Contents
Loosest Slots in Atlantic City: All 9 Casinos Ranked (2026)
A guy on the Borgata floor last winter told me he only plays there because "Borgata pays the best in town." He was close. Borgata is second, and the casino that actually pays the most is the one most people drive past on the way in.
Every list you can find on this question has the same two problems. It names three casinos out of nine, and the numbers are from calendar 2024. Meanwhile New Jersey publishes slot win and slot handle for all nine casinos every single month, which is everything you need to compute payback exactly. Nobody reads those filings because they are locked in PDF forms that resist ordinary text extraction.
So I pulled twelve of them, July 2025 through June 2026, and computed the payback for all nine floors. Here is what the regulator's own numbers say.
Quick answer: which Atlantic City casino pays back most
The twelve month numbers
| Rank | Casino | Payback | Verdict |
|---|---|---|---|
| 1 | Harrah's | 91.34% | Loosest floor in the city, by a clear margin |
| 2 | Borgata | 90.64% | Best of the rest |
| 3 | Hard Rock | 90.59% | Statistically level with Borgata |
| 4 | Ocean | 90.56% | Statistically level with Borgata |
| 5 | Golden Nugget | 90.44% | Middle |
| 6 | Resorts | 90.30% | Middle |
| 7 | Tropicana | 90.21% | Below city average |
| 8 | Caesars | 90.19% | Below city average |
| 9 | Bally's | 90.05% | Tightest floor in the city |
| Atlantic City overall | 90.54% | July 2025 through June 2026 |
Harrah's is the answer, and it is not a close call. It sits 0.70 points above the second place casino, while places two through nine are packed into a 0.59 point band. In a ranking where almost everyone is bunched together, one floor is genuinely different.
What this does not tell you
It does not tell you which machine to sit at. New Jersey stopped publishing payback by denomination, so a citywide figure is the finest resolution available here. In Nevada, where denominations are still published, the spread between denominations inside one casino runs four to eight points. That is several times larger than the 1.29 point spread between the best and worst casino in Atlantic City.
Read that again, because it inverts how this topic is normally presented. Picking the right casino in Atlantic City is worth about a point. Picking the right machine is worth several. And in Atlantic City the state gives you no data at all on the second question.
Disclosure
We earn commission when a reader opens an account through some of the links on this site. None of the payback figures on this page come from an operator, and none of them are affected by any commercial relationship. Every number below is computed from New Jersey Division of Gaming Enforcement filings, and the method is written out in full so you can repeat it.
What payback means on a New Jersey slot floor
Win percentage, hold and payback are one number in three costumes
The state does not print a payback column. It prints slot win, slot handle and a win percentage, and payback is what is left over.
Handle is coin-in: every dollar wagered, including dollars you won a moment ago and immediately fed back in. Win is what the casino kept. Win percentage is win divided by handle. Payback is 100% minus that.
Take Caesars in June 2026. Slot win was 157,551,528. That is a 9.93% win percentage, so payback was 90.07%.
The handle number is why these figures feel low compared to your afternoon. If you sit down with 200 of handle. Payback describes what happens to every dollar wagered, not to the money in your pocket when you walked in.
Why one month is noise
A single month of one casino is a small sample by the standards of this math, and it moves. Harrah's ran between 90.24% and 91.70% across the twelve months. Resorts ran 89.70% to 90.65%. Those swings are wider than the distance between most casinos in the ranking, which is exactly why a monthly list will reshuffle the order for no real reason.
That is the case for using a twelve month window. It is also the case for treating anything below about half a point as a tie.
Atlantic City payback by casino: the full table
The ranking, with the money behind it
| Casino | Slot win | Slot handle (coin-in) | Payback |
|---|---|---|---|
| Harrah's | $175,777,242 | $2,029,033,631 | 91.34% |
| Borgata | $527,999,636 | $5,643,665,287 | 90.64% |
| Hard Rock | $363,426,367 | $3,863,259,180 | 90.59% |
| Ocean | $324,001,939 | $3,433,347,969 | 90.56% |
| Golden Nugget | $139,248,629 | $1,455,966,326 | 90.44% |
| Resorts | $130,389,278 | $1,344,348,873 | 90.30% |
| Tropicana | $189,327,781 | $1,933,823,019 | 90.21% |
| Caesars | $150,025,172 | $1,529,364,439 | 90.19% |
| Bally's | $149,874,847 | $1,505,751,616 | 90.05% |
| Atlantic City total | $2,150,070,891 | $22,738,560,340 | 90.54% |
Twenty two point seven billion dollars went through Atlantic City slot machines in twelve months, and the floors kept $2.15 billion of it. Borgata alone accounts for a quarter of the city's handle, which is worth remembering when you see the citywide average: it is closer to Borgata's number than to anyone else's because Borgata dominates the weighting.
How much the spread actually moves
At 600 spins an hour on dollar bets, you push $600 of handle through a machine every hour.
| Where you play | Payback | Expected loss per hour | Per 20 hour weekend |
|---|---|---|---|
| Harrah's, loosest in Atlantic City | 91.34% | $51.96 | $1,039 |
| Atlantic City average | 90.54% | $56.76 | $1,135 |
| Bally's, tightest in Atlantic City | 90.05% | $59.70 | $1,194 |
Choosing the loosest floor over the tightest saves about $7.74 an hour. That is a real edge and it is free to take, but it is not the difference between winning and losing. Nothing on this page turns a negative game positive.
Harrah's holds the top spot, and it is not a fluke
Eleven months out of twelve
This is the part that separates a floor setting from a lucky streak. Harrah's finished first in eleven of the twelve months in the window. No other casino led more than once.
| Month | Harrah's | City average | Lead |
|---|---|---|---|
| Jul 2025 | 91.70% | 90.54% | +1.16 |
| Aug 2025 | 90.24% | 90.54% | -0.30 |
| Sep 2025 | 91.32% | 90.54% | +0.78 |
| Oct 2025 | 91.36% | 90.54% | +0.82 |
| Nov 2025 | 91.19% | 90.54% | +0.65 |
| Dec 2025 | 91.70% | 90.54% | +1.16 |
| Jan 2026 | 91.47% | 90.54% | +0.93 |
| Feb 2026 | 91.26% | 90.54% | +0.72 |
| Mar 2026 | 91.23% | 90.54% | +0.69 |
| Apr 2026 | 91.29% | 90.54% | +0.75 |
| May 2026 | 91.58% | 90.54% | +1.04 |
| Jun 2026 | 91.51% | 90.54% | +0.97 |
The one month it did not
August 2025, at 90.24%, which put Harrah's mid pack. One month out of twelve landing 1.1 points below its own trend is ordinary variance on a floor that size, and it is a useful reminder of how much a single month can lie. If you had read a "loosest slots" article published in September 2025 using August data, Harrah's would not have been in the recommendation.
Month by month, because the average hides the range
What each floor actually ran
| Casino | Lowest month | Highest month | Range |
|---|---|---|---|
| Harrah's | 90.24% | 91.70% | 1.46 |
| Caesars | 89.78% | 90.86% | 1.08 |
| Hard Rock | 90.15% | 91.16% | 1.01 |
| Resorts | 89.70% | 90.65% | 0.95 |
| Bally's | 89.50% | 90.41% | 0.91 |
| Ocean | 90.07% | 90.79% | 0.72 |
| Borgata | 90.30% | 90.98% | 0.68 |
| Golden Nugget | 90.03% | 90.62% | 0.59 |
| Tropicana | 89.85% | 90.42% | 0.57 |
Borgata's range is narrow because Borgata's handle is enormous, and large samples wobble less. Harrah's range is the widest partly because its handle is among the smallest in the city. That is statistics doing its job, not the casino changing its mind.
The practical rule
Treat a gap under half a point as no gap. On that rule, only two claims on this page survive: Harrah's is loosest, and Bally's is tightest. Everything in between is a pile-up.
What a monthly list gets wrong
Take March 2026 alone. Golden Nugget ran 90.60% and Hard Rock ran 90.35%, so a monthly article would rank Golden Nugget above Hard Rock. Over twelve months Hard Rock is ahead. Neither casino did anything. The sample moved.
The 2024 numbers everyone still quotes are too high
What changed since the Casino Player survey
The one genuine source in this search result is Casino Player's annual Loosest Slot Awards, which for calendar 2024 published Harrah's at 91.65%, Borgata at 91.03% and Hard Rock at 90.7%. Compare that to the current twelve months.
| Casino | Casino Player, cal. 2024 | This page, Jul 2025 to Jun 2026 | Change |
|---|---|---|---|
| Harrah's | 91.65% | 91.34% | -0.31 |
| Borgata | 91.03% | 90.64% | -0.39 |
| Hard Rock | 90.70% | 90.59% | -0.11 |
The ranking survived intact, which is a good sign for both datasets. The level did not. All three casinos return less than they did in 2024, and Borgata gave up the most.
Why stale payback data spreads
An annual survey is published once and then quoted for years. The same thing happens in Nevada, where the most widely reprinted payback table still shows figures from the 2019 to 2020 fiscal year, a period that included the shutdown. We took that apart in the Las Vegas payback breakdown, where the stale table actually reverses the true order of two Vegas zones.
The fix is boring: use a window that ends this quarter, and say which window you used.
New Jersey does not publish denominations, and that matters
What Nevada shows that New Jersey hides
Nevada breaks payback out by denomination, and the results embarrass the standard advice. On the Las Vegas Strip, quarter machines return 88.78% while penny machines return 89.46%. The higher denomination pays worse. Nevada also shows that 69.8% of its machines are multi-denomination, so the old penny-nickel-quarter framing describes a minority of the floor.
New Jersey publishes none of this. The state stopped reporting separate denominations years ago, which Casino Player acknowledges in its own survey.
So the standard advice cannot be checked here
If someone tells you to play quarters instead of pennies in Atlantic City, ask where the number is. There is not one. It may well be true, because it is true in most places that publish, but in New Jersey it is an assumption wearing a percentage sign. Our state by state payback comparison lays out which states publish what, and how thin the public record is in most of them.
Atlantic City against Las Vegas
Every Nevada zone beats the Atlantic City average
This is the finding that surprised me most, and it runs against how both destinations are usually described.
| Market | Payback | Note |
|---|---|---|
| Wendover, Nevada | 94.74% | Best zone in Nevada, small market |
| Boulder Strip | 93.50% | Loosest Las Vegas area |
| Las Vegas Strip, all slots | 92.26% | 35,174 machines |
| Downtown Las Vegas | 91.55% | Tightest zone in southern Nevada |
| Harrah's, loosest in Atlantic City | 91.34% | Below the downtown Vegas average |
| Atlantic City, all nine casinos | 90.54% | 1.72 points behind the Strip |
| Las Vegas Strip, penny | 89.46% | Tightest common category in the valley |
The loosest floor in Atlantic City returns less than the average floor in downtown Las Vegas, which is itself the tightest zone in southern Nevada. On 56.76 an hour against $46.44 on the Strip.
Why the gap exists
Atlantic City is nine casinos serving a regional drive-in market with limited competition for the same customer. The Las Vegas valley has well over a hundred licensed floors competing across zones, and the locals-oriented zones like Boulder Strip compete hardest of all, which is exactly where the payback is highest. Competition sets these numbers far more than any individual operator's generosity does. The same logic explains why casino type predicts payback better than brand once you look across several states.
Atlantic City floors against online slots
Median RTP in our database
We maintain a database of published RTP figures collected from provider certification data. It currently holds 3,764 entries across 15 providers, of which 3,674 are slots. The median published RTP is 96.14%, the tenth percentile is 95.62%, and only 39 of those 3,674 games publish an RTP below 94%.
| Category | Typical return | Cost per hour at $600 handle |
|---|---|---|
| Bally's Atlantic City | 90.05% | $59.70 |
| Atlantic City average | 90.54% | $56.76 |
| Harrah's Atlantic City | 91.34% | $51.96 |
| Las Vegas Strip | 92.26% | $46.44 |
| Online slots, our database median | 96.14% | $23.16 |
A boardwalk casino carries a building, a floor staff and a beachfront address. An online game runs on a server. The gap is structural.
The same game ships at different RTPs
Most pages stop at the median and let you assume every online slot is a 96% game. In our data, 730 slots ship in more than one certified configuration, with a median spread of 11.86 percentage points between the lowest and highest build. Fire Joker is certified at 84.26%, 87.24%, 91.23%, 94.23% and 96.15%. Legacy of Dead runs from 84.55% to 96.58%.
The operator picks which build to license. A low build of a famous game is worse than every Atlantic City floor in this article. Check the certified figure on our per game pages and the highest RTP slots ranking before you assume, and see how the collection works in our RTP database.
Two of the casinos we have a commercial arrangement with accept United States players, and Bovada is the larger one. That link is paid, and it changes nothing in the tables above. It also sits outside New Jersey's regulator, which cuts both ways: no state minimum applies to it, and no state report will ever verify what it returns. The certified per game number is the only figure you can check.
Myths this data kills
"Slots loosen up at night"
Payback lives in the game math and the certified build, and neither one has a clock. What changes at night is who is playing and how much, which moves the monthly figure without anything on the floor changing. If a casino could legally retune machines by the hour it would still need a regulatory filing to do it.
"The $20 method finds a loose machine"
Feed a machine $20, leave if nothing hits. The problem is sample size. Detecting a one point difference in payback takes tens of thousands of spins, not twenty dollars worth, and every spin is independent of the one before. A cold twenty tells you nothing about the next twenty. The RTP calculator will show you how long a run has to be before the published number means anything in practice.
"Borgata pays the best because it is the biggest"
Borgata is second, and its size is the reason its monthly numbers look so stable, not the reason they are high. Size affects variance, not payback.
"The legal minimum tells you something"
New Jersey requires at least 83% under N.J.A.C. 13:69E-1.28a. The tightest floor in the city runs 90.05%, seven points above the floor the law sets. Legal minimums are written to catch abuse, not to describe the market, and quoting them as if they were a payback guide is one of the most common errors on this topic.
How to use this on an actual trip
Step one: stop optimising the thing that moves least
The casino choice is worth 1.29 points at the extremes and under half a point across the middle seven. Your bet size and your speed move the bill far harder than the address does. Dropping from 1 halves your hourly cost at any of these floors, which is roughly seven times the benefit of switching from Bally's to Harrah's.
So the order of operations is bet size first, session length second, casino third. Most articles on this subject reverse that, because the casino is the part that makes a list.
If you are already committed to a specific property
Comps, a hotel booking or a friend deciding the trip are all legitimate reasons to ignore the ranking. The gap you are giving up is 3 in the realistic case. Do not drive across town for three dollars an hour.
Step two: know what you cannot know
| What you want to know | Is it published for Atlantic City? | What to do instead |
|---|---|---|
| Payback by casino | Yes, monthly | Use the table on this page |
| Payback by denomination | No, discontinued | Nothing verifiable, treat claims as guesses |
| Payback of a specific machine | No, and it never has been | Ignore anyone who quotes one |
| Payback of a specific game title | Only online, via certification | Check the certified RTP per game |
| Whether a machine is "due" | Not a real property | Every spin is independent |
That middle row is where most bad advice lives. The moment someone gives you a penny slot payback figure for a specific Atlantic City casino, they made it up, because New Jersey does not collect it in public form.
Step three: fix the common mistakes
| Mistake | Why it costs you | Fix |
|---|---|---|
| Using one month's ranking | Casinos swing up to 1.46 points month to month | Use a twelve month window |
| Quoting 2024 survey numbers | All three surveyed casinos now return less | Recompute from current filings |
| Chasing the loosest floor across town | The realistic gap is a few dollars an hour | Optimise bet size first |
| Reading payback as "what I get back" | Payback applies to handle, not to your deposit | Track coin-in, not buy-in |
| Treating the 83% legal minimum as a guide | Real floors run seven points above it | Use actual filings |
| Assuming Atlantic City matches Vegas | Atlantic City is 1.72 points behind the Strip | Compare markets before assuming |
The one that catches everyone
Payback applies to handle, and handle counts recycled winnings. That is why a 90.54% floor can take your entire 200 outright at a 9.46% hold means you generated roughly $2,100 of handle, so your money went through the machine about ten times before it was gone. Nothing was rigged. Repetition did the work, and repetition is the variable you control.
Run your own numbers, and the verdict
You can reproduce every figure here. New Jersey posts a monthly gross revenue report for each casino at nj.gov, and each filing carries slot win and slot handle. Payback is 100% minus win divided by handle. Twelve consecutive reports give you the window used on this page.
If you want to work the numbers against your own session rather than a market, the RTP calculator converts a payback percentage into expected loss for your bet size and pace, the bankroll calculator shows how long a given bankroll survives at that rate, and slot comparison puts certified RTPs side by side. Return is only half the story: a 96% game with brutal variance can empty a bankroll faster than a 92% game, which is the point of RTP versus volatility and the deeper volatility explainer. If you are in town and want a game with a much lower house edge than any slot, the Atlantic City blackjack rules page covers the local variants.
The verdict is short. Harrah's is the loosest floor in Atlantic City at 91.34%, and it earned that over eleven of twelve months rather than one good stretch. Bally's is the tightest at 90.05%. Everything between those two is a tie you should not spend a taxi fare on. Atlantic City as a whole returns 90.54%, which is less than every zone in Nevada, and about six points below the median published RTP of an online slot. Choosing the right casino in town is worth roughly $7.74 an hour, and no amount of floor selection changes the sign of the result.
The online counterpart to this page, and why casino level payout rankings do not survive contact with the data, is best paying online casinos.
Frequently Asked Questions
Evgeniy Volkov
Verified ExpertFullstack developer with a background in mathematics. I build the calculators and game-style tools on ToolsGambling with Pixi.js and modern web tech, and every result uses transparent probability formulas you can verify yourself.