House edge for common (poor) strategy

I'm having an argument with a coworker regarding the house edge of Blackjack. My coworker keeps talking about the low house edge of the game, and I'm arguing that the house edge is actually in practice much higher given the strategy people play.

Our rules, for context, are as follows:

-- 6-deck blackjack
-- Dealer hits soft 17
-- Double after split is allowed
-- Dealer checks for blackjack
-- Surrender to any 10 (before dealer checks for blackjack)
-- Surrender to any ace AFTER dealer checks for blackjack
-- Only one card may be taken when splitting aces (a subsequent ace can be split again)

Perfect strategy is easy enough to look up for such a game, but I've been unable to find house edge for common strategy in this game.

Local players typically:

-- Never double vs a ten
-- Do not surrender
-- Do not hit 12 against a 2 or 3
-- Double 9 vs a 2
-- Always stand on soft 18
-- Double 8 against a 4, 5 or 6 (even with two fours)
-- Do not split 9's
-- Do not split 8's vs a 9, 10 or A
-- Often do not hit 16 regardless of dealer card
-- Often do not hit 15 vs a 7

I've been unable to find any information on what the house edge is when people employ these strategies. Given such personal rules, how much can people expect to lose?

Some of your questions can be answered here: for any of these, you'll have to pick the rule change they're not following, and add or subtract that amount from your optimal strategy calculation (that you should plug in to the wizard's HE calculator to determine the HE of your game).

Most of the rest you can pull out of these tables, which list the relative value of standing, hitting, doubling, or splitting any given hand. You'll have to do some math, with what they're doing in one column, vs. the correct strategy for each play in the other, and figure out how much they're giving up in each case.

Late surrender to dealer's ace, early surrender to all other up-cards? Where is this game?

Quote: Elastoid

Local players typically:

"Typically" is not sufficient to do this analysis. You need to provide an exact list of strategy deviations. Every single one. Also, you must assume that players make this mistake every time the hand occurs.

That said, I have done this type of work before. Bill Zender provided me with a list of 9000 hands he observed and the mistakes made. We grouped the mistakes into highly likely, common and rare. I then ran three simulations, assuming a player made the group of errors. What we found is that the usual edges that are used by casinos to evaluate players significantly overstate the house edge. Not taking Surrender into account, so that the game starts at 0.61%, the player who makes the common errors plays with a house edge of about 1.00% to 1.25%. The list you provide looks like it consists of the "common" errors ... I just double checked. Yes. We got 1.19% based on 1B hands.


Here are the common deviations we assumed to get 1.19%:

9 doubles only against 5,6, otherwise hit
T doubles against 2-8, otherwise hit
11 doubles against 2-8, otherwise hit
12 stands against 2, 3
15 stands against T
16 stands against T
A,2, A,3, A,4, A,5, always hit
A,6 doubles only against 5,6, otherwise hit
A,7, A,8, A,9 always stand
don't split 2,2, 3,3, 4,4, 6,6, 7,7 or 9,9
double 5,5 against 2-8, otherwise hit
Always split 8,8 and A,A

Quote: DiceDealer

Late surrender to dealer's ace, early surrender to all other up-cards? Where is this game?

Normally I'd avoid stating which casino I work at, but... I only work here for another 4 days before I switch casinos (I put in my two weeks almost two weeks ago). I work at the River Cree Casino in Enoch, Alberta (near Edmonton).

That said, dealer hits on soft 17, so it's not the best game.

Quote: teliot

"Typically" is not sufficient to do this analysis. You need to provide an exact list of strategy deviations. Every single one. Also, you must assume that players make this mistake every time the hand occurs.

What I mean at "typically" is that most players follow that strategy. I don't want to categorically say that all local players play poorly -- just most of them.

Quote: teliot

That said, I have done this type of work before. Bill Zender provided me with a list of 9000 hands he observed and the mistakes made. We grouped the mistakes into highly likely, common and rare. I then ran three simulations, assuming a player made the group of errors. What we found is that the usual edges that are used by casinos to evaluate players significantly overstate the house edge. Not taking Surrender into account, so that the game starts at 0.61%, the player who makes the common errors plays with a house edge of about 1.00% to 1.25%. The list you provide looks like it consists of the "common" errors ... I just double checked. Yes. We got 1.19% based on 1B hands.


Here are the common deviations we assumed to get 1.19%:

9 doubles only against 5,6, otherwise hit
T doubles against 2-8, otherwise hit
11 doubles against 2-8, otherwise hit
12 stands against 2, 3
15 stands against T
16 stands against T
A,2, A,3, A,4, A,5, always hit
A,6 doubles only against 5,6, otherwise hit
A,7, A,8, A,9 always stand
don't split 2,2, 3,3, 4,4, 6,6, 7,7 or 9,9
double 5,5 against 2-8, otherwise hit
Always split 8,8 and A,A

Really, so low? From this page https://wizardofodds.com/games/blackjack/basics/#toc-BadStrategies I assumed that the house edge would increase more significantly when poor strategy is employed.

Quote: Elastoid

Really, so low? From this page https://wizardofodds.com/games/blackjack/basics/#toc-BadStrategies I assumed that the house edge would increase more significantly when poor strategy is employed.

Yes. Most people overestimate the house edge that comes from weak play.

Quote: beachbumbabs

Most of the rest you can pull out of these tables, which list the relative value of standing, hitting, doubling, or splitting any given hand. You'll have to do some math, with what they're doing in one column, vs. the correct strategy for each play in the other, and figure out how much they're giving up in each case.

Those tables assume dealer stands on soft 17. Any place I can find those tables, but for dealer hits on soft 17?