How We Test: Devices, Deposits and Payout Timings

The test hardware

Two handsets, chosen to bracket the phones Indian players actually own rather than to flatter an app. A flagship would hide every performance problem this category has.

Both handsets are wiped and re-imaged before a full test cycle so that cached data from a previous build cannot flatter a cold-start figure.

The six-step process

The sequence runs the same way for every application we cover, Crorebet included, which is what makes the comparison between them meaningful.

1. Install on two phones

   The package goes on both handsets in the same session. We log the download size, the install duration and the cold-start time separately for each, because a 43 MB file behaves very differently on 2 GB of RAM.

2. Fund the account with our own money

   Real rupees, through at least two UPI routes, timed from confirmation to credited balance. We do not accept operator-provided balances, because a comped account can be flagged for special handling.

3. Compare 40 game categories across six apps

   The same categories, sampled within a few minutes of each other on six applications, then compared by published RTP and measured house edge. This is where the 4.8% live-table figure comes from, against a 6.9% six-app average.

4. Run graded withdrawal tests

   Three payouts at different sizes and different hours: ₹500 landed in 22 minutes on 14 August 2026, ₹2,500 in 41 minutes on 16 August, ₹7,000 in 2 hours 10 minutes on 18 August. Each is timed from request to bank credit.

5. Contact support cold

   We open live chat as an ordinary user with a real problem, in Hindi and in English, and record the first-response time — 3 minutes 40 seconds on 15 August 2026 — plus whether the answer resolved anything.

6. Re-verify quarterly

   Every table on this site carries a last-verified date, and the measurement behind it is repeated at least quarterly, or immediately when a new build ships.

Scoring weights, and what moves each one

Weights are fixed before any app is scored and are identical across the products we compare. The columns below are the part that decides a number.

Scores are multiplied by these weights and summed. The published Crorebet casino review carries the resulting number and the per-criterion breakdown.

How is a house edge calculated?

By reading the published RTP on each game category, converting it to the operator's retained share, and comparing that figure across apps on the same titles. That retained share is what the platform keeps, and it is the single most useful number for comparing two casino apps in the same category.

The conversion: house edge = 100% − RTP. Take a crash game with a stated 97% RTP.

• 100% − 97% = 3% house edge on that title.
• On ₹10,000 of turnover, the expected cost is roughly ₹300 before variance.
• A rival title at 94% RTP carries a 6% edge — ₹600 on the same turnover.

We repeat this across 40 categories so that one generous title cannot carry an average, and we sample all six apps within the same short window because campaign RTP panels can change. The category detail behind the figures sits on the casino games guide.

What this method cannot prove

Stating the ceiling is part of the method. Four limits apply to everything we publish, and a reader who ignores them will over-read our scores.

• Prices are single snapshots. A 40-market sample taken over one August window is a photograph, not a film. Margins widen close to an event and on less liquid markets, and our figure will not capture that.
• The payout sample is small. Three withdrawals in one week, from one verified account, in one city. They tell you the rail works and roughly how fast; they cannot tell you what happens to a ₹2,00,000 payout or to an account flagged for review.
• We have no server-side visibility. We see what the client shows us. Bet placement logic, risk management, account limiting and how randomness is generated in casino titles are all invisible from a handset.
• We hold no audit access. We cannot inspect an operator's books, segregation of player funds, or the testing certificates behind any game. Where we describe a licence, we are describing a public register entry, not an audit we performed.

What would move a score?

Scores are not decorative, so they change when evidence changes. The Crorebet numbers published today are provisional in exactly that sense. A payout that misses its stated window in two consecutive test cycles cuts the payments figure. A house edge that drifts above the six-app average cuts the casino lobby figure hardest, because it carries 30% of the weight. Crashes on the 2 GB handset that survive an update cut stability.

In the other direction, a payout rail that adds UPI, a licence in a jurisdiction with player recourse, or wagering terms shown before deposit rather than after would each raise a component. Reader reports feed into this: if several people document the same failure and we reproduce it, the score moves at the next quarterly pass. How to send one is set out in our corrections process, and the reasoning behind our independence is on the about page.