Can I reuse these figures?

Yes — journalists, researchers and educators are welcome to quote any number on these pages with attribution and a link. Each study publishes its random seed, exact rules and payout tables, so every figure can be independently reproduced or challenged. We will also run custom cuts of a simulation on request.

Why should anyone trust numbers published by an affiliate site?

You should not trust them — you should check them. That is why we publish the seed, the model and the code notes rather than only the conclusions. We have a commercial interest in this subject, and the correct response to that is verifiability, not assurances. Several of our findings are actively unhelpful to gambling operators.

What have these studies found that contradicts industry marketing?

Three things in particular. Bonuses are not free money: only 16.16% clear a 40x requirement and the average claimant loses 30.8% of their deposit. Betting systems do not work: all four tested lost exactly the house edge per dollar wagered. And RTP does not tell you how long your money lasts: identical 96% RTP models differed 16x in session length.

Are the simulations verified against known mathematics?

Where a theoretical value exists, yes, and we publish the comparison. Our measured baccarat banker edge was 1.04% against a published ~1.06%; blackjack basic strategy measured 0.44% against an analytic 0.43–0.45%; Plinko's 1000x slot hit 31 times in a million drops against a binomial expectation of 30.5. Agreement with known maths is how you know the engine is sound before trusting it where no closed form exists.

data

Our Casino Data Research: Every Study, Every Number, Every Seed

An index of every Monte Carlo study we have published: 44 million game rounds, 1.3 million bonus claims, 400,000 betting-system sessions and 450,000 slot bankroll sessions. Every model, seed and limitation is public, and the figures are free to reuse with attribution.

By Casino Aurum Editorial Team · How We Rate

Published: 2026-08-04 · Updated: 2026-08-04 · 3min read

Key takeaways

We have published four original Monte Carlo studies covering 46 million simulated rounds, sessions and claims — each with its random seed, exact rules and limitations stated so anyone can reproduce or refute them.
Blackjack basic strategy measured 0.44% over 8 million hands, and 'never bust' tested worse (5.94%) than blindly copying the dealer (5.68%) — the cautious-feeling strategy is the most expensive one.
Only 16.16% of players clear a 40x bonus wagering requirement on slots, and the average claimant ends 30.8% down on their deposit.
Every betting system tested — Martingale, D'Alembert, Fibonacci — lost 2.68–2.72% of everything wagered, which is European roulette's 2.70% house edge. Progressions move turnover, never the rate.
At identical 96% RTP, $100 lasted a median of 2,315 spins on a low-volatility slot model and 147 on a high-volatility one — a 16x gap from variance alone.

Most gambling sites publish opinions. We publish simulations, with the seed attached, so you can check us instead of believing us.

This page indexes every study we have run. Together they cover around 46 million simulated rounds, sessions and claims. Each one states its model, its random seed and — importantly — what it does not prove.

The studies

Study	Scale	The headline finding
44 million casino bets	44,000,000 rounds	Published RTP held up everywhere we could measure it. Blackjack basic strategy: 0.44%; playing by feel costs 13x more.
1.3 million bonus claims	1,300,000 claims	Only 16.16% clear a 40x wagering requirement; the average claimant ends 30.8% down on their deposit.
400,000 betting-system sessions	400,000 sessions	Martingale, D’Alembert and Fibonacci all lost 2.68–2.72% of turnover — the house edge, unchanged.
450,000 slot bankroll sessions	450,000 sessions	At identical 96% RTP, $100 lasted 2,315 spins or 147 depending only on volatility.

The measured house edge of every major game, pulled from these studies into one ranked table, is in casino house edge statistics.

Findings that work against our own industry

An affiliate site has an obvious incentive to make gambling look better than it is. These are the results that cut the other way, and they are the ones we consider most worth citing:

Bonuses are a discount with conditions, and often a bad one. At 40x on slots the average claimant finishes with $69.23 of the $100 they deposited. A $20 no-deposit bonus is worth about $7.99 in practice.
The reason your bonus excludes blackjack is measurable. The same 40x offer returns +22.8% on a 1% house-edge game and −30.8% on 4%-edge slots. That 53-point swing is what game weighting exists to prevent.
No staking system has ever worked and now there is a number for it. Every progression we tested lost the house edge per dollar wagered, to within 0.02 points.
“Never bust” is the worst blackjack strategy we tested — 5.94%, worse than blindly mimicking the dealer at 5.68%, and 13x the cost of basic strategy.
A 97.95% hit rate can still be a losing game. Most of those hits return half the stake.

How we run them

Every study follows the same rules, because they are what make the numbers checkable:

Fixed seed, published. Same seed, same figures, on any machine.
The model is stated in full — exact payout tables, probabilities, drawing rules, stopping conditions. Where a target RTP or house edge is claimed, it is asserted in code, not assumed.
Validate against known maths first. Baccarat, blackjack and Plinko all have closed-form or published values; our engine reproduces them before we trust it anywhere else.
Single runs, reported whole. No re-rolling for a better headline.
A limitations section on every study, stating what the model omits.

Using this research

Quote anything here with attribution and a link. If you need a cut we have not published — a different wagering multiplier, another volatility profile, a specific rule set — ask and we will run it.

Disclosure: Casino Aurum is an online-casino review site with commercial relationships with some operators. That is precisely why the methodology is public: our conclusions should be checkable rather than taken on trust, and several of the findings above are unhelpful to the businesses we partner with. A commercial relationship never changes a rating or a measured figure.

FAQ

Can I reuse these figures?: Yes — journalists, researchers and educators are welcome to quote any number on these pages with attribution and a link. Each study publishes its random seed, exact rules and payout tables, so every figure can be independently reproduced or challenged. We will also run custom cuts of a simulation on request.
Why should anyone trust numbers published by an affiliate site?: You should not trust them — you should check them. That is why we publish the seed, the model and the code notes rather than only the conclusions. We have a commercial interest in this subject, and the correct response to that is verifiability, not assurances. Several of our findings are actively unhelpful to gambling operators.
What have these studies found that contradicts industry marketing?: Three things in particular. Bonuses are not free money: only 16.16% clear a 40x requirement and the average claimant loses 30.8% of their deposit. Betting systems do not work: all four tested lost exactly the house edge per dollar wagered. And RTP does not tell you how long your money lasts: identical 96% RTP models differed 16x in session length.
Are the simulations verified against known mathematics?: Where a theoretical value exists, yes, and we publish the comparison. Our measured baccarat banker edge was 1.04% against a published ~1.06%; blackjack basic strategy measured 0.44% against an analytic 0.43–0.45%; Plinko's 1000x slot hit 31 times in a million drops against a binomial expectation of 30.5. Agreement with known maths is how you know the engine is sound before trusting it where no closed form exists.

The studies

Findings that work against our own industry

How we run them

Using this research

FAQ

Sources

Related guides