculation into a usable budget

The formula becomes useful once you stop treating the output as a prediction. Use it to estimate long-run cost across many sessions, then apply a variance multiplier for the session itself. A rough working rule: for high volatility titles, plan to absorb three to five times the calculated expected loss without changing your stake, and for very high volatility titles, more.

Take the £39.60 expected loss from earlier. A high variance player should hold a session bankroll of roughly £120 to £200 to ride out the barren stretch without cutting stake mid-session. That is not a guarantee of survival, just a practical buffer that keeps your bet sizing stable and your decisions rational.

Bonus purchase menus deserve separate treatment. The Nolimit Bonus buy prices are set so the feature’s expected value sits at or below the base game’s RTP. Buying a bonus is a variance choice, not a value choice. If you buy, account for the full purchase price in your session budget rather than treating it as a separate pot.

Before staking real money, use slot demo play. The site provides free demos and a mobile app for every slot, which lets you run the mechanics and see the barren stretches for yourself. Demo play will not reveal the RTP in a session, but it will show you how long the dry spells run and whether your planned bankroll survives them. A slot promo code or a welcome offer such as WELCOME2026 at a Nolimit City Casino extends playtime but does not change the underlying maths.

Safe online play comes down to three habits: set the session budget from the variance-adjusted figure, stop when it is gone, and never chase a session that has drifted below the median. The expected loss formula is a planning tool for the long run and a poor guide to tonight. Use it to know what the game costs over hundreds of sessions, then budget for the session itself as though the median is the reality, because for most players it is.