How specific is it?? And insight to my sidecounting - (long)
In other words, how close to a "match" does it have to be for the alarm to be set off? I have several ideas on varying one's play that may or may not be successful.
Please note that these are just theories. I know nothing about the computer technology mentioned other then what has been posted. I also know next to nothing about computers in general. However, like most who post on these pages...I feel I know a fair amount about BJ.
Example - Spreading 10-80...let's say my count is at +4 and thus I should bet 4x or $40. What happens if I bet $30...or if I bet $42 or $59? Or two hands of $23?
It seems to me the computer would have to be pretty sharp to detect that a player is spreading to 2 hands.
How would it compensate for that?? Is there a butt-detector in the seats?? Can I defeat this by standing at the table (which I occassionally do)? Just kidding. But this is similar to the team play idea mentioned previously. The pit has to watch for players going from 1-2 hands because I doubt the computer can do it.
Is it really programmed with every kind of count? That seems unlikely to me also. I side-count Aces and usually another card if I find a demonination in abundance or lack thereof at the beginning of the shoe (if I see four 8's on the first hand then I'll start counting 8's and adjust play accordingly) so if it's analyzing my play I should be able to throw it off pretty easily.
Lets say I hit on my 16 v. 10 despite the slightly plus count because I have noticed an abundance of 5's remaining (or there has been a depletion of 6's which adversely effects the count for purposes of this playing decision).
Here is another example...on a DD let's say I have a 12 v. 3 and the count is even. Obviously this is normally a hit. But if I have also seen 7 or even all 8 of the 8's come out of the DD then I should stand because for these purposes the 8 counts as a 'plus' card (helps both the player and the dealer make their hand). So if I stand on this one instance of 12 v. 3 perhaps that would be enough to tell the computer that even though I may be counting I am obviously not very good at it (unless it side-counted the 8's also, which strikes me as unlikely).
Conversely, if the count is slightly plus on 12 v. 3 I may decide to hit based on my chances of pulling an 8 (have seen only 1 and we're near the end).
I don't deviate from my count strategy too often but it seems to me that just 1 or 2 of these types of plays would rapidly get the computer heat off of you. At least, I know it works with the human survelliance (they typically stop paying attention to me when they determine that my winnings must be 'lucky'). Plus, it adds to my overall EV by giving me a couple of wins that I normally would not have won if I had played my hand strictly by the count.
Lets say I'm at a pitch game and I'm seeing 2 or 3 players' hands to help me with my decisions....but I'm not able to see their hands all the time. This inconsistency should throw off the computer too...how does it know who's hand I successfully saw?
Lets say I make an assumption about the 2 players at the end of the table who both tuck their cards against an Ace....You can usually assume a couple of 10's over there between the 2 players...maybe even 3-4 but I'm guessing one on each (10, 7 and 10, 8 perhaps). So I decide not to take insurance based on this only to see that they both had A, 9. With this knowledge I would have taken insurance....but since I didn't perhaps I should breathe a sigh since this tells the computer I'm not very good at when to take insurance because the count clearly dictated that I should have.
All right...maybe it doesn't analyze how the player actually plays. Maybe it just analyzes the count and whether the bets are being raised accordingly....
What if I leave my high bet out there after a push because I'm afraid of drawing attention despite the rapidly plummeting count. Won't the computer detect that I am betting abnormally high here?? (which I am but I am sometimes not comfortable pulling my bet back after a push).
here's another one...What if I change which count I'm using from shoe to shoe (which would be a problem for some but is fairly easy for me, especially against a slightly slower dealer in a face-up game).
I go from a count that considers 3-6 + and 10's a minus on one shoe. Then I add in 2's and 9's on the next one. Or on one shoe I decide not to raise on a slighty plus count because no 5's have come out yet and I know that in the actual percentages...the 5's are more valuable to the dealer than the 3's (even though they are counted the same in most level-1 counts). This one play on it's own takes me to a level-2 count just for that hand (essentially counting the 5's as a +2...or at least as being more valuable then the other low cards.... for the time being).
Same thing goes if I raise slightly (from $10 to $15) on a +1/2 TC based on an extremely favorable A's to 5's ratio.
Except for directly changing my count in mid-session, these are all things I typically do anyway.
This doesn't even account for purposeful camo plays to throw off the scent such as doubling hard-12 or not splitting 8's or that type of thing.
Conclusion - based on these ideas, I think you would have to play exactly by the "book" that the computer was programmed with to be caught. Some players already do this of course. But it should not be too hard to throw in an occassional camo play...and with side-counts or changing up the kind of count used then it's possible to have camo plays that actually help your EV.