See the top rated post in this thread. Click here

Results 1 to 11 of 11

Thread: Software skill assessments

  1. #1


    Did you find this post helpful? Yes | No

    Software skill assessments

    In Gambling With an Edge's interview with former surveillance operator Junior, he mentions multiple times that skill assessments these days are done with software. How exactly does this get inputted? Video recognition? Or is someone sitting there inputting a player's decisions one by one into a computer? Are index plays taken into account? Would using a different count than the software throw it off?

  2. #2


    Did you find this post helpful? Yes | No
    A remarkably interesting question. It should be answered with the best of technology in mind. The software wouldn’t care about count system used. It would care about determining bet size to advantage, though it would likely be pre programmed with numerous count systems to determine what a player is using.

    It would evaluate human play against a score of 100 for perfect computer play. It would incorporate every split imaginable (there’s an index for absolutely everything), though there are obvious human limitations. One example would be a 77 split v 9 when an obvious surplus of 3’s, 4’s and 10’s are present.

    Against 1 or 2 shoes, it would assign a general level of ability (applicable to most players). For those with demonstrated abilities over and above the norm (pun not intended), it would take several shoes to determine a proper score (not to be confused with SCORE) measurement.

    Regarding my own style of play, I think I would be evaluated as average or limited ability over several average type shoes. This would be achieved through limited type spreads with sometimes less than optimal ramping (stores where I’m known) - probably 2/3 of shoes. However, should that technology catch me in a superior shoe composition - where the fearsome dual ramp FBM ASC Halves Advanced takes over, the alarm bells would start to ring. It would evaluate comparable true counts of all shoes defining average against superior compositions. It would note substantial bet differentials dependent on Quality of True Count, which it could easily determine.

    The most sophisticated versions (doubt they’re available) would incorporate card readers. I digress - likely some guy inputting plays into a keyboard, which ties up staff for lengthy periods. A poor level of sophistication.

  3. #3
    Random number herder Norm's Avatar
    Join Date
    Dec 2011
    Location
    The mote in God's eye
    Posts
    12,461
    Blog Entries
    59


    Did you find this post helpful? Yes | No
    It would make no sense to test against perfect play as the purpose is to find counters. The head of surv at one of the largest casinos on the Strip (who I've chatted with a couple times), stated that the software is rarely used as it is so time-consuming and awkward to use. That was a while back and may have changed. Generally, it would take more than a couple shoes to make a correct determination. DD would be easier. But, how many times have you sat there with a shoe waiting forever for the count to significantly matter. One of the reasons for short sessions. I think it is unusual for the surv room to use such techniques without a call from the pit. They spend most of their time watching the money. I could be wrong and casinos differ. But, I think a well written program would simply wouldn't have enough data in a short session. Although, I'm sure they have chased out some non-APs using it.
    "I don't think outside the box; I think of what I can do with the box." - Henri Matisse

  4. #4


    Did you find this post helpful? Yes | No
    Quote Originally Posted by Norm View Post
    It would make no sense to test against perfect play as the purpose is to find counters. The head of surv at one of the largest casinos on the Strip (who I've chatted with a couple times), stated that the software is rarely used as it is so time-consuming and awkward to use. That was a while back and may have changed. Generally, it would take more than a couple shoes to make a correct determination. DD would be easier. But, how many times have you sat there with a shoe waiting forever for the count to significantly matter. One of the reasons for short sessions. I think it is unusual for the surv room to use such techniques without a call from the pit. They spend most of their time watching the money. I could be wrong and casinos differ. But, I think a well written program would simply wouldn't have enough data in a short session. Although, I'm sure they have chased out some non-APs using it.
    I'm commenting on the potential ultimate in evaluation, further stating that I doubt that it’s there.

    As far as stores chasing out ramping ploppies - Ma Nishtana. Don says it so eloquently - something about lobotomized Neanderthals. It has a certain something.

  5. #5
    Random number herder Norm's Avatar
    Join Date
    Dec 2011
    Location
    The mote in God's eye
    Posts
    12,461
    Blog Entries
    59


    Did you find this post helpful? Yes | No
    Fact is, it can't be there.
    "I don't think outside the box; I think of what I can do with the box." - Henri Matisse

  6. #6


    Did you find this post helpful? Yes | No
    I had dinner one-time with a blackjack hall of famer in Vegas, within 5 mins we had a guy that looked like a Casino backroom staff at the table right next to us even though the restaurant was virtually empty, and he started eating by himself, sadly for him all we talked about was pure math and probabilities. I am guessing facial recognition wasn't looking for me :grin:
    Chance favors the prepared mind

  7. #7


    1 out of 1 members found this post helpful. Did you find this post helpful? Yes | No
    As someone who works with data and analysis in my day job, I can say...
    Quote Originally Posted by Freightman View Post
    A remarkably interesting question. It should be answered with the best of technology in mind. The software wouldn’t care about count system used. It would care about determining bet size to advantage,
    ... this is probably the approach a contemporary system would take. And ...

    Quote Originally Posted by Freightman View Post
    though it would likely be pre programmed with numerous count systems to determine what a player is using.

    It would evaluate human play against a score of 100 for perfect computer play.
    ... this would probably be irrelevant.

    I'm not a Machine Learning expert by any means, but models are built by looking for correlations. Parameters/conditions that create the correlations are irrelevant to the ML tool. One relatively famous example is some banks have gotten in hot water for their "fair" loan approval models based on machine learning analysis of their history of successful loans. What the programmers didn't take into account was that "baked into" the history was previous racially discriminatory lending practices. So, citizen advocacy watchdog groups were able to demonstrate the "new" models were being racially discriminatory, though that certainly was no ones intent while building them.

    So, a ML developed model would simply look for bet size and strategy (splitting, doubling, hit/stand) changes that had strong correlations to other changing factors in the game, no matter how "perfectly" matched to a given system. The programmers wouldn't even need to know what factors to look at for correlations -- the recursive nature of contemporary ML model building would find them.

  8. #8


    Did you find this post helpful? Yes | No
    Arnold Snyder's new book mentions a dialogue with a surveillance agent regarding his use of the blackjack software Blackjack Survey Voice. The agent states that he inputs the player bets and cards played manually into the computer, so it's rarely used in real time to catch counters. He used the software to evaluate counters during a video review after the play. So it's not likely you will get backed off during the play but instead if you return after the play.

    I doubt if using a different count will throw it off and I would guess index plays of different counting systems are programmed in the software.

  9. #9
    Random number herder Norm's Avatar
    Join Date
    Dec 2011
    Location
    The mote in God's eye
    Posts
    12,461
    Blog Entries
    59


    Did you find this post helpful? Yes | No
    IIRC, the software was designed for live use; hence the word "voice". But, this was quite a while back. In any case, I think it was next to useless against an AP.
    "I don't think outside the box; I think of what I can do with the box." - Henri Matisse

  10. #10


    Did you find this post helpful? Yes | No
    In the book, the agent said Blackjack Survey Voice was described to them as something that could be used in real time but he said "...that is realistically impossible with the speed of the game, especially since you have to input all the cards of all players at the table for the program to get an accurate count...".

    Also he described the difficulty of determining a player's bet size looking at the cameras in real time. It was too difficult to judge the height of the chip stack or whether different chip denominations were used. On a video review he was able to see the dealer spreading the chips on a payout or placing the player's chips to the dealer chip tray after a player loss to determine the player's bet size.

  11. #11
    Random number herder Norm's Avatar
    Join Date
    Dec 2011
    Location
    The mote in God's eye
    Posts
    12,461
    Blog Entries
    59


    Did you find this post helpful? Yes | No
    Yep. Voice meant voice input instead of keyboard. It was a stupid selling point since a good typist can type faster than voice recognition, particularly at that time. (Although a limited vocabulary helps.) And yes, it was basically useless at finding actual APs. Pretty good at finding unready new counters that would probably lose money if allowed to stay.
    "I don't think outside the box; I think of what I can do with the box." - Henri Matisse

Similar Threads

  1. Other players skill?
    By hellohilo21 in forum General Blackjack Forum
    Replies: 20
    Last Post: 04-17-2013, 03:32 PM

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

About Blackjack: The Forum

BJTF is an advantage player site based on the principles of comity. That is, civil and considerate behavior for the mutual benefit of all involved. The goal of advantage play is the legal extraction of funds from gaming establishments by gaining a mathematic advantage and developing the skills required to use that advantage. To maximize our success, it is important to understand that we are all on the same side. Personal conflicts simply get in the way of our goals.