WE CAN LEAVE IT AT THAT.gronbog said:Agree to disagree. See the underlined part of the quote. Everyone should agree that full Kelly gives you the mathematically optimal bankroll growth. What they're trying to tell you is that, despite that, it may not be the best choice under real-life conditions. If you get backed off too many places, you can no longer play. If you can no longer play, your bankroll growth will be zero. But then on the other hand so will your risk of ruin
Bet spread
- Thread starter SlyPooch
- Start date
aceside
Active Member
This is interesting. I haven’t learned very much about Kelly bet spread but I have been an AP for a few years doing fine. I also don’t have CVCX either. Is this Kelly thing the only approach for me to turn pro?gronbog said:Everyone should agree that full Kelly gives you the mathematically optimal bankroll growth.
Don't overthink it.
Kelly betting is a mathematically optimal method to grow your bankroll. However, it comes with a rather high (~13%) risk of ruin, so it's probably not suitable for most people, especially pro's. (Yes, you can resize your top bet as bankroll changes)
It really just comes down to betting proportionally to your advantage. From there, you have to balance your risk tolerance given your bankroll, vs. hourly EV, with heat, longevity and similar considerations included. CVCX is helpful for much of this.
Kelly betting is a mathematically optimal method to grow your bankroll. However, it comes with a rather high (~13%) risk of ruin, so it's probably not suitable for most people, especially pro's. (Yes, you can resize your top bet as bankroll changes)
It really just comes down to betting proportionally to your advantage. From there, you have to balance your risk tolerance given your bankroll, vs. hourly EV, with heat, longevity and similar considerations included. CVCX is helpful for much of this.
No human can recalculate full Kelly every hand. At least I have met no math genius so far, who has tried or says they do. If you are that "unicorn", so much power to you. BUT if you have NEVER been backed off in a casino, then you have not played enough to have personally experienced why this is not possible. And to be honest, typing IN ALL CAPS IS CONSIDERED RUDE! So be more circumspect.moraine said:
Thanks for pointing out "Both are important." I totally agree. And that was also implied when I wrote: "Substance, not style, is more important."DSchles said:
PS -- I did not address the members' earlier points on "unicorn, etc.," it is NOT BECAUSE I disagreed with those points or thought they had "missed the point." It is because I thought the discussions should stay more focused. Sorry for the neglect.
Last edited:
aceside
Active Member
Let me recap. If at the HiLo TC of +2, I have a player’s advantage of 1% in blackjack and my bankroll is $10000, then I should bet 0.77*1%*$10000=$77. I recall that a couple of months ago I had this discussion with Don S at blackjacktheforum, and the conclusion was that my bet amount also needed to be inversely proportional to the variance at that moment. How does variance come into play then?moraine said:
- Terrible Misunderstanding here (which might have been caused by my use of "The player's POSITIVE EXPECTED VALUE" in the formula I gave.
- At Hi-Lo TC +2, any player has only about +0.5% EV, NOT +1% EV, under common S17 blackjack rules. (The card counting player's skill level is irrelevant AS LONG AS THE PLAYER IS USING THE BASIC STRATEGY.)
- One Kelly Bet for a $10,000 bankroll = 0.77 x 0.005 x $10,000 = $38.5 or $35 for convenience (betting less is better than over-betting).
- The 0.77 coefficient takes care of the ordinary "variance" of blackjack games under ordinary blackjack rules.
Last edited:
aceside
Active Member
This is way too reserved. First of all, I treat the TC of +2 as TC>=+2. Second, based on the the true count advantage graph on blackjackincolor.com, the player’s EV increases about 0.65% per true count when TC>=+1 for 6-deck stand-17 games. I am a little more aggressive, using $77 at TC >=+2.moraine said:
KewlJ
Well-Known Member
So where, or maybe more accurately WHEN are you playing that S17 is a common blackjack rule for shoe blackjack games that you are talking about. I see S17, but not at the limits you are speaking of, usually either $50 or 100 minimum.moraine said:
Next, you are figuring the advantage gained (from the house edge starting point) or IHA, wrong for each TC. You are figuring IHA .5% with a gain of roughly .5% for each TC, which gets you to your +.5% at exactly TC of +2. Problem is the true counts are more like a bucket than exact number and you are not accounting for the higher parts to the bucket.
So lets take your 6 deck game, 1 deck played. Running count +10 = TC +2. What about running count +12, +13, +14? They all fall into the TC +2 betting bucket, but the advantages are greater than .5%. So you have to figure an average advantage for that entire TC +2 bucket and it will be higher than what you are figuring for the lowest point in the bucket (exactly TC +2). In short, you are underestimating your advantage at each TC.
I have a couple more thoughts on your comments on different bet amounts, but I will save that for later. I will just tease that you don't want to have too many different bet amounts for each slightly different advantage. That really is a big give-away. In this day and age, 2021, there is not a whole lot new about HOW to gain an advantage counting cards. That has all been done and laid out for you by people like Don S. The trick today is to get away with it. Just have 4-5 at most, different bet amounts, The most important being max bet and minimum bet.
Edit: Aceside touched on some of what I said about the buckets, while I was typing.
- Last time when I was in Pennsylvania (about 2 years ago), I found even a $5 or $10 table was a S17 table.
- The relationship between the change in TCs and EVs may not be linear, but a ball-park average number is good enough for all practical purpose. I believe an AVERAGE OF 0.5% EV per each Hi-Lo true count has been generally accepted by Hi-Lo users (not including me because I don't use Hi-Lo for shoe games).
- Good luck, good luck, if you want to over bet. REMEMBER KELLY CRITERION !!!
Just look at any chart of the hundreds in Chapter 10. Pick your game conditions. For most, +2 <= TC < +3 edge is surely greater than 0.5% and in S17, DAS, LS games exceeds 1.00%. We're talking about card counters, and indices add to the edge. Also, the method for calculating TC matters. I use flooring, which as Kewl mentions, implies an entire interval from 2 to 3.
Don
Don
KewlJ
Well-Known Member
So you have mentioned Kelly Criteria, or Kelly betting or the formula for Kelly betting in almost every post. I am guessing you just fairly recently learned about Kelly and are hell bent on showing us all how smart you are.moraine said:
If I had to guess, I would say you read a book or two from about 40 years ago. Maybe Wong's Professional Blackjack by use of your benchmark $10,000 bankroll. But that is just a guess. What we do know is it wasn't Don Schlesinger's BJA3.
2 things you might consider. Some of the things you are reading are outdated for todays conditions and games. And second, on this forum, in particular, you are mostly talking to some pretty experienced and successful players. While we welcome your participation, it really isn't productive to come on here, challenging everyone because you recently learned a little something about Kelly Betting.
Let me repeat, winning blackjack via card counting today, is not about any of the stuff you read in a book. That is all elementary, despite what you think. Nothing new. The real key is applying it in a way that allows you to achieve some kind of longevity and be welcome to keep playing.
- I mentioned K**** criterion because of the topic of bet-spread, which two topics should be inter-related.
- I used a $10,000 bankroll as an example, because IT WAS THE ROUND NUMBER ORIGINALLY USED BY member "aceside" in an earlier post.
- Again, I always like to restrict my discussions to the "SUBSTANCE" of he topic being discussed, but would like to express my sincere appreciation of whatever comments, including the GUESSES, that you may have.
moraine, ok you want to talk about substance let's start with two posts by you:
"Whatever you try to accomplish by card counting, A SELF-IMPOSED BET BREAD WOULD NOT HELP IN SHOE GAMES, EXCEPT HELP SETTING YOURSELF UP AS A LOSER WITH A PREDICTABLE RISK OF RUIN. If you still cannot see the point, Good Luck, Good Luck. "
" BUT MY POINT IS that a pre-selected bet sp is neither needed nor helpful in shoe games. USING KELLY BET IS BETTER UNDER ANY CIRCUMSTANCES. "
I will repeat what I said earlier:
You have a lot to learn.
"Whatever you try to accomplish by card counting, A SELF-IMPOSED BET BREAD WOULD NOT HELP IN SHOE GAMES, EXCEPT HELP SETTING YOURSELF UP AS A LOSER WITH A PREDICTABLE RISK OF RUIN. If you still cannot see the point, Good Luck, Good Luck. "
" BUT MY POINT IS that a pre-selected bet sp is neither needed nor helpful in shoe games. USING KELLY BET IS BETTER UNDER ANY CIRCUMSTANCES. "
I will repeat what I said earlier:
You have a lot to learn.