# Is it possible to count cards by literally counting cards?

Typical card-counting, which treats 2,3,4,5,and 6 as "+1" and 10's as "-1", requires card counters to keep track of each card as it's dealt. I wonder if you can get a reasonable approximation of this by just counting the number of cards that are on the table and ignoring what cards they are.

For example, when a player stands after exactly two cards, they usually have something like 10-8, 10-10, 9-7 (against dealer 6), etc. so the count is usually negative. If a player stands after three or more cards, there's usually a 2 thru 6 in there, often more than one, so the count is usually positive.

Through simulation, general averages could possibly be determined; for example, hypothetically, Two cards = -0.75; Three cards = +0.6, Four cards = +1.8, Five or more cards = +3.25.

I'm envisioning a system where instead of counting +1, -1 as each card gets dealt, you begin with a negative "x" count (under the assumption that each player will use exactly 2 cards), and then increase the count by "y" for each card beyond the second that any player receives, regardless of its value.

Edited to add: This would be a cumulative count, and at the beginning of each new hand you'd continue by again subtracting "x" from your current count.

Could such a system provide enough advantage to be useful to a beginner? Thank you very much.

The big problem I see right off the bat is that just counting the two card hands wouldn't give you an understanding of the cards left in the shoe, it would be too imprecise: Here's the 2 card hands you could (ever) possibly stand on:

12 (vs some dealer up cards) 13 (vs some dealer up cards) 14 (vs some dealer up cards) 15 (vs some dealer up cards) 16 (vs some dealer up cards) 17 (vs any) 18 (vs any 19 (vs any) 20 (vs any) 21 (vs any)

for a 12 the combo of cards is 10+2, 9+3, 8+4, 7+5, 6+6. So one hand with a +0 count, and 3 hands with a +1 and one hand with a +2 all on top of the fact that the dealer would have a bust card (3-6) face up in any instance that a player would stand on a 12, so if people stand on a lot of 12s, the deck is getting heavier, or good for the player.

For a 20 however, the combo is A+9 or 10+10 one hand at a -1 or one hand at -2. both make the deck lighter or negative for the bettor. So if people are standing on a lot of 20s then the deck is generally getting lighter or worse for the player.

So, there's a marked difference from one two card hand to another, so in order for this system to have any chance at all, you'd actually have to remember how many 12's stand, how many 13's stand etc. And it's not going to be as accurate as a hi-lo count, because sometimes in every hand the deck isn't actually getting better or worse. Consider an A+9 vs a dealer up card of 5. That's a net zero change in weight.

At this point I'd say the Hi-lo count is better because it's simpler. A simple system that gives you an edge is all you ever need.

The disadvantage of your system is that it is imprecise, as discussed below; the advantage is that it is a "nonstandard" method of card counting, and you may get less "heat" from the casino. But I think you're onto something here.

First, there are 20 10- and 11- valued cards (counting aces) in the deck, and 32 other cards, in a ratio of nearly 2 to 3.

Second, if someone stops at two cards, especially against a dealer's 4,5, or 6, it's likely that their total count is the the 16-21 range, meaning that at least one card is a "10." And if the values are 10-8 or 9-9, the second card is "close enough.

If someone draws a third card, that suggests at least one, probably two non-ten cards, originally. Note that it is possible to "bust" with something like 9-8-7. or even 6-7-9.

If someone draws a fourth card, that suggests an average value for the first three cards of four or less, meaning all low cards (except possibly the last).

So a bunch of two card draws are not good for the player, three card draws are normal to good, and four card draws are great.

A player playing perfect strategy and "flat" betting is about a 1% disadvantage to the house, but if you can use the above to vary your bets, it could cancel out the house advantage.

If the cards are randomly shuffled and if you're just counting the number of cards without regard for their values, then you effectively have a constant "count" of 0. The whole point of counting is to monitor the values of the played cards to determine what values are left in the deck.

• You seem to have misunderstood the question, the values of individual cards are statistically inferred from how many other cards a player has in a hand. May 29 '16 at 6:07
• @hkBst Give it a shot. I have two cards, and I stand. Dealer has a 5 showing. What are my two cards? May 29 '16 at 16:38

The answer is almost certainly no. Card counting is used primarily to determine how much to bet - a standard strategy might look like:

• -1 or smaller: find a better table
• -1 to +1: bet table minimum (still in casino's favor but worth waiting around)
• +1 or higher: start betting big (the game has tilted in your favor)

Now I haven't run the simulations, but my guess is that the correlations you'd find are too small, e.g. Two cards = -0.05; Three cards = 0, Four cards = +0.05. Unless any of the numbers are greater than 1 or less than -1, you can't do anything useful with the information.

Ultimately, I think if it were as easy as counting the cards on the table, everyone would be doing it.

• One possible source of confusion: "the count" that we care about is for all cards that have ever been dealt from that shoe, so even if standing with 2 cards usually means that the count currently on the table is negative, it will have a much lower correlation to the overall count of the deck. May 25 '16 at 17:45
• Thanks Benjamin, I've edited the original question to specify that the count should be cumulative. I'm not suggesting that you just do it for the current hand and expect that to be enough. May 25 '16 at 18:08