Andar Bahar & Dragon vs Tiger on GoPlay: Rules, Real Odds & Bets to Avoid

Andar Bahar and Dragon vs Tiger look like coin flips with cards: pick a side, watch a reveal, get paid or not, all in well under a minute. Because the rules take two minutes to learn, most players assume the maths is just as simple. It is not, and the speed is where the real cost hides. This guide works out the actual odds behind both games, what each bet costs per hour, and which bets are worth skipping.

Why the Fastest Card Games Deserve the Most Caution

The house edge applies to every rupee you stake, not to the rupees you deposit. A game that resolves a round every 20 to 30 seconds lets you stake the same money over and over, so the edge is applied dozens of times in a few minutes. That is why a small percentage can add up to a large hourly cost.

Both games sit at the quick end of GoPlay's games library. The site's existing figures put Andar Bahar at roughly 30 seconds a round, and Dragon vs Tiger is described as rounds "in seconds". Compare that with a Rummy hand that takes several minutes, and you can see why the same deposit disappears at very different speeds.

RTP (return to player) is the published long-run figure: of every ₹100 staked over a very large number of rounds, about ₹97.10 comes back in Andar Bahar and about ₹96.30 in Dragon vs Tiger. The house edge is 100% minus that number, so 2.9% and 3.7%. It is a long-run average, not a promise about your next hour, and a short session can land far above or below it.

Why near-even bets feel safer than they are

When a bet wins almost half the time, losses arrive as small, frequent jolts and wins arrive just as often. That rhythm feels like control, even though every round is independent. The edge does not announce itself in one big loss; it shows up as a slow drift over hundreds of rounds.

• Small stakes hide big turnover. ₹20 a round at 120 rounds an hour is ₹2,400 staked in an hour, from money you may feel you have barely touched.
• Wins get re-staked. Winnings that go straight back into the next round count as turnover again, so a good start does not slow the edge down.
• There is no natural stopping point. A slot has a bonus round and Rummy has a declaration; these games simply deal the next card.

Andar Bahar: The Rules, Step by Step

Andar Bahar is a traditional Indian card game. A single card is turned face up, and you bet on which of two piles the next card of the same rank will land on. The exact table rules, payouts and betting limits vary, so treat the description below as the common structure and confirm the details on the game's rules screen in the app.

Terms you will see on the table

Typical bet types

The core bet is the simple one: Andar or Bahar. Some tables add side bets on top. I cannot confirm which side bets GoPlay offers or how they pay, so check the rules screen for the list and the payout table before you use any of them.

You can judge any bet yourself with one formula. If a bet wins with probability p and pays b-to-1 (you receive b times your stake plus the stake back), the house edge is 1 − p × (b + 1). A bet that pays a lot for a rare result almost always has a bigger edge than the main bets, because the payout is usually set below the true odds.

The Maths Behind Andar Bahar

The short answer: the two sides are not equal. The side dealt first wins about 51.5% of rounds and the other side about 48.5%. The gap is small, but it is the reason many tables pay the two sides differently.

Where 51.5% comes from

Once the joker card is turned up, 51 cards remain, and exactly 3 of them share its rank. Those three matching cards can sit in any of the 51 positions, and every arrangement is equally likely. The number of ways to choose 3 positions from 51 is 51 × 50 × 49 ÷ 6 = 20,825.

The round ends at the first matching card, so ask: what is the chance that the first match is at position 1, position 2, and so on? If the first match is at position k, the other two matches must be somewhere in the 51 − k positions after it. That can happen in (51 − k) × (50 − k) ÷ 2 ways. Divide by 20,825 for the probability.

Earlier positions are always a little more likely than later ones, because there is more deck left behind them to hold the other two matches. The first-dealt side gets the odd positions (1, 3, 5 and so on up to 51), and the other side gets the even positions. Add up the odd-position ways and you get 10,725. Divide by 20,825 and the first-dealt side wins 51.50% of rounds; the other side wins the remaining 10,100 arrangements, or 48.50%.

How a payout difference produces an RTP near 97%

If both sides paid an even 1:1, the side dealt first would have a positive expected value, which no operator would allow. So the favoured side typically pays a bit less. Here is an illustration only, not GoPlay's confirmed payout table: suppose the favoured side pays 0.9:1 and the other side pays 1:1.

The two illustrative results, 97.85% and 97.0%, bracket the listed 97.1%, which shows how a payout structure of this kind produces an RTP in the high nineties. The actual payouts and dealing order in the app may differ, so use your own numbers with the formula above. The point that holds either way: neither side beats the house, and the game's own designers have priced the small first-dealt advantage into the payout.

Dragon vs Tiger: The Rules, Step by Step

Dragon vs Tiger is the simplest card game on the app: one card for Dragon, one card for Tiger, and the higher card wins. The site describes it as one card each, highest wins, with side bets available and rounds in seconds. As with Andar Bahar, confirm the payouts and tie rules on the game's own rules screen.

Card ranking

How ties are handled

Ties matter far more in Dragon vs Tiger than most players realise, because they happen about once every 13 or 14 rounds. Many versions of the game return half of a Dragon or Tiger stake when the round ties, but this is not universal, and I cannot confirm GoPlay's rule. The next section shows why the listed 96.3% RTP points to the half-back rule.

The Maths Behind Dragon vs Tiger

The short answer: Dragon and Tiger are perfectly symmetrical, about 46.27% each, and the tie takes the remaining 7.47%. The tie is what creates the house edge on the main bets and makes the Tie bet a poor deal.

The probability of a tie

The calculations here assume an 8-deck shoe (416 cards, 32 of each rank), which is a common casino setup; check the rules in case the game uses a different number of decks. Give Dragon any card. Tiger's card is one of the 415 cards left, and 31 of them share Dragon's rank. So P(tie) = 31 ÷ 415 = 7.47%.

The other 92.53% of rounds have a different rank for each side, and by symmetry Dragon and Tiger are equally likely to hold the higher one. So each wins 92.53% ÷ 2, or about 46.27%. With an infinitely large deck the tie chance would be exactly 1 in 13, or 7.69%, so the 8-deck figure is a little lower.

Why the listed 96.3% is consistent with half back on a tie

Say you bet ₹100 on Dragon at even money. You win ₹100 in 46.27% of rounds and lose ₹100 in 46.27%, so those two cancel out. That leaves the ties, 7.47% of rounds. If a tie returns half your stake, you lose ₹50 in those rounds, so the expected loss per ₹100 is 0.0747 × ₹50 = ₹3.73 and the RTP is 100 − 3.73 = 96.27%, or about 96.3%.

That matches the 96.3% listed for the game. It is an inference, not a confirmed rule, but it is a neat fit. If a tie lost the whole stake instead, the RTP would be about 92.5%, far below the published number.

What the Tie bet really costs

The Tie bet wins only 7.47% of the time. At 11:1 (you win ₹11 for every ₹1 staked) the expected value per ₹1 is 11 × 0.0747 − 0.9253 = −0.1036, a house edge of about 10.4%. At 8:1 the same working gives 8 × 0.0747 − 0.9253 = −0.3277, a house edge of about 32.8%.

To make the bet fair you would need the payout to be roughly 12.4:1, because (1 − 0.0747) ÷ 0.0747 = 12.39. No table pays that, so the bet always favours the house. With an infinite deck the 11:1 edge would be 1 ÷ 13 = 7.69%, so the 8-deck maths is a little worse for the Tie bet.

What each bet really costs

The Tie bet's edge is roughly three times the main bets at 11:1, and almost nine times at 8:1. It is the bet to avoid. It also has the most violent variance: you lose about 92.5% of rounds and occasionally win 11 times your stake, so a short session can look brilliant or brutal without telling you anything about the edge.

What an Hour of Play Actually Costs

Multiply your stake by the rounds you play and by the house edge, and you get the expected loss. The formula is expected loss = stake × rounds × house edge. It is an average over the long run, not a prediction for your session, but it is the honest price of the entertainment.

The speed assumptions

Andar Bahar runs at about 30 seconds a round, which is 120 rounds an hour if you bet every round. The site says Dragon vs Tiger rounds take "seconds" but gives no figure, so as an assumption I use about 20 seconds, or 180 rounds an hour. If the real timer is longer, the Dragon vs Tiger costs scale down proportionally.

A quick rule of thumb from the table: every ₹1 of stake costs about ₹3.48 an hour in Andar Bahar, ₹6.66 in Dragon vs Tiger on the main bets, and ₹18.72 on the Tie bet, if you bet every round at these speeds. Because the minimum deposit is ₹100, a first deposit funds only five ₹20 rounds, which is about two and a half minutes of Andar Bahar.

The average hides the swing

These are averages, and the actual result of any single hour moves around them by much more than the average itself. A single ₹50 bet has a standard deviation of about ₹47 in Andar Bahar. Over 120 rounds the swing in one hour is roughly ₹47.5 × √120, about ₹520, against an expected loss of ₹174. In Dragon vs Tiger over 180 rounds it is roughly ₹650 against an expected loss of ₹333.

So an hour can genuinely end ahead, and many do. But that happens because the swing is large, not because the edge is small, and the edge is still working in the background of every session.

Streaks, Roads and the Gambler's Fallacy

The short answer: every round is independent, so no history can tell you what comes next. Streaks look meaningful but they are exactly what chance produces. The mistake of thinking a result is "due" after a long run is called the gambler's fallacy.

How long a streak should you expect?

Treat each round as a fair 50/50 flip, ignoring ties, which is close enough for these games. A run is a set of identical results in a row. The table shows the chance of seeing at least one run of that length, of either result, in a given number of rounds. I calculated it by stepping through the rounds one at a time and tracking the current run length.

In 100 rounds, a run of six or more of the same result is far more likely than not (about 81%), and the longest run in 100 fair flips is typically around six or seven. If you watch a table for a while and see seven Dragons in a row, nothing has broken. Random data looks streaky to people who expect it to alternate.

What "roads" and history boards claim

Some card tables show a history strip or a "road": a grid of coloured dots that records recent winners in columns. Players read these as trends ("Tiger is running") or as reversals ("Dragon must be due"). If your game screen shows a history display, treat it as decoration.

• Independence. In Dragon vs Tiger the cards come from a large shoe whose make-up barely changes from round to round, and it may be reshuffled often. Last round's result tells you almost nothing about the next.
• Andar Bahar resets every round. The positions of the three matching cards are random each time, so previous rounds tell you nothing about where they will land.
• Trend-following and reversal-betting are both wrong. "It's on a streak, ride it" and "it's overdue, bet the other way" each assign a memory to something that has none.

Could card counting help in Dragon vs Tiger?

In principle a shoe that has lost many cards of one rank slightly changes the tie chance. In practice, the Tie bet at 11:1 breaks even only if P(tie) reaches 1 ÷ 12, or about 8.33%, compared with 7.47% for a full shoe. You would need to know the exact composition of the remaining cards, the app does not show it, and the table may reshuffle every round. Check the rules screen, but do not build a plan on it.

Systems That Do Not Work: Martingale, Predictors and Signals

The short answer: no staking system, prediction tool or signal group changes the expected value of a bet with a house edge. A system can change how the losses arrive; it cannot remove them. This section covers the four ideas that come up most often.

The Martingale (doubling after every loss)

The Martingale says: double your stake after each loss, and the first win recovers everything plus one unit of profit. On an even-money bet that is true. The catch is what the ladder demands, shown here with a ₹20 starting stake at 1:1 (the actual payouts in the app may be slightly different, which makes recovery even less complete).

Every win pays the same ₹20 however deep the ladder goes, while the money at risk grows exponentially. With ₹1,260 in your session budget, you can fund six bets, and six straight losses wipe it out. Losing six in a row at odds close to 50/50 has a chance of about 1.6% per attempt, which sounds remote until you play many rounds.

Stepping through 120 rounds (one hour of Andar Bahar), I calculate about a 61% chance of at least one run of six straight losses on a 50/50 bet, and about 67% if you are backing the side that wins only 48.5% of the time. Over three hours (about 360 rounds) the chance is roughly 95% to 97%. The Martingale is not a way to win; it is a way to win small amounts often and lose a large amount eventually.

Table limits make it worse. If the app's maximum bet is below the next rung of the ladder, the system stops working even if you have the money. Check the minimum and maximum bets on the rules screen. And because each bet in the ladder has a negative expected value, a bigger total turnover means a bigger expected loss, not a smaller one.

Predictor apps and Telegram or WhatsApp signals

Some groups sell "predictions" for Andar Bahar and Dragon vs Tiger, often with screenshots of winning runs. The maths explains why they cannot work. The outcome is decided by a shuffled deck or a random number generator after bets close, and no outside tool can see it in advance.

• Survivorship bias. A group that posts hundreds of predictions will have a few good streaks to screenshot, and quietly deletes the rest.
• The business model is you. Subscription fees, "VIP" upgrades and referral commissions pay the group, not a winning method.
• Malware risk. "Predictor" APKs and "hack" files are a classic way to steal UPI details and OTPs. Never install one, and never share an OTP or UPI PIN. The APK install guide explains what a safe install looks like.

"The dealer rigs it after you win"

Players often feel that a table turns cold after a big win or a large deposit. That feeling comes from a very human habit: we remember the runs that hurt and forget the ones that did not. A game that changed its odds mid-session would show an RTP far below its published figure, and you cannot see that from your own results, which vary by hundreds of rupees even when nothing is wrong.

I cannot verify the fairness of any operator from the outside. What you can do is check whether the app or the game screen names a testing body, keep round IDs and screenshots of any dispute, and raise it with the app's support. Judging fairness by a bad hour is not evidence in either direction.

RNG tables versus live tables

Some card games are run by software (a random number generator) and others by a live dealer on a video stream. I cannot tell you which format each GoPlay table uses, so check the game screen or rules. The difference for your maths is nil: in both cases each round is independent, and the house edge is set by the payout table, not by whether the cards are physical.

• Live dealer: you can see the cards, but you are still trusting the operator's shuffle and stream.
• RNG: nothing is physical, and fairness rests on the software and its audit.
• Neither is easier to beat. No format makes a fixed-edge bet profitable.

What Actually Helps: Habits, Limits and Bankroll

The short answer: these habits lower your cost and protect you from your worst decisions, but they do not create an advantage. Think of them as a way to keep a small edge from becoming a big loss. Here they are in order of impact.

The six habits that matter

1. Skip the Tie bet. It costs about 10.4% at 11:1 against 3.7% for the main bets. If you like the excitement of it, treat it as a small entertainment purchase and cap it, say at one bet in twenty.
2. Stake the same amount every round. Flat staking keeps your cost predictable; raising or lowering after wins and losses only changes the swings.
3. Sit out rounds. Nothing forces you to bet every round. Betting one round in three cuts the hourly turnover, and therefore the hourly expected loss, by two thirds.
4. Set limits before you start. The responsible gaming page describes a daily deposit limit, weekly and monthly limits, session-time reminders, reality-check pop-ups and cooling-off periods. Lowering a deposit limit is immediate; raising one needs a 48-hour cooling-off.
5. Treat the RTP as a long-run number. 97.1% does not mean you get ₹97.10 back from ₹100 in a session. Over 20 rounds the swing is many times the expected cost, as the worked session below shows.
6. Use side bets only after doing the maths. If you cannot state the win chance and the payout, you cannot know the edge.

Bankroll rules that fit these games

A worked example: with a ₹1,000 session budget and the 2% rule, the stake is ₹20. That gives 50 rounds of Andar Bahar, about 25 minutes at the site's pace. The expected cost at 2.9% is ₹1,000 × 0.029 = ₹29, and the standard deviation of the result is about 0.95 × ₹20 × √50 ≈ ₹134. So even this modest plan can end ₹100 or more up or down.

If you are considering the welcome bonus, remember that a bonus is credit with wagering requirements, not cash, and Blackjack is typically weighted at a fraction of the stake, with other card and table games possibly the same. Read the welcome bonus wagering guide and check the Promotions screen for the weighting, maximum bet and expiry. You can also decline the bonus and play with cash only. The wagering maths for a low-weighted game is rarely favourable.

How These Two Games Compare With Blackjack, Teen Patti, Aviator and Rummy

The short answer: Andar Bahar has a modest edge and no skill element; Dragon vs Tiger costs more; Blackjack is the cheapest card game on the app if you play it correctly; and Rummy is the one where decisions really move results. All RTPs below are the published figures.

Speed changes the ranking of these games by cost. At ₹50 a round, Aviator's 3% edge and roughly 300 rounds an hour (from a 12-second round) implies about 50 × 300 × 0.03 = ₹450 an hour, more than Andar Bahar's ₹174 on the same stake. Blackjack's 0.5% edge is the lowest on the app, but only if you follow the strategy in the Blackjack guide. A "low edge" game played fast and badly can cost more than a "high edge" game played slowly.

A Worked 20-Round Session: Expected Versus Actual

Numbers make the gap between a long-run average and a real session clear. Below are three short examples, all clearly hypothetical; they are illustrations built from the maths above, not records of anyone's play.

Example 1: Andar Bahar, ₹50 flat for 20 rounds

Assume you back the favoured side every round at the illustrative 0.9:1 payout. A win pays ₹45 and a loss costs ₹50. The chance of winning is 51.5%, so you expect 10.3 wins. The expected result is 20 × ₹50 × (−0.0215) ≈ −₹21, and at the game's listed 2.9% edge it would be −₹29. Here is one plausible sequence with 10 wins and 10 losses.

You were ₹80 up after round 6, ₹95 down after round 19, and finished ₹50 down. The total staked was ₹1,000. A perfectly average session would be about ₹21 down, so this one was a little worse than average, and nothing unusual happened. The standard deviation for 20 rounds is roughly 0.95 × ₹50 × √20 ≈ ₹212, which is about ten times the expected loss.

To finish ahead you need 11 or more wins, since 11 × ₹45 = ₹495 beats 9 × ₹50 = ₹450 by ₹45. That happens in roughly 46.5% of 20-round sessions. So about half of short sessions end in profit, and yet the long-run average is a loss. Those two facts are not in conflict: the game is priced so that the average of many sessions sits below zero.

Example 2: the Tie bet, ₹50 for 20 rounds at 11:1

In 20 rounds you expect 20 × 31/415 ≈ 1.5 ties. Each tie pays ₹550 and each non-tie loses ₹50, so the net result is 550 × ties − 50 × (20 − ties) = 600 × ties − 1,000. The expected result is 20 × ₹50 × (−0.1036) ≈ −₹104.

More than one time in five you lose the entire ₹1,000, and about 45% of the time you finish ahead. That is what a high-variance, high-edge bet looks like: the frequent big loss and the occasional big win average out to a much larger long-run cost than the main bets. A good result here is not a sign that the bet is fair.

Example 3: Dragon, ₹50 for 20 rounds

With the tie returning half, the expected loss is 20 × ₹50 × 0.0373 ≈ ₹37, and the standard deviation is about ₹217. A typical result might be 9 Dragon wins, 9 Tiger wins and 2 ties: 9 × ₹50 − 9 × ₹50 − 2 × ₹25 = −₹50. Small expected loss, wide swings, and a result you cannot forecast.

Who These Games Suit, and Who Should Avoid Them

These games suit players who want a quick, simple bet with no skill component and who can keep the pace under control. They are a poor fit if you are hoping to outsmart the game or if fast rounds tend to pull you into playing longer than you planned.

If the right-hand column sounds familiar, slow down or step away. The responsible gaming page lists the warning signs (chasing losses, hiding play, losing track of time, playing to cope, borrowing money, irritability when you stop) and helplines including iCall (TISS) on 9152987821 (Mon-Sat, 8am-10pm), Vandrevala Foundation on 1860-2662-345 (24/7) and NIMHANS on 080-46110007. If you would like a card game where skill does move results, the Rummy guide is a better starting point, and the download page covers the install steps for Android.

Frequently Asked Questions