Why does Bloch subtract 1?

**counter19** · 07-07-2014, 10:01 PM

Andy Bloch on his video says that after calculating the true count (using Hi-Lo), he subtracts 1, then plays two hands. So if the true count is 5, he will play 2 hands with 4 units on each (he was playing 6 deck in the example). Why does he subtract 1? Also, since he has two hands out there, he really has 8 units at risk. What is the thinking here?
I noticed that his minimum bet is 1/1000 of his bankroll. This sounds like a very conservative calculation for a minimum bet to me, which could be why he can justify risking 8 units in this example.

**seriousplayer** · 07-07-2014, 10:40 PM

Originally Posted by counter19

Andy Bloch on his video says that after calculating the true count (using Hi-Lo), he subtracts 1, then plays two hands. So if the true count is 5, he will play 2 hands with 4 units on each (he was playing 6 deck in the example). Why does he subtract 1? Also, since he has two hands out there, he really has 8 units at risk. What is the thinking here?
I noticed that his minimum bet is 1/1000 of his bankroll. This sounds like a very conservative calculation for a minimum bet to me, which could be why he can justify risking 8 units in this example.

How the MIT version of Hi-lo works is they convert the running count to true count then they subtract the TC from the "true offset". They defined the true offset as 2 x the players disadvantage. For example, a typical disadvantage for a prefect basic strategy player is .5%. To get the true offset you multiply .5 x 2 = 1. Then he subtract 1 from the TC. If a blackjack game has a player disadvantage of .25% (6 decks, S17, DAS, LS, RSA) then the true offset will be .25 x 2 = .5.

The MIT bet units are determine using the Kelly criterion. They usually set $50 as half of unit and $100 as one unit. They take their entire bankroll and divide into 1/4 of a Kelly. If the true count is 5 their bet basis of fthe Kelly criterion if setting $50 as half of unit will be $400.

**bigplayer** · 07-08-2014, 07:06 AM

Each TC is worth 0.5% but the game has an 0.5% edge against you. So you bet the TC -1 to account for the house advantage.

**bigplayer** · 07-08-2014, 07:08 AM

Originally Posted by counter19

I noticed that his minimum bet is 1/1000 of his bankroll. This sounds like a very conservative calculation for a minimum bet to me, which could be why he can justify risking 8 units in this example.

For any player, but particularly the big teams, the minimum bet is always as low as you can get away with. If you play shoes, your minimum bet is actually $0 considering you exit the table when the count drops below a certain point. You base your Risk of Ruin and Bankroll on your Kelly betting unit not your minimum bet.

**counter19** · 07-08-2014, 08:44 AM

Thanks guys. This all makes a lot of sense.

**TeepsConPeeps** · 07-08-2014, 10:48 AM

Originally Posted by bigplayer

Each TC is worth 0.5% but the game has an 0.5% edge against you. So you bet the TC -1 to account for the house advantage.

This doesn't mean at tc1 you're not upping your bet, does it? According to CVCX in any reasonable game the player has an advantage at tc1, so obviously you would up your bet, right?

**code red** · 07-08-2014, 10:52 AM

Yeah I would expect you do whatever CVCX tells you to

**bigplayer** · 07-08-2014, 12:58 PM

Originally Posted by TeepsConPeeps

According to CVCX in any reasonable game the player has an advantage at tc1, so obviously you would up your bet, right?

CVCX gives an edge at +1 because that bucket includes all counts from +1.0 to +1.99. While most players floor or tuncate to make their index decision, I think most players tend to calculate an exact TC (or have an idea when they're at +1, +1.5, +2, etc) and thus if the TC is +2.5 you'd make an in-between wage Most players should just know that their betting unit is two hands of $xxx for every 0.5% worth of advantage so to figure out what to bet you just take the TC x 0.5% and subtract out the house advantage, which is most easily accounted for by just calling it 0.5%. If your house edge lower such as a classic strip rules shoe (0.26%) you'd adjust your betting accordingly by using a bigger betting unit and making a corresponding half unit bet at +1.

For example your betting unit is 2@$200 or 1@300. So your ramp would be something like

+1 or less minimum
+1.5 2@$100
+2 2@$200
+3 2@$400
+4 2@$600
+5 2@$800

Where you stop is up to the team but some teams would just continue to raise their bets at higher and higher True Counts until infinity (or table max). Some might cut the ramp off at +5 or +6. It all depends on how much spread you want to show. Using a quick method like TC-1 is the KISS (Keep It Simple Stupid) principle in action...and it works.

**DSchles** · 07-08-2014, 06:52 PM

I've always had a bit of a problem with blindly using the -1 offset, because, well, it turns out that, quite often, it's just not right!

Quick: 4.5/6, S17, DAS (I won't even throw in surrender or RSA, both common in Las Vegas). Hi-Lo, I18.

1. What's your edge at TC = +4?
2. At TC = +5?

Answers: 1) 1.97%. 2) 2.56%. What happened to the offset?

Do you see the problem?

Don

**seriousplayer** · 07-08-2014, 07:45 PM

Originally Posted by DSchles

I've always had a bit of a problem with blindly using the -1 offset, because, well, it turns out that, quite often, it's just not right!

Quick: 4.5/6, S17, DAS (I won't even throw in surrender or RSA, both common in Las Vegas). Hi-Lo, I18.

1. What's your edge at TC = +4?
2. At TC = +5?

Answers: 1) 1.97%. 2) 2.56%. What happened to the offset?

Do you see the problem?

Don

Don,

Could you kindly explain how you got the answer?

**bigplayer** · 07-09-2014, 08:28 AM

Originally Posted by DSchles

I've always had a bit of a problem with blindly using the -1 offset, because, well, it turns out that, quite often, it's just not right!

Quick: 4.5/6, S17, DAS (I won't even throw in surrender or RSA, both common in Las Vegas). Hi-Lo, I18.

1. What's your edge at TC = +4?
2. At TC = +5?

Answers: 1) 1.97%. 2) 2.56%. What happened to the offset?

Do you see the problem?

Don

Of course the edge does not rise and fall in a perfectly linear fashion, but unless you plan to memorize 20 different betting ramps for 20 different games and rules sets you need something a bit more "user-friendly".

Another option is to just know your max and half-max bets and at what point you're gonna make them and then just interpolate in between. What really matters is getting the money on the money on the table within your betting constraints when called for. I think Bloc's group just needed an objective method for putting out the money that could be easily quality controlled that was close enough. A stylized inexact betting method is fine as long as it does a fairly good job of reflecting reality.

**Nyne** · 07-10-2014, 08:29 AM

To Don's point, I found that for most of the games I've played, the difference in edge per TC is closer to 0.6% than 0.5% within the range I'm concerned with, so I calculated my unit bet per 0.6% advantage, then applied the appropriate offset. Inputting this betting strategy into CVCX shows virtually no difference in SCORE from what CVCX calculates as optimal betting.

**seriousplayer** · 07-10-2014, 09:23 AM

Originally Posted by Nyne

To Don's point, I found that for most of the games I've played, the difference in edge per TC is closer to 0.6% than 0.5% within the range I'm concerned with, so I calculated my unit bet per 0.6% advantage, then applied the appropriate offset. Inputting this betting strategy into CVCX shows virtually no difference in SCORE from what CVCX calculates as optimal betting.

What about the N0 and risk of ruin? Is it higher or lower than what CVCX calculates?