Listed below is an excerpt of an article from Blackjack Forum, Spring 1997, The High-Tech Eye by Arnold Snyder.
This column contains comments by Oliver Schubert, the author of Blackjack Survey Voice. Read the paragraph to the left first, then read his comments.
The High-Tech Eye
" Blackjack Survey Voice" is the most sophisticated of the software packages currently being sold to casinos for the express purpose of identifying professional blackjack players.
Blackjack Survey Voice is an excellent tool for identifying professional blackjack players, but one major feature is initially often overlooked: Player Ratings. BJSV will tell you exactly what kind of advantage (in percent) the casino has over a particular player (which might be negative against a card counter.)
The 
speech recognition
feature of BJSV allows the operator to vocally enter the action occurring
instead of using a keyboard for entry. This feature appears
to work well, ...
Although BJSV follows the play of the hands, and rates players' talents according to both basic strategy and counting strategy, the determination of whether or not a player is a card counter is primarily weighted -- as it should be -- towards the betting strategy. In my opinion, any player following a traditional card counting approach to betting -- even if "cover" bets are employed along with parlaying, camouflage, etc. -- would likely be accurately identified as a card counter by the BJSV software within 75-125 hands of play in a 6-deck shoe game.
And, unfortunately, I believe it would be impossible to camouflage a traditional card counting betting strategy from BJSV. This is because BJSV can translate the betting history of any player to graphs.
Graphs are not only available to visualize card counting, but also for shuffle tracking, ace counting and ace location.
To make it even simpler for a casino to analyze
betting data, BJSV will produce a graph
which ignores the chronology of the bets entirely, but merely shows the
player's bet size at each true count which occurred during the course
of the analysis.
No player employing a traditional card counting strategy can hide from this type of analysis.
A complete BJSV report on an identified card counter is
reproduced at the right. Note the incredible
amount of detail included in this report. Note also that although
the counter lost money (at the rate of 17.77%!), and made 8 deviations
from basic strategy
(only two of which were correct!), the BJSV software correctly identified
the player as a card counter because of the player's accurate betting
strategy (0.0% Counter Deviations). BJSV estimates that despite this player's
lack of talent, the house has only a 0.08% long run advantage over him.
See the Synopsis Report on this same
sample player, reproduced on page 27, which BJSV also provides. Even the
dumbest pit boss can understand this report! This black
chip player may pose no real threat to the house, but he may find it difficult to get his share of comps
after the pit reads this synopsis.
Arnold Snyder, the author of this article,
misinterpreted one of the numbers in the Summary
Report. The 8 deviations listed, reflect how many times the player
had the opportunity to deviate from Basic Strategy,
of which the player did deviate 2 times. The player actually made 9 mistakes
against Basic Strategy, of which 2 were correct according to the count.
Unfortunately, Blackjack Survey Voice is being used in a growing number of big Las Vegas casinos.
Unfortunate for Card Counters, Shuffle Trackers, and Hole Card players.
Arnold Snyder
Blackjack Forum
Oliver Schubert
Casino Software & Services