The independent South Africa guideof all returned credits came from the five largest payouts in this sample.
Game Lab / Evidence before claims
Wealth Inn: 699 demo rounds tested.
How often did a payout cover its stake? How different were the sessions? Explore the completed-round records alongside our payline and paytable explanations. This sample does not measure theoretical RTP.
Recorded provider demo · 2026-10-13
91 rounds paid something; 75 returned more than the stake. A payout and a profitable round are different results.
93,7%Observed sample returnNot theoretical RTP
75Rounds returning more than stakeOut of 699 completed rounds
8Sessions with completed roundsReset balances kept separate
Fixed stake: 4,00 demo credits. A small observed sample, not a forecast. Credits have no cash value. Partial sample: the planned 1 000-round run was not completed.
Profitable historical patterns / 108,035 nearby settings
Which profitable historical fit was less fragile?
Our local search found a maximum of +2 577,37 hypothetical credits, but its bank fell to 48,77. The lower-drawdown 1/18-credit rule made +1 118,42, with drawdown 293,25 and a minimum bank of 814,39. All eight tested neighbouring settings also completed in profit.
Compare the settings, drawdowns and neighbour checks →
Green: net profit. Red: the largest drop from a previous bank peak. All variants paid for 699 recorded spins with one 1,000-credit bank and no top-ups. Settings were selected using these same outcomes; new data have not been tested. Hypothetical credits, not rand or observed winnings.
Recorded outcomes, not a prediction
Long gaps. Uneven payouts.
Read the recorded sequence rather than assuming that a loss makes the next win due. Each mark below is one completed round. The gold spikes show returned credits divided by that round’s stake.
Payout by recorded roundReturned ÷ stake
How each session moved.
3 of 8 sessions ended above their starting point in the recorded completed rounds. Each graph starts at zero and shows cumulative payouts minus stakes within that session only. The longest session is expanded first. Session sizes differ, and the vertical scales differ; compare the labelled credit amounts, not the shapes alone.
session-0110 rounds+1 credits
Net change from this session’s startDemo credits
session-02210 rounds-319,5 credits
Net change from this session’s startDemo credits
session-032 rounds-0,5 credits
Net change from this session’s startDemo credits
session-0425 rounds-38,5 credits
Net change from this session’s startDemo credits
session-0525 rounds-88 credits
Net change from this session’s startDemo credits
session-06117 rounds-165,5 credits
Net change from this session’s startDemo credits
session-0732 rounds+189,5 credits
Net change from this session’s startDemo credits
session-08278 rounds+245,5 credits
Net change from this session’s startDemo credits
Five payouts contributed 27,9% of all returns
Top five: 730 creditsOther 694 rounds: 1 890 credits
| Round | Returned credits | Return ÷ stake |
|---|---|---|
| 562 | 232 | 58× |
| 416 | 142 | 35,5× |
| 687 | 132 | 33× |
| 218 | 112 | 28× |
| 411 | 112 | 28× |
A payout is not always a profitable round.
Every completed round falls into exactly one category.
consecutive rounds without a payout in the longest observed within-session streak.
Inspect a recorded round.
These rows come from the downloadable dataset.
Choose a round to inspect its stake, return and session.
Different sessions, different results.
| Session | Rounds | Net credits | Sample return |
|---|---|---|---|
| session-01 | 10 | 1,00 | 102,5% |
| session-02 | 210 | -319,50 | 62,0% |
| session-03 | 2 | -0,50 | 93,8% |
| session-04 | 25 | -38,50 | 61,5% |
| session-05 | 25 | -88,00 | 12,0% |
| session-06 | 117 | -165,50 | 64,6% |
| session-07 | 32 | 189,50 | 248,0% |
| session-08 | 278 | 245,50 | 122,1% |
Session sizes can differ. These are separate demo balances, not a continuous bankroll or equal-size experiments.
A recorded example · Round 441
One Wild. Two winning lines.
The Wild expanded over the right-hand reel. Two ingot combinations paid 9.00 credits each: the bottom row and a bottom, top, bottom route. Together they returned 18.00 credits on a 4.00-credit stake.
Returned9 + 9 = 18 credits
Net result after the stake18 − 4 = +14 credits
The routes above were checked against the official payline diagrams and the recorded winning coordinates. They are explanatory diagrams, not additional gameplay screenshots. This one example shows an expansion; it does not measure how often the Wild appears.
Find round 441 in the public CSV, or use the round explorer above. Read all eight line routes and the Wild expansion rule.
01
What the source comparison establishes.
The provider specifies several RTP builds; operator listings establish where the name appears; a screenshot illustrates the interface. These answer different questions. None of them proves a winning strategy or a player’s expected next result.
02
Read the Wild rule we captured.
- The condition matters
The small Wild must participate in a winning combination. Its appearance alone is not the rule for expansion.
- Expansion comes before payout
The expanded Wild covers the reel. This is the rule text, not a screenshot of a winning round we recorded.
03
Returned money is not profit.
Change the hypothetical stake and multiplier. Returned = stake × multiplier. Net = returned − stake. The calculation is original site functionality; the inputs are illustrative, not recorded spins.
04
A paytable number needs a denominator.
The help PDF expresses a line win in coins, not as a multiple of the total stake. With eight lines, those are different amounts. Our current demo paytable confirms the example below: three peaches return 29 credits at a 0.50 coin value.
Illustrative calculation using the 2020 help document. Not a captured win or current operator limit. The paytable in your selected version takes priority.
05
A compact, downloadable source register.
Download the source comparison as CSV or JSON. It records the source, claim, evidence category and limitation. It is a source-comparison dataset, not a gameplay outcome dataset.
06
How we recorded and checked the sample.
We recorded 699 completed official-demo outcomes across 8 sessions on 13 October 2026, at a fixed 4.00 demo credits per round. The planned 1,000-round run was not completed. Credits have no cash value and are not South African rand.
What we checked
For every included round, we checked its stake, total payout and terminal game state against the saved response. The closing balance must equal the opening balance minus the stake plus the payout. Initialization records and subsequent balances were checked separately for each session. We do not join reset balances into one bankroll.
How the outcomes were recorded
530 included outcomes came through the ordinary game interface. Another 169 were recorded through sequential requests in legitimate public demo sessions. The combined figures are descriptive observations, not a representative random sample or a comparison of collection methods.
What is missing from the sample
Two earlier request errors and four later connection failures produced no usable outcome. We cannot establish whether those requests were charged or settled; they are not counted as losses, wins or completed rounds. Three session-start attempts contained no completed rounds. Rules-only visits are separate.
The accepted sample ends at round 699. A further 27 recorded outcomes after that analysis cutoff are excluded. The interruptions and exclusions mean this is not an uninterrupted record of every attempted wager. Session sizes differ, so their returns should not be compared as equal experiments.
What these results cannot establish
Observed sample return is not theoretical RTP, a forecast or evidence for a winning strategy. We have not classified symbol features across the full sample, so the dataset does not establish Wild frequency. The illustrated round 441 is a recorded example, not a typical result. Provider rules, illustrative calculations and observed outcomes remain separate.
Recalculate the published numbers
Download the CSV or JSON, and our standalone analysis script. It needs Node.js 18 or later, no packages or network connection.
node reproduce-study.mjs rounds-699.csvThe script recalculates payouts, profit, session results and losing runs from the public rows. It does not reproduce the random outcomes or authenticate the original server responses. Compare downloaded files against the SHA-256 manifest. Session access credentials and complete raw network responses remain private.
13 Oct 2026
Provider specifications and public operator listings. No funded gameplay test claimed.
Provider specification ↗