Blackjack switch expected values table question

A question about the BJ Switch Expected value tables has begun to nag me.

https://wizardofodds.com/games/blackjack/switch/appendix/1/

I'm wondering why the EV for some of the soft totals doesn’t match the hard ones when both the hard and soft will stand. Most obvious to me are the soft 20 and 19 versus their hard counterparts. With a 19 and 20, we will stand against every dealer upcard regardless of if the total is soft or hard. Hence, I would assume that the EVs would be practically the same.

I do note that there are differences in the compositions of a soft versus a hard and that affects the dealer's drawing, but it surprises me to see the differences are so pronounced. Wizard, if you would be so kind, please explain the differences. How many decks was your table based on? Is it possible this is an error, or maybe the noise of simulation?

I really appreciate these sites. Thank you for them.

The reason is because the hand composition may affect whether or not you switch.

The hand composition does affect the switch decision. However, what I'm wondering is why a hard 20 and a soft 20 are given significantly different EVs in the table. The hand will be played the same way against all upcards (stand). As I understand, the table is the value of a hand once set (after the switch decision). This would mean the values should be the same (or at least much closer).

As far as I know, Switch is always dealt from a shoe, and I’m surprised that the EVs would diverge that much due to solely compositional factors. I’m suggesting there may be an issue with the table (or at least my understanding of it).

Are you sure you're looking at the correct rows?

Most of the differences between hard 19/soft 19 and hard 20/soft 20 are around .001.

Quote: endermike

The hand composition does affect the switch decision. However, what I'm wondering is why a hard 20 and a soft 20 are given significantly different EVs in the table. The hand will be played the same way against all upcards (stand). As I understand, the table is the value of a hand once set (after the switch decision). This would mean the values should be the same (or at least much closer).

As far as I know, Switch is always dealt from a shoe, and I’m surprised that the EVs would diverge that much due to solely compositional factors. I’m suggesting there may be an issue with the table (or at least my understanding of it).

Is it really played the same way against all upcards? If I have A9 and K2, I'm switching the K and 9 most of the time, aren't I?

[reply directed at wudged]

Yes that is the difference I'm talking about. First, remember this is not .001% this is .001 in unit EV.

Second the differences vary. I have listed them below:
[column one is dealer upcard, column two is EV(19, hard)-EV(19, soft), and column three is EV(20, hard)-EV(20, soft)]

2 -0.0008 -0.0008
3 -0.0034 -0.0026
4 -0.0035 -0.0015
5 -0.0024 -0.0015
6 -0.0016 -0.0011
7 -0.0008 -0.0013
8 -0.0035 -0.0003
9 -0.0034 -0.0024
X 0.0054 0.0035
A 0.0021 -0.0038

Let’s take the largest one, 19 vs dealer 10. The difference in the shoe:

*Soft 19 = full shoe except one 8, one X, and one Ace
*Hard 19 = full shoe except one 9 and two Xs
=> difference in remaining shoe: 8 and Ace removed vs 9 and X removed

(we can skip to this since play decisions are the same, stand)

The table (from the wizard) shows that:
EV(soft 19 vs 10) = 0.0049, EV(hard 19 vs 10) = 0.0103 => difference in EV = .0054

This difference is approximately the difference between winning and losing one hand in about 400. Hmm…that sounds reasonable (based on the size of a shoe). I just looked at the table for returns based on a 6 deck shoe for standard blackjack.

https://wizardofodds.com/games/blackjack/appendix/9/6dh17r4/

The differences seem to be there as well. I guess I was underestimating the compositional effect even in large shoes. A valuable lesson I doubt I will forget.

Thanks for your responses.

[reply directed at rdw4potus]

I am speaking to the differences when the hands will be played the same way. Of course we would hit a soft 17 against an X but not a hard one? The table I cited in my first post is to be used for making the switching decision. For the table to be valid it assumes optimal strategy from there on out. Its use is best explained by the example given at the below it.

My confusion, was due to the suprising (at least to me) difference in EV when comparing soft totals to their hard counterparts when the optimal play both ways was the same for a fixed dealer upcard.

Quote: endermike

[reply directed at rdw4potus]

I am speaking to the differences when the hands will be played the same way. Of course we would hit a soft 17 against an X but not a hard one? The table I cited in my first post is to be used for making the switching decision. For the table to be valid it assumes optimal strategy from there on out. Its use is best explained by the example given at the below it.

My confusion, was due to the suprising (at least to me) difference in EV when comparing soft totals to their hard counterparts when the optimal play both ways was the same for a fixed dealer upcard.

I played the Switch simulator at Switch's home website the other day (really neat game, Switch!). It may just be that I'm misunderstanding the question, but is it possible the tables pertain to which cards can be switched? The example above, K2 v. A9, assuming the K and A are the first cards in each hand, have to stay where they are: you can only trade the 2nd cards received. As I said, it may not be pertinent, but I didn't understand this rule until I played the sim.

deleted