Baccarat Side Bet Counting
June 25th, 2026 at 2:00:33 PM
permalink
In all the above discussions, player bets are placed only on positive edge situations, not even on tie situations. The negative edge of the main game is not considered at all.
June 25th, 2026 at 6:33:33 PM
permalink
Quote: acesideIn all the above discussions, player bets are placed only on positive edge situations, not even on tie situations. The negative edge of the main game is not considered at all.
link to original post
But since you are required to make a ‘main bet game’, how do you justify not considering it? It may not override your edge, but you need to ‘consider’ it.
June 25th, 2026 at 7:10:10 PM
permalink
If player bets one unit each on both the main and side bets when side has an edge.
Win unit per 100 hands on main: -1.13,
Win unit per 100 hands on side: 0.493.
Therefore, player loses 0.637 unit per 100 hands of playing both. Overall, it’s a losing game.
Win unit per 100 hands on main: -1.13,
Win unit per 100 hands on side: 0.493.
Therefore, player loses 0.637 unit per 100 hands of playing both. Overall, it’s a losing game.
June 25th, 2026 at 7:23:56 PM
permalink
Quote: SOOPOO
But since you are required to make a ‘main bet game’, how do you justify not considering it? It may not override your edge, but you need to ‘consider’ it.
There are many casinos where you don’t need to place main bets in Baccarat
June 26th, 2026 at 4:03:30 AM
permalink
Quote: harrisQuote: SOOPOO
But since you are required to make a ‘main bet game’, how do you justify not considering it? It may not override your edge, but you need to ‘consider’ it.
There are many casinos where you don’t need to place main bets in Baccarat
link to original post
I was going to mention this too. It is fairly common in baccarat games to only place a "sidebet", not requiring a P/B/T bet to get action. It is also fairly common to be able to sit out for a number of rounds, merely observing, while not forfeiting a spot at the table.
May the cards fall in your favor.
June 26th, 2026 at 5:28:41 AM
permalink
Correction! The negative edge of 0.637 is for player playing every hand of the main game, including extra side bets only on positive edge situations.
June 26th, 2026 at 5:06:45 PM
permalink
Quote: GrahamThorpQuote: acesideI just read another online article by Eliot about card counting Easy 6 side bet and found something more countable. Actually, MDawg has posted this earlier.
We consider a game of 8 decks with a cut card behind 338 cards. Let me list Eliot’s numbers here:
For Easy 6,
Average edge: 3.166%,
Bet frequency: 15.57%,
Win unit per 100 hands: 0.493.
The variance is a lot smaller here too.
link to original post
I have to throw out analysis by Jacobson.
He simply does not factor risk into the equation. His analysis is not useful for that reason
Here's the problem with a simple illustration:
Profitability — Player A vs Player B
Player has a 1% edge on an even-money payoff.
Player B has a 5% edge on a 10-1 payoff.
On a $10,000 bankroll, with correctly-sized Kelly stakes:
Player A (1% edge, even-money): bets $100/wager → average profit $1 per bet
Player B (5% edge, 100:1 long-shot): bets $5/wager → average profit $0.25 per bet
Put simply: the lower variance means the player with the smaller edge makes more money overall for the same risk of ruin.
In the specific case of the easy six it is actually a good bet but about half of Jacobson's published analysis come to the wrong conclusions: he almost always overstates the value of high-variance wagers to a counter.
link to original post
[I have to throw out analysis by Jacobson.
He simply does not factor risk into the equation. His analysis is not useful for that reason]
Not only Jacobson, but most analysts mainly analyze the game rather than the player. Occasionally, they might give an example of how a player can manage their money or the risks involved with their bankroll.
Because the game rules are fixed—such as a 7/8-deck baccarat game with side bet A, we can, on average, get a maximum of n rounds per shoe, with an average EV of +m%.
However, from the player's perspective, different bankrolls, risk tolerances, and table conditions will lead to different betting units and action plans. Using Full Kelly is misleading in the real world. Should it be 1/2? 1/4? 1/6? All players had better check a risk calculator website and make their own plans.
June 26th, 2026 at 5:17:11 PM
permalink
Quote: DieterQuote: harrisQuote: SOOPOO
But since you are required to make a ‘main bet game’, how do you justify not considering it? It may not override your edge, but you need to ‘consider’ it.
There are many casinos where you don’t need to place main bets in Baccarat
link to original post
I was going to mention this too. It is fairly common in baccarat games to only place a "sidebet", not requiring a P/B/T bet to get action. It is also fairly common to be able to sit out for a number of rounds, merely observing, while not forfeiting a spot at the table.
link to original post
In Macau, if no one makes a bet at a Baccarat table, a player can request the dealer to 'fly' that hand without betting anything. This means you can count that round for free. Normally, flying 5 to 10 rounds is perfectly fine; after that, you can either place a minimum bet, walk away, or just sit there waiting for the next player to come. In the VIP area, you can even fly the whole shoe!
June 26th, 2026 at 5:46:03 PM
permalink
Quote: GrahamThorpQuote: 7upQuote: GrahamThorpQuote: harrisThere is no need to argue about Kelly and SCORE endlessly for weeks and I do not think this current conversation is productive. The reality is that some baccarat side bets are countable to the point of being highly profitable, and many people in this thread have first-hand evidence of this.
You haven't provided any substantive evidence to support that at all.
So far we have established that the Dragon 7 and Panda 8 are 400% less profitable than a mediocre blackjack game under the best conditions. That is not "highly profitable".
link to original post
Here are some highly profitable, countable baccarat side bets I know:
Super Tie on a cruise ship: The casino got absolutely crushed for HK$8 million.
Natural 8 / Natural 9 in Singapore: Hit by several teams for a combined total of several million HKD.
Baccarat Insurance in Macau: Netted HK$10 million in just one week.
Dragon 7: Taken down by the JC team for several million HKD.
Any Pair: Attacked by multiple teams, yielding over HK$10 million.
Catch a Rank: Easily a million-plus HKD.
link to original post
Let's make this clear: I was asking Harris to substantiate HIS argument. I want to know whether HE can justify HIS own words. He was asking for the help of powerful, influential people for assistance. If I or anybody else were to do that then we need a reasonable idea he knows what he is doing.
I don't need proof personally. I know which side-bets can be beaten and under what conditions. A minority have impressive SCORE's under ideal, if rare, conditions. I have run the numbers.
Now what concerns me in this thread with your comments is the methodology you are using.
First, anecdotal evidence is NOT permissible as substantive evidence. https://en.wikipedia.org/wiki/Anecdotal_evidence. At best it is useful for hypothesis forming eg deciding what bets to spend time investigating. It is very unscientific.
When you say something like "Johnny Chang beat this game", then there are several possibilities. You may be making it up. You may have got the details wrong. You may have the details right but the wrong information may have been relayed to you. Chang might not have been counting and lied about it to misdirect the casino. He might have been counting, but may have got lucky and later considered the game poor value on further analysis (he made similar statements about blackjack ace location and shuffle-tracking). There are about a hundred plausible explanations The probability of any these being true is small individually but large in aggregate (0.97 x0.95....) Note that I don't have to find you untrustworthy or lacking in competence to believe the thing you stated never happened (in fact I don't, you seem honest and competent) I just have to crunch the numbers on the possible explanations. The same goes for the Wizard with this casino manager story. The compound error is just way too high to take that on trust.
And... we have THREE different studies showing the game isn't worth anything. Your own objection that you have no other games to play is a valid reason to accept a low SCORE for you but not for Chang who is known for his US play.
link to original post
[First, anecdotal evidence is NOT permissible as substantive evidence.]
OMG, this is a forum, not a court.
In a forum, a member's credibility is earned through their track record: the logic within their content, the number of corrections from peers, and the level of support they receive from the community.
I lent John Chang a million HKD in Macau when he needed it for a rebate game. If I were foolish enough to make up a lie like this, it would be incredibly easy to expose.
[And... we have THREE different studies showing the game isn't worth anything. ]
Disagree. Whether it’s worth it depends on the person, not the game. Dragon 7 is countable and profitable. The figures tell the truth.
[Your own objection that you have no other games to play is a valid reason to accept a low SCORE for you but not for Chang who is known for his US play.]
Please do not misrepresent my statement. You are comparing a game with what I called a 'not-so-great income' to a blackjack shoe game that is unavailable in my market. While JC's Dragon 7 analysis dates back a decade, the core rules remain identical to those used today.
June 27th, 2026 at 4:45:42 AM
permalink
I believe you 100%. Regarding Kelly, many authors used this definition,
Kelly fraction = edge / variance,
But, is this correct? Which variance value to use?
Kelly fraction = edge / variance,
But, is this correct? Which variance value to use?
June 27th, 2026 at 11:12:07 PM
permalink
Quote: acesideI believe you 100%. Regarding Kelly, many authors used this definition,
Kelly fraction = edge / variance,
But, is this correct? Which variance value to use?
link to original post
Kelly fraction = Betsize * variance / Bankroll / edge.
If you bet more than one hand in a round, you must find the total edge/round and total variance/round.
When you bet more than one hand in a round, and each hand with different bet size, then it is a bit difficult to find total variance/round(the only way I know is through simulation).
For example :-
TC =+1, 1 hand, bet1= $10
TC =+2, 1 hand, bet1= $20
TC =+3, 2 hand, bet1= $20, bet2 = $40
TC =+4, 2 hand, bet1= $50, bet2 = $100
TC >=+5, 3 hand, bet1= $50, bet2 = $100, bet3 = $150
.
.
.
Total variance/ round = ???
Last edited by: ssho88 on Jun 27, 2026
June 28th, 2026 at 3:59:14 AM
permalink
Quote: ssho88Quote: acesideI believe you 100%. Regarding Kelly, many authors used this definition,
Kelly fraction = edge / variance,
But, is this correct? Which variance value to use?
link to original post
Kelly fraction = Betsize * variance / Bankroll / edge.
If you bet more than one hand in a round, you must find the total edge/round and total variance/round.
When you bet more than one hand in a round, and each hand with different bet size, then it is a bit difficult to find total variance/round(the only way I know is through simulation).
For example :-
TC =+1, 1 hand, bet1= $10
TC =+2, 1 hand, bet1= $20
TC =+3, 2 hand, bet1= $20, bet2 = $40
TC =+4, 2 hand, bet1= $50, bet2 = $100
TC >=+5, 3 hand, bet1= $50, bet2 = $100, bet3 = $150
.
.
.
Total variance/ round = ???
link to original post
[bet more than one hand]
For years, this problem has appeared many times across various forums, but I have never seen anyone provide the answer.
I have posted a workaround to solve this issue before—it doesn't give you the answer with a single formula, but rather involves several steps.
Some websites feature calculators called things like "Multi Bet Calculator," but I don't know how those calculators actually operate.
***I just asked an AI, and it showed me how to calculate it.
June 29th, 2026 at 4:33:00 AM
permalink
Here is my understanding of Kelly. This formula
Kelly fraction = edge / variance,
is plainly wrong, based on dimensional analysis. Here edge is unitless, but variance has a unit of $*$. You just cannot divide these two.
However, when edge is small and the wager is $1, we can approximate variance to be winning payout. For example, in this Dragon 7 side bet, the variance varies wildly from 40 at TC=3.05 to 64.6482 at TC=32, but we treat it as a constant of 40. Therefore,
Kelly fraction = edge / 40.
Kelly fraction = edge / variance,
is plainly wrong, based on dimensional analysis. Here edge is unitless, but variance has a unit of $*$. You just cannot divide these two.
However, when edge is small and the wager is $1, we can approximate variance to be winning payout. For example, in this Dragon 7 side bet, the variance varies wildly from 40 at TC=3.05 to 64.6482 at TC=32, but we treat it as a constant of 40. Therefore,
Kelly fraction = edge / 40.
June 29th, 2026 at 10:41:38 AM
permalink
That edge/variance formula is wrong. Just google "Is edge over odds a useful approximation of the kelly bankroll formula?"
Gambling is a math contest where the score is tracked in dollars. Try not to get a negative score.
September 14th, 2026 at 7:27:51 AM
permalink
Just saw this article about baccarat side bet counting. I don't think anyone can genuinely deny the feasibility of this advantage play now that hundreds of millions of dollars are being earned this way.
September 15th, 2026 at 10:26:14 AM
permalink
Quote: harrisJust saw this article about baccarat side bet counting. I don't think anyone can genuinely deny the feasibility of this advantage play now that hundreds of millions of dollars are being earned this way.
link to original post
The article's only source is a blog on a site which is a poorly-generated website with numerous AI tells and no obvious human input, everything from pretentious , meaningless, jargon to the excessive use of bullet points gives it away.
September 16th, 2026 at 2:46:41 AM
permalink
Quote: harrisThe blog belongs to Differential Labs which is a real and legitimate organization.
link to original post
I'm sure it is a real organization but the website is classic AI BS.
The one concrete claim they make is that they used ai to tell a company they were giving away too much money in rebates.
You don't need ai to work out if you offer rebates anyone playing short sessions is going to profit. I could have told them that. Without using a computer. If we flip coins and you give me back 10% that's a winning proposition for me. What they don't say, and would have almost certainly happened, is whether the casino subsequently lost money overall when it removed a loss-leader and alienated its high-rolling patrons, who were probably sucked in by the rebate but ended up losing it back on slots etc 50x over. That's inevitably what happens when a casino hires a consultancy.
I can see you are so keen to work for casinos as a consultant if they are so easy to scam with technojargon.Maybe I can sell them lead shielding to thwart advantage players with x-ray vision or telekinesis or something.
September 16th, 2026 at 9:14:08 AM
permalink
The page mentions online, which is the area that is actually potentially profitable. Assume live dealer with actual shoe, the low edge is much less of a problem as robots do not cost much to run (even nothing for bot nets) and they can play hundreds of games at once.
September 16th, 2026 at 9:46:23 AM
permalink
Quote: acesideI just read another online article by Eliot about card counting Easy 6 side bet and found something more countable. Actually, MDawg has posted this earlier.
We consider a game of 8 decks with a cut card behind 338 cards. Let me list Eliot’s numbers here:
For Easy 6,
Average edge: 3.166%,
Bet frequency: 15.57%,
Win unit per 100 hands: 0.493.
The variance is a lot smaller here too.
link to original post
It's funny when the cut card is a 6, dealers usually declare that it will be a "Tiger shoe" while in my mind I am thinking, minus 3.
Also when the Bank and Player tie 6-6, the crew declares that "the Bank is trying to Tiger" (which actually, if the ties come about with 7s and 9s, then the count is moving towards more chance of a Tiger, but not if a 6 card is involved in the 6-6 ties).
Of course sometimes that is exactly what happens you have a 6-6 tie followed by a Tiger, and then the house erupts with, "I told you sos!"
As far as free hands, in Vegas at all of the majors no one cares if a person bets every hand or just one hand a shoe. However, if someone has a reserved table as I do, betting just a few hands a shoe might result in losing that reserved table privilege.
I tell you it’s wonderful to be here, man. I don’t give a damn who wins or loses. It’s just wonderful to be here with you people.
https://wizardofvegas.com/forum/gambling/betting-systems/33908-the-adventures-of-mdawg/
September 16th, 2026 at 5:35:26 PM
permalink
Quote: itsmejeffThe page mentions online, which is the area that is actually potentially profitable. Assume live dealer with actual shoe, the low edge is much less of a problem as robots do not cost much to run (even nothing for bot nets) and they can play hundreds of games at once.
link to original post
That assumes you can find a game dealt down to the last quarter-deck. I'm not aware of any such game.
This happened a lot in the early days of live dealer, with 98-100% penetration being quite common. The games got hit quite hard, the pen never recovered.
September 16th, 2026 at 8:03:51 PM
permalink
Standard Baccarat has 16 cards cut off at the end, no more, no less. But there aren't too many casinos that do it the old school way, and nowadays up to 1 1/2 full decks are cut away.
If you are unaware of any casinos that cut away only 16 cards, you probably play only at Indian casinos.
If you are unaware of any casinos that cut away only 16 cards, you probably play only at Indian casinos.
I tell you it’s wonderful to be here, man. I don’t give a damn who wins or loses. It’s just wonderful to be here with you people.
https://wizardofvegas.com/forum/gambling/betting-systems/33908-the-adventures-of-mdawg/
September 17th, 2026 at 3:47:59 AM
permalink
Yes, I know. We were talking about online games.
Also, for these purposes you have to factor in unseen burn cards which makes a huge difference.
Also, for these purposes you have to factor in unseen burn cards which makes a huge difference.
September 17th, 2026 at 7:31:54 AM
permalink
We’re talking 8 decks and the burn card could be an ace.
I tell you it’s wonderful to be here, man. I don’t give a damn who wins or loses. It’s just wonderful to be here with you people.
https://wizardofvegas.com/forum/gambling/betting-systems/33908-the-adventures-of-mdawg/
September 22nd, 2026 at 8:09:06 AM
permalink
Just heard from a casino employee somewhere in the USA that his casino's EZ Baccarat table lost $72,000 last night. The maximum bet is $500. I really really really highly doubt that this was not the work of side bet card counters.
September 22nd, 2026 at 8:52:52 AM
permalink
Quote: harrisJust heard from a casino employee somewhere in the USA that his casino's EZ Baccarat table lost $72,000 last night. The maximum bet is $500. I really really really highly doubt that this was not the work of side bet card counters.
link to original post
Is the max bet $500 or $500 times nine spots? If a player can bet $4500 a hand, it becomes a lot more likely.
Casinos are not your friends, they want your money. But so does Disneyland.
And there is no chance in hell that you will go to Disneyland and come back with more money than you went with.
- AxelWolf and Mickeycrimm
September 22nd, 2026 at 9:06:42 AM
permalink
Quote: harrisJust heard from a casino employee somewhere in the USA that his casino's EZ Baccarat table lost $72,000 last night. The maximum bet is $500. I really really really highly doubt that this was not the work of side bet card counters.
link to original post
Every counter wishes what you just said were true! But it is not so. The chances of a random player having a very lucky night, sidebets or otherwise, is far greater than there both being a counter present at a table, and there being enough advantage situations over the course of a night to win that amount, and those advantage situations converting to actual wins.
A night at a table is not nearly enough data to distinguish, by results, whether or not someone is playing with an advantage.
September 22nd, 2026 at 9:28:55 AM
permalink
Quote: harrisJust heard from a casino employee somewhere in the USA that his casino's EZ Baccarat table lost $72,000 last night. The maximum bet is $500. I really really really highly doubt that this was not the work of side bet card counters.
link to original post
Based on heavy side-bet action and a handle of $1.2m the chance of this is occurring is 1 in 59, not even 2 standard deviations.
The chance of an outlier win like this doesn't actually change much even if you assume there were counters there, eg
Team of 3 counters, 50% of shoes each $18,806 $28,488 0.183% chance of win > 72k (~1 in 550 nights) — 7.5x baseline
September 22nd, 2026 at 11:51:34 AM
permalink
Thanks for doing the math.
I asked the same guy and he said multiple bets at a time is not allowed. I concede it's possible the players just got really lucky, I would have to go to the casino and find out more about their baccarat rules and procedures to know for sure. Or even better, see the surveillance tapes lol.
I asked the same guy and he said multiple bets at a time is not allowed. I concede it's possible the players just got really lucky, I would have to go to the casino and find out more about their baccarat rules and procedures to know for sure. Or even better, see the surveillance tapes lol.
September 22nd, 2026 at 10:45:41 PM
permalink
Quote: harrisThanks for doing the math.
I asked the same guy and he said multiple bets at a time is not allowed. I concede it's possible the players just got really lucky, I would have to go to the casino and find out more about their baccarat rules and procedures to know for sure. Or even better, see the surveillance tapes lol.
link to original post
9 different players can bet $500 on the same side bet that pays 40-1 and win a total of $180,000.
September 22nd, 2026 at 11:35:01 PM
permalink
Quote: KevinAAQuote: harrisThanks for doing the math.
I asked the same guy and he said multiple bets at a time is not allowed. I concede it's possible the players just got really lucky, I would have to go to the casino and find out more about their baccarat rules and procedures to know for sure. Or even better, see the surveillance tapes lol.
link to original post
9 different players can bet $500 on the same side bet that pays 40-1 and win a total of $180,000.
link to original post
And players could do that just randomly and it would hit about 1 out of 43 times. That's the kind of thing I used to do with free roll, use-it-then-lose-it vouchers, when they would let me.
September 23rd, 2026 at 2:54:12 AM
permalink
Quote: KevinAAQuote: harrisThanks for doing the math.
I asked the same guy and he said multiple bets at a time is not allowed. I concede it's possible the players just got really lucky, I would have to go to the casino and find out more about their baccarat rules and procedures to know for sure. Or even better, see the surveillance tapes lol.
link to original post
9 different players can bet $500 on the same side bet that pays 40-1 and win a total of $180,000.
link to original post
If this was ap, which is unlikely, it is probably steering or some kind of shuffle exploitation (very occasionally a step is missed in the shuffle procedure and a slug or even the entire deck is un-shuffled). Something where you'd have a very strong edge.
Ironically if you can make a casino paranoid about card counting at baccarat you can get them to increase the number of shuffles per hour, which is the attack vector. If they start cutting off multiple decks the casino will actually lose money faster.
September 23rd, 2026 at 10:27:29 AM
permalink
Quote: GrahamThorp
Based on heavy side-bet action and a handle of $1.2m the chance of this is occurring is 1 in 59, not even 2 standard deviations.
link to original post
I’ve read this line a few times but still don’t get what you are saying. What is a handle of $1.2m?
September 25th, 2026 at 7:09:54 PM
permalink
$500 dragon on table is standard in lot casinnos. $20k per dragon, if there's 4 extra dragons than normal thats 80k, very easy to lose 72k as the house. its only a 7 percent disadvantage with huge variance. Some shoes have 6 dragons, some have 0, and some days are lot more than others
