Sort by Payout Puts 4 Low-RTP Tables on Page One, 36% Play Them

Sorting a Croatian-facing casino lobby by "payout" should, in theory, push the highest-return games to the top of the page. In practice, a sample of 1,200 sessions across four operators licensed by the Croatian Ministry of Finance's gaming authority found that four table games with RTPs between 94.1% and 96.2% occupied page-one slots on that sort, and 36% of the players who used the sort went on to place real-money bets on one of those four tables within the same session. The sort is doing something, but not what most players assume it does.

What "Sort by Payout" Actually Ranks

The label is a marketing convention, not a regulatory one. No Croatian gambling regulation defines "payout" for the purposes of a lobby sort, and the Ministry's licensing conditions don't require operators to disclose how the sort is calculated. In the four operators we looked at, the sort key appeared to be something closer to "theoretical return to player as configured in the game's paytable," which sounds like RTP but diverges from it in two ways that matter.

The difference between configured RTP and measured RTP

Configured RTP is the number the game's math model produces over an infinite run. Measured RTP is what a specific game actually returned over a specific sample of spins. For slots with high variance, these can diverge by several percentage points over 100,000 spins. For table games with fixed rules, they converge quickly — which is exactly why the four tables in question are a problem.

The four tables that landed on page one were:

• A European roulette variant with a single-zero wheel but a 94.1% RTP because of a rule modification on even-money bets (la partage removed, en prison added in a way that penalises the player on a double-zero-style outcome).
• A blackjack table paying 6:5 on naturals with a 96.2% RTP under basic strategy.
• A baccarat table with a 95.1% RTP driven by a reduced banker commission.
• A Caribbean stud variant with a 94.8% RTP and a progressive side bet that pulls the blended figure down further.

None of these are bad games in isolation. All four are worse than the alternatives sitting on page two — a 99.5% blackjack table, a 98.6% baccarat table, a 97.3% single-zero roulette table, and a 99.5% video poker variant. The sort put the weaker options first.

Why the sort key produced this order

The most plausible explanation is that the sort ranks by a composite of theoretical RTP and operator-configured "house edge display," where games with visible side bets or progressive components get a blended figure that the sort treats as a single number. A 94.1% roulette table with a side bet that adds 0.4% to the blended figure can outrank a clean 97.3% table if the sort is weighting recent player returns rather than the paytable.

We can't confirm the mechanism without operator disclosure. What we can confirm is the outcome: the sort's page-one ordering did not correlate with the highest available RTP in the lobby.

The 36% Number and What It Suggests About Player Behaviour

The 36% figure comes from session-level tracking across the 1,200 sessions: 432 sessions included at least one real-money bet on one of the four low-RTP tables after the player had interacted with the payout sort. That's not the same as saying the sort caused the bets. Players who sort by payout are already payout-conscious, and they may have chosen those tables for reasons unrelated to the sort — a familiar interface, a lower minimum bet, a bonus that restricted them to certain tables.

But the correlation is uncomfortable. Among players who did not use the sort, the equivalent figure was 19% — 228 of 768 sessions. The gap of 17 percentage points is large enough that it warrants scrutiny, even allowing for selection effects.

The confound we couldn't rule out

The cleanest confound is bonus restriction. Several of the operators in the sample restrict bonus wagering to specific table games, and those tables are often the lower-RTP ones — a 94.1% roulette table is cheaper for the operator to expose to bonus play than a 99.5% blackjack table. A player with an active bonus who sorts by payout may be sorting within a restricted set, not across the full lobby. That would explain both the page-one placement and the betting behaviour without the sort being misleading in the way it first appears.

We flagged this and re-ran the analysis excluding sessions with an active bonus. The gap narrowed to 11 percentage points — still present, still worth explaining, but smaller. That's the honest version of the number.

What the 36% doesn't tell us

It doesn't tell us whether players understood the RTP of the tables they chose. It doesn't tell us whether they would have chosen differently without the sort. And it doesn't tell us whether the four tables were the worst available — only that they were worse than the page-two alternatives in the same lobby. A player who lands on a 94.1% roulette table after sorting by payout has lost roughly 5.9% of turnover in expectation, against 2.7% on the 97.3% table. Over a €500 session at €2 spins, that's a difference of about €16 in expected loss. Not catastrophic. Not nothing.

The Croatian Regulatory Context

Croatia's online gambling market is regulated by the Ministry of Finance under the Law on Games of Chance, with licensed operators required to display game rules and RTP information in a accessible form. In practice, RTP disclosure is uneven. Some operators publish a per-game RTP figure in the game info panel; others publish it only in the general terms. The payout sort sits in a gap: it's a player-facing feature that implies RTP transparency without being bound by the disclosure rules that apply to the games themselves.

That gap matters more in Croatia than in some markets because the domestic licensed pool is small. Players have fewer operators to compare against, and the sort becomes a de facto comparison tool. If the sort is unreliable, the comparison it enables is unreliable too.

What the regulator requires — and doesn't

The licensing conditions require operators to:

• Display the rules of each game.
• Make RTP information available.
• Provide responsible gambling tools, including deposit limits and self-exclusion.

They do not require the lobby sort to reflect the highest-RTP games first, and they do not define what "payout" means when used as a sort label. An operator could legally sort by "payout" using a metric that ranks a 94.1% table above a 99.5% table, provided the games themselves are correctly disclosed. That's not a loophole most players would expect.

The responsible gambling angle

The 36% figure is a behavioural signal, not a harm signal. But it's the kind of signal that responsible gambling frameworks are designed to catch: a player-facing feature that nudges players toward lower-return products without their knowledge. Croatia's responsible gambling requirements focus on deposit limits, time limits, and self-exclusion — all necessary, none of which address a misleading sort. The sort isn't a harm in itself. It's a small, systematic drag on expected value for the players most likely to be paying attention to it.

How to Read a Payout Sort Without Being Misled

The practical takeaway is not to avoid the sort — it's to treat it as a starting point rather than an answer. A few habits close most of the gap.

Check the game info panel, not the lobby label

Every licensed operator in Croatia is required to make RTP available. Open the game, find the info or rules panel, and read the RTP figure there. That number is bound by the disclosure requirement; the lobby sort label is not. If the panel says 94.1% and the sort put the game on page one, the panel wins.

Know the table-game benchmarks

For fixed-rules table games, the RTP is a matter of public arithmetic, not operator discretion:

• Single-zero roulette with la partage: 98.6%.
• Single-zero roulette without la partage: 97.3%.
• Double-zero roulette: 94.7%.
• Blackjack, 3:2 naturals, basic strategy: 99.5%.
• Blackjack, 6:5 naturals: 96.2%.
• Baccarat, standard commission: 98.9% on banker, 98.8% on player.
• Baccarat, reduced commission: 95.1%.

If a table game on page one of a payout sort sits below the benchmark for its ruleset, the sort is not ranking by RTP. That's the whole test.

Watch for blended figures

Progressive side bets and side-bet variants blend a high-variance component into the headline RTP. A Caribbean stud table at 94.8% base RTP with a progressive side bet might display a blended figure that looks competitive while the base game is weak. The base game is what you play most hands on. Read the base RTP separately.

Treat bonus-restricted tables separately

If you have an active bonus, check which tables the wagering requirement allows before you sort. A 94.1% table that's bonus-eligible may be the correct choice under the restriction even though it's the wrong choice in a vacuum. The sort doesn't know about your bonus; you do.

The Open Question

The interesting question isn't whether four low-RTP tables appeared on page one of a payout sort — it's whether the sort is designed to produce that outcome or whether it's an artefact of a blended metric nobody audited. The 36% figure suggests players trust the sort enough to act on it. The 11-point gap that survived the bonus-restriction control suggests the trust isn't fully explained by circumstance. What we don't have is operator disclosure of the sort key, and without that, the most we can say is that the label "payout" is doing work the underlying calculation doesn't support.

Croatia's regulator could close this by defining "payout" for lobby-sort purposes the way it defines RTP for game disclosure — a single, auditable number, ranked honestly. Until then, the sort is a convenience with a known failure mode, and the players most likely to use it are the ones most likely to be caught by it. The next time you sort by payout, check page two before you bet. The 97.3% table is probably there, and it's probably the one you wanted.