David Spence: Cheating via Selective Up Card. For Don or Norm

[David Spence]

Whenever I experience a significant losing session, part of me invariably suspects that the loss was due to cheating. I realize this suspicion is driven entirely by my ego (okay, maybe my id plays a role too :-) ), and that the likelihood that I was actually cheated is roughly 0%.

Nevertheless, I have seen what was probably cheating on one occasion. The dealer, after looking at both cards, sometimes turned up the first card and sometimes the second. This dealer also showed 10s and aces far more than what would be expected.

This procedure affects neither the dealer's hand nor the players' hands, but it does have a huge effect on win (and, in this case, loss) rates. If the dealer RANDOMLY turns up one card or the other, naturally, it shouldn't make any difference. However, if the dealer looks at both cards, then turns up a 10 or ace whenever possible, this will have an impact on the player. Player decisions are based on the fact that a PARTICULAR dealer card is a given value. This is very different from AT LEAST ONE of the dealer's cards having that value. As Wong mentions, players will bust far too often.

So here's the question: what is the precise effect on win rate if the dealer selectively turns up a 10 or ace whenever possible? Assume that if neither dealer card is a 10 or ace, the dealer turns up a random card.

The effect will be huge regardless of the rules, but, for the sake of argument, assume 6D, S17, DAS, 75%pen. Even the effect on a basic strategy player would be interesting to know, but assume the player uses the I18 for high-low and a 1-8 spread.

David Spence

[Norm Wattenberger]

Assuming that the dealer shows the Ten when he has BJ, 6D, S17, DAS, 75%, Ill18, 1-8:

Honest deal SCORE: 13.8 EV: 0.6%
Crooked deal SCORE: -490 EV: -3.6%

Now if you can get the dealer to put up the low card instead:

Honest deal SCORE: 13.8 EV: 0.6%
Crooked deal SCORE: 340 EV: 3.2%

[David Spence]

Wow! That was fast! Thanks, Norm. How did you do that so quickly?

I was surprised to see that selectively showing the low card actually helps the player, despite the fact that strategies are based on an essentially random choice of upcard. But I suppose it makes sense--knowing the dealer has at least one low card ensures he doesn't have a great hand. Knowing the dealer has at least one 10 or ace doesn't ensure anything.

> Assuming that the dealer shows the Ten when he has BJ,
> 6D, S17, DAS, 75%, Ill18, 1-8:

> Honest deal SCORE: 13.8 EV: 0.6%
> Crooked deal SCORE: -490 EV: -3.6%

> Now if you can get the dealer to put up the low card
> instead:

> Honest deal SCORE: 13.8 EV: 0.6%
> Crooked deal SCORE: 340 EV: 3.2%

[Norm Wattenberger]

Would be even better if I favored 4, 5 and 6 over other non-high cards. Knowing the dealer doesn't have 20 makes a big difference.

> I was surprised to see that selectively showing the
> low card actually helps the player, despite the fact
> that strategies are based on an essentially random
> choice of upcard. But I suppose it makes
> sense--knowing the dealer has at least one low card
> ensures he doesn't have a great hand. Knowing the
> dealer has at least one 10 or ace doesn't ensure
> anything.

[Don Schlesinger]

> Wow! That was fast! Thanks, Norm. How did you do that
> so quickly?

After screwing around over at BJ21, you've finally figured out where to get REAL information from! :-)

> I was surprised to see that selectively showing the
> low card actually helps the player, despite the fact
> that strategies are based on an essentially random
> choice of upcard.

Why be surprised? If the dealer's hand is T,6, you don't see what a huge difference it is to play against the 6, rather than against the T??

> But I suppose it makes
> sense--knowing the dealer has at least one low card
> ensures he doesn't have a great hand. Knowing the
> dealer has at least one 10 or ace doesn't ensure
> anything.

You never bust you hand against the dealer's 6. You bust repeatedly against his T. You split and double hands vs. the 6 that you can't vs. the T. HUGE difference.

Don

[David Spence]

> After screwing around over at BJ21, you've finally
> figured out where to get REAL information from! :-)

Without a doubt...though bj21.com has its merits, the signal/noise ratio here is markedly higher.

> Why be surprised? If the dealer's hand is T,6, you
> don't see what a huge difference it is to play against
> the 6, rather than against the T??

Okay, you've put me in my place...I had no business being surprised :-)

My misplaced surprise stemmed from the fact that strategies are based on an assumption (random choice of upcard) not true in the "crooked" game. It just so happens that this form of crookedness (selectively showing a low card) helps the player to such an extent that it overshadows the invalid assumption.

> You never bust you hand against the dealer's 6. You
> bust repeatedly against his T. You split and double
> hands vs. the 6 that you can't vs. the T. HUGE
> difference.

> Don

[Sun Runner]

> My misplaced surprise stemmed from the fact that
> strategies are based on an assumption (random choice
> of upcard) ...

That was an excellent question by the way.

Not to drag you down to my level of seeing through glass darkly, I thought the same thing you did initially and was having trouble getting away from the randomness aspect of it. Don's T6 example was spot on of course.

[Magician]

> Would be even better if I favored 4, 5 and 6 over
> other non-high cards. Knowing the dealer doesn't have
> 20 makes a big difference.

Are you saying that the figures above assume the player knows that the game is crooked and is altering his strategy to compensate?

[Norm Wattenberger]

[Don Schlesinger]

> Without a doubt...though bj21.com has its merits, the
> signal/noise ratio here is markedly higher.

There is no noise at all here. We don't tolerate it.

> Okay, you've put me in my place...I had no business
> being surprised :-)

> My misplaced surprise stemmed from the fact that
> strategies are based on an assumption (random choice
> of upcard) not true in the "crooked" game.
> It just so happens that this form of crookedness
> (selectively showing a low card) helps the player to
> such an extent that it overshadows the invalid
> assumption.

Sometimes dealers cheat for the player to win, because they are in collusion with him, no? And, continually showing a low card is a huge edge for the player.

Don

[Magician]

I get what you were saying now. If the dealer were to choose the upcard most favourable to the player (first 6 then 5, 4, 3, 2, 7, 8, 9, T, A) the edge would be even greater than if she chose randomly from two low cards.

Back to the original post, if the player knew that the dealer were cheating by always exposing an ace or ten if possible, he could use this information to his advantage by altering his strategy, correct? More information should translate into a higher edge.

[Norm Wattenberger]

If you saw a Ten up, dealer would have 20 or 21. But you would still have a lousy game.

> I get what you were saying now. If the dealer were to
> choose the upcard most favourable to the player (first
> 6 then 5, 4, 3, 2, 7, 8, 9, T, A) the edge would be
> even greater than if she chose randomly from two low
> cards.

> Back to the original post, if the player knew that the
> dealer were cheating by always exposing an ace or ten
> if possible, he could use this information to his
> advantage by altering his strategy, correct? More
> information should translate into a higher edge.

[victoria]

> Whenever I experience a significant losing session,
> part of me invariably suspects that the loss was due
> to cheating. I realize this suspicion is driven
> entirely by my ego (okay, maybe my id plays a role too
> :-) ), and that the likelihood that I was actually
> cheated is roughly 0%.

> Nevertheless, I have seen what was probably cheating
> on one occasion. The dealer, after looking at both
> cards, sometimes turned up the first card and
> sometimes the second. This dealer also showed 10s and
> aces far more than what would be expected.

> This procedure affects neither the dealer's hand nor
> the players' hands, but it does have a huge effect on
> win (and, in this case, loss) rates. If the dealer
> RANDOMLY turns up one card or the other, naturally, it
> shouldn't make any difference. However, if the dealer
> looks at both cards, then turns up a 10 or ace
> whenever possible, this will have an impact on the
> player. Player decisions are based on the fact that a
> PARTICULAR dealer card is a given value. This is very
> different from AT LEAST ONE of the dealer's cards
> having that value. As Wong mentions, players will bust
> far too often.

> So here's the question: what is the precise effect on
> win rate if the dealer selectively turns up a 10 or
> ace whenever possible? Assume that if neither dealer
> card is a 10 or ace, the dealer turns up a random
> card.

> The effect will be huge regardless of the rules, but,
> for the sake of argument, assume 6D, S17, DAS, 75%pen.
> Even the effect on a basic strategy player would be
> interesting to know, but assume the player uses the
> I18 for high-low and a 1-8 spread.

> David Spence

Funny seeing this post now.
Later in the same day that you began this thread and not reading your thread till today, I left a blackjack table for just that reason. Experienced dealer was sometimes choosing his first card (proper procedure) and at other times choosing his second card. So I wong out in a positive count and watch the next three hands but actually I am watching his eyes. He is not only picking different cards to turn but his eyes lead me to believe he is peeking at the card as it comes out of the shoe. Looked a while longer and any doubt in my mind was gone, high card showed almost every time but more devious than just that, he showed his 3 when he had 11.

Had I not been an AP I might have asked the dealer about it or asked the pit, but felt it was just better to leave.

Victoria