What's the best count system?
  • I started out learning basic high/low and recently heard about red seven count and have been practicing that system lately. But before I get confused from changing systems, can I get some feedback from some experienced counters on what is the most effective count system?I figure it will be easier on my brain if I pick a system and stick with it.
  • Red-7 is Arnold Snyder's (the bishop) counting system. It's unbalanced, which eliminates the true-count conversion that confuses many. But it is less powerful than Hi-Lo from a playing efficiency point of view.

    There are several stronger counting systems than Hi-Lo, but these level-2/level-3 counting systems are much harder to use, and often require a side-count as well, which totally kills the accuracy of most counters.

    Bottom line is to use the best you can use. Hi-Lo is the most popular counting system on the planet. There's a reason for that. It's what I use myself, in fact, from single-deck through shoes. Yes there are better systems for single-deck (better playing efficiency). But I've learned to count with Hi-Lo without worrying about inaccurate counts. And I've not seen any reason to change, as the difference between a basic unbalanced count like KO or red-7 and hi-lo is not that big, nor is the difference between hi-lo and a level-2 or 3 counting system very big. In fact, hi-lo played with 100% accuracy is better than any level-2 counting system if you can't keep an accurate count and make mistakes...

    Rather than changing counting systems, get one down pat and then use it...
  • Actually, Red 7 is a tad bit better than H/L...same for betting...better for
    playing.

    For betting, most come in at about 97%. So take your pick and stay
    with it. (betting is the main consideration, especially for shoe games)

    Level one

    KISS........96-55-UB
    Red 7.......97-53-UB
    High/Low..97-51-B
    KO...........97-55-UB

    Level two

    Revere.....98-55-B
    Zen..........96-63-B
    High Op II.91-67-B-A/SC

    Level three

    W halves...99-56-B

    etc
  • the problem I have with those numbers is that I do not believe that an unbalanced level-1 count is going to beat hi-lo in playing efficiency. If you look at Arnold's "Blackbelt in BlackJack" he clearly agrees and says his red-7 count is not as strong as hi-lo, but is easier to use (no TC conversion)...

    one easy example is that on 16 vs 10 I know when to hit or stand (index=0) but with an unbalanced count, until I get to some key count (like the pivit for betting) I don't. Unless I do a TC conversion, which the unbalanced counts don't do...

    For example, arnold also discusses the "Zen count", and points out that it will add about .1% to your advantage in a shoe game, or about .2% to your advantage in a SD game. This from a level-2 count that is significantly more difficult to use...

    I don't know what's up with those PE numbers (I find the same numbers on the q-fit web site). But everyone I know (including Arnold) says that Hi-Lo is better than red-7, yet those numbers suggest red-7 is better...
  • This one may not be the "best", but it is very accurate and consistent :
    2,3,4,5,6,7 = +1
    8 = 0
    9, 10, Ace = -1
  • Surfteg- Let me try this again. I don't know what I did, but the count
    system you described is called the "Silver Fox System". I have been
    using it for many years now, but I still call it H/L because when I added
    the 7 and 9 I didn't know what to call it. The numbers I have on it are
    96 and 54. Also, I might add that the numbers on these systems will
    vary some depending on who does the testing and what ever other things
    that can be applied. At any rate, they are so close that I don't think it
    matters.
  • How about the UBZ2?
  • Ray...I first "discovered" THAT count in Ed Thorp's Beat The Dealer book almost 18 years ago fwiw...also, for me, it definitely works best in a DD game...
  • Alex, you got me there.....What are the card values?
  • that is George C's, "Unbalanced Zen II" system. Don't know the tags offhand...
  • Ray,
    Check out www.qfit.com/cvstrat.htm .There's a list of #'s for a lot of the different systems.
    Vytas
  • I've got Unbalanced Zen @ 96/62...just didn't make the connection. A
    sure sign missing neurons.
  • The reason an balanced count with a lower BC and PE might outperform an unbalanced count with better numbers is that the balanced count implements its BC and PE more accurately than the unbalanced count. This is particularly true when it comes to PE (playing efficiency). Unbalanced counts have an error range in implementing their index numbers -- especially at negative counts.
    This is a good time to point out that unbalanced counts whose pivot is +4 true have a greater error range than those with a pivot of +2 true. That's why Hi/Lo (97 & 51) outperforms KO (97.5 & 55). Yet KISS III (96 & 55) slightly outperforms KO and falls just a hair shy of Hi/Lo.
  • I use Fred Renzeys' kiss III count and it's been great to me. Learned it in his fantastic book Blackjack Bluebook II
    Good luck
  • What are the card values for the KISS III system?
  • KISS III card tags are:

    red 2 = 0
    blk 2 = +1
    3 = +1
    4 + +1
    5 + +1
    6 + +1
    7 + +1
    8 = 0
    9 = 0
    10= -1
    A = -1

    It's numbers are: BC = 96.3, PE = 55.1. The IRC for a six deck game is "9" and you begin ramping up your bets at "20RC". It comes with a table of 21 index numbers individually activated by the running count.
  • BC & PE- It would seem to me that those systems that are rated
    higher did perform better over the long term(sims). Now, it may be
    true that H/L will perform better for any given session due to better
    accuracy on a given night. I would think that single or DD would be
    the best opportunity for that to happen.

    Is it related to the number of variations? KISS has 22 vs H/L 40-50.
    Can you have a better BC and win less money under the same playing
    conditions? I can see that happening in the short term, but how is it
    possible in the long term?

    If KISS scores 55 on its 22 variations and H/L scores 51 on its 50, that
    would explain the difference in $$$$$$$ due to the greater number
    of variations.
  • Ray -- The way I see it is that balanced counts do in fact implement their index numbers with greater accuracy than unbalanced counts due to the true count conversion. Unbalanced counts merely use one "compromise" index number that is accurate at some penetration point, but inaccurate elsewhere.
    Now any count system's theoretical PE is arrived at by computing its card tags in conjunction with a straightforward PE equation. But that equation doesn't know that there will be error in implementing that PE with an unbalanced count. I believe that's why I've seen that balanced and unbalanced counts with identical BC's and PE's have different long term sim performances -- with the balanced count usually outperforming the unbalanced count by about 0.05% in net EV (i.e. +.70% vs. +.75%)
  • Fred-Thanks for the reply and I do see your points. I think you touched
    on this a little in the KISS presentation (16 vs 10 and counts above 25).

    What I think I get from this experience is that count systems are measured on how well they perform their design and not some standard.
    We can't say that Red 7 at 97/53 is a tad better than H/L at 97/51.
  • Why don't I ever hear mention of High Opt I or II? Is this system considered inferior in some way?
  • Why not Hi Opt? It's a matter of very limited application. This system
    has a very high PE and is best applied in single deck games. Also, the
    BC is somewhat lower than most and is not recommended for shoe
    games. The numbers I have are 91/67.
  • I took a look at http://www.qfit.com/cvstrat.htm. It rates Hi-Lo at an ease of 6 (with lower number meaning harder), and a BC/PE/IC of .97/.51/.76.

    However, it rates KO as easier with a 7.5, and a BC/PE/IC of .98/.55/.78, beating Hi-Lo in every category and in ease of use.

    Can I trust this? Why would anyone use Hi-Lo then over KO?
  • Because the numbers are sometimes misleading. Hi-Lo is most certainly better than KO in most anything you can compare, excepting ease of use (KO has no true count conversion). In shoe games, the difference is very minor. In SD/DD games, I'd take Hi-Lo any day...
  • > Hi-Lo is most certainly better than KO in most anything you can compare

    Well, except playing efficiency, betting efficiency, and insurance. You don't have me convinced. Everytime anyone asks about certain systems over others, the response is always to look at these numbers. You seem really knowledgeable, but can you give me something more than just "trust me"? How can the numbers be misleading? I guess I don't really understand where these numbers come from really.

    Also, is KO or Hi-Low (or any other system) better than the other in terms of risk? I assume some systems have a greater or lesser risk reward?

    Thanks for your help! I've just learned a bit Hi Opt I, but before I study anything too in depth I started looking at the board and noticed no one is using it so I'm trying to find a good replacement.
  • You will notice that unbalanced count systems have a limited number of
    index numbers. The reason for this is the accuracy of some quanity of
    additional numbers is not trust worthy. As you move away from the pivot
    point the accuracy tends to be less than what one would like.

    When you include a limited number of index numbers that are close to the
    pivot point, the math is more likely to give a better overall reading as
    compared to H/L which has approx 2.5 times the number of index no's.
    This is the reason that H/L will out perform unbalanced counts in single
    or double deck. There are ways to overcome this "drift in accuracy", but
    why bother if you play shoe games where the very best balanced counts
    with an excellent PE gives very little additional advantage.

    If you play shoe games, KISS, KO or Red 7 will do the job about as well
    as most level one count systems.

    If you play single and DD, which would you rather have an unbalanced
    sys with 20 variations or a balanced system with 80 variations?
  • Whoa. That went right over my head. I think you are assuming a lot of knowledge on my part that I don't have.

    My basic question is how one system can be better than the other when PE/BE/and IC are all worse. Maybe you answered it right on and I just need to learn a lot more before I can understand the answer to my question?

    My other question, though, that I don't think you addressed, is if there is a difference between the systems in risk percentage.
  • Here's two examples, using Hi-Lo and KO. Both assume typical rules and penetration, and a starting bankroll of $3000. CVCX then calculates the optimal "unit" based on the "spread" I supplied (1-4 for SD, 1-20 for 6D).

    SD game, bet spread $10-$40, Hi-Lo hourly win rate is $47.68, RoR is 12.5%. KO Preferred is $42.74 per hour, Ror is 14.7%.

    6D game, bet spread $5-$100, Hi-Lo hourly win rate is $21.78, RoR is 16%. KO preferred is $19.21 per hour, RoR is 16.5%

    If you compare, Hi-Lo has a better hourly win rate, and a lower RoR in both cases. The difference is not huge, but even at $2 per hour for 6D and $5 per hour for SD, the difference is significant. Note that in both cases, Hi-Lo also has a lower RoR. Assuming the same bankroll and bet spread for the comparisons.

    There you have it. These are from CVCX, 2 _billion_ rounds simulated for each, so the numbers are going to be quite accurate. This shows that the stuff showing KO is has a better PE than Hi-Lo is simply flawed in some basic way, because in the real world, a balanced count is superior in all regards except for ease of use.

    I can provide more sim data if needed...

    The bottom line is buy CVCX, for under $100, and you can answer these questions to 3 decimel places, whenever you have one. How PE, BC, IC are calculated is hard to say based on some of the numbers published. But what counts is how the systems perform under _equal_ comparisons. The above gives that...
  • Now money I understand. Thanks.

    You don't want to run Hi-Opt I through the ringer for me just to satisfy my curiosities do you? :)
  • Sure. This is for the same games as before, same bank, same bet spreads. Also this assumes the ace side-count which makes hi opt I much more difficult to play accurately...

    SD, hourly win rate is $46.97, with a RoR of 11.6%. Note that Hi-Opt is good enough that even though the spread is 1-4, the dollar spread is $15-$60, but produces a similar RoR.

    6D, hourly win rate is $20.12 for a spread of $5-$100, RoR is 14%.

    Bottom line? HiOpt crushes a SD game. But for 6d it just tires out the old noggin' for no additional gain except for a headache...
  • Muchas gracias. My last thought, though -- a 1-20 spread is almost definitely going to be picked up by the casinos, correct?
  • Depends. at $5 to $100, in vegas, no. That's chump change since your average bet is going to be maybe $10. At black-chip level, it will be noticed. But if you are going to "play all" then you either spread or lose. Otherwise, you are left with back-counting and wonging in when the count rises. At crowded games, this doesn't work, of course. So it depends on when you play. At nearly empty games, you might stand out while back-counting, so there are risks all around.

    If you learn to vary your bets based on something that looks like a losing progression... or something similar, then you can stay under the radar. Or you can vary your spread and accept a lower hourly win rate. For example, in a 6D shoe, using Hi-Lo, here is what you can expect:

    spread win rate
    1-4 5.25 (flip burgers instead)
    1-8 11.81
    1-12 16.67
    1-16 20.03
    1-20 21.74

    Note that in the spreads, there are multiple ways to increase your bet according to the count, some are more aggressive and increase your win rate and variance, others reduce your win rate and variance (and hence RoR).

    You can use a smaller spread with a larger betting unit size, but that increases your bankroll requirement significantly to avoid an impossible RoR.
  • Thanks for the info. You are one knowledgeable rat.

    Will casinos really not care at al lif I am only betting the $5 to $100 spread? I got the impression that if they see any counting at all they will chuck you out no matter what.

    Also, don't forget the fun of playing. I'd rather play blackjack for $5.25 an hour than flip burgers.
  • Funckychicken-Sorry about that. Try this on for size: We are playing a
    single deck game with H/L and an unbalanced count. The unbalanced
    count has a PE of 55 and H/L is 51. The less decks you play, the more
    important is card removal in how you play your hands. The unbalanced
    system has 20 variations that are better than basic strategy and most
    of these are very close to the pivot point and accurate. H/L has the
    same 20 variations plus 30 more.

    The unbalanced count is measured on those 20 variations and so is H/L,
    along with 30 others. H/L has a greater number of variations that are
    better than basic strategy and that is where the performance difference
    plays-out.

    The net of all this is, if H/L and the unbalanced count were both measured
    on the first 20 variations, H/L would have the higher PE. We know this
    must be so because there will still be some accuracy drift in the UB count
    and none in H/L.
  • Thanks, but I am not a "particularly knowledgable rat".

    Our edge in AP blackjack is _soooo_ thin, you really have to be careful or you can be playing right into the house's pockets. It doesn't take much to whittle that 1% advantage down to zero or worse. And if you don't realize that has happened, and you sit down at such a game, you end up stuck.

    I don't try to memorize every game option (DA2, D910, D10, etc) as to how it affects my "edge" as I just 'look 'em up" on my handy laptop. I don't try to guess the best betting ramp for a particular game, I look it up with CVCX after I've finished my scouting trips and taken notes about the different 21 offerings available...

    Most people would not think about saving 1/2% on a grocery purchase because it might be too much trouble. But in 21, that 1/2% might be the difference between winning and losing for that particular game. So every .01% adds up... don't forget to add 'em up. :)
  • Well, you're the smartest rat I ever met (and I used to work in a lab, so I knew a lot of rats).

    Now that you've sold me on Hi-Lo, what's the best book that explains it? Professional Blackjack? I've read "The World's Best Blackjack Book" already, so I think I'm set with basic strategy and the likes.
  • You have two choices.

    1. Pro BJ is "the Hi-Lo bible" obviously. You can't go wrong.

    2. Mike Shackleford has a "blackjack school" that is free. A series of something like 24 lessons you grab off the web, and they take you from basic strategy, to counting drills for Hi-Lo, to basic strategy departure (index) plays, and also covers bankroll, betting ramp, how different rules affect the house edge, and so forth.

    You can find links to #2 all over the place or you can search for "blackjack school online" and print the whole thing out free...

    All you need then is something like CVBJ to drill you until you can count, do the TC conversion and bet/play properly without having to think about it. Then CVCX (I bought 'em as a "package" which is somewhat cheaper than buying each one separately, to really let you zero in on how rules and penetration and bet ramp are all inter-related... not to mention bankroll and RoR...
  • or you can develop/check your own counting/betting system with BJ777 simulator
    http://www.aceandten.com

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!