# Thread: MJ: Real world play vs simulated play question

1. ## MJ: Real world play vs simulated play question

I am the type of player that likes to know how my performance at the tables compares to simulated win rates with CVCX.

One assumption of BJ simulators is that the number of players at the table remain constant for play all. Suppose in a game with 4 players, a ploppy leaves in the middle of the shoe. Wouldn't this player's departure alter the TC frequencies for the shoe and make them different from the calculations which the simulation was based on?

Same question applies when another player enters the shoe at random. Again, the TC frequencies would be altered.

This type of random player entry/departure occurs in at least 15% of shoes. I guess my concern is whether or not the simulated WR, variance, SCORE, etc are still close to accurate despite these unavoidable, real world factors.

Another question with respect to this topic is regarding penetration for the simulation. Most players select one level of penetration which they feel is a close approximation to their game, and then use a bet schedule based upon that level of pen. But, what happens if the estimate of pen is off by 1/4 deck? Now, the player would be using a suboptimal bet ramp in their game. Is this anything to be concerned about?

At the end of the day, I guess one should realize that simulations are just estimates of real world play. I mean there are just sooo many variables that can be slightly different than that which was simulated when one sits down at a BJ table.

MJ

2. ## Norm Wattenberger: Actually

CVData does not make these assumptions if you don't want them made. You can set it to vary both the number of players and the penetration during a sim. You can also combine multiple sims with different parameters. It will merge the thousands of stats and recalculate DI, std. dev., SCORE, etc.

Serious Blackjack Software

> I am the type of player that likes to know how my
> performance at the tables compares to simulated win
> rates with CVCX.

> One assumption of BJ simulators is that the number of
> players at the table remain constant for play all.
> Suppose in a game with 4 players, a ploppy leaves in
> the middle of the shoe. Wouldn't this player's
> departure alter the TC frequencies for the shoe and
> make them different from the calculations which the
> simulation was based on?

> Same question applies when another player enters the
> shoe at random. Again, the TC frequencies would be
> altered.

> This type of random player entry/departure occurs in
> at least 15% of shoes. I guess my concern is whether
> or not the simulated WR, variance, SCORE, etc are
> still close to accurate despite these unavoidable,
> real world factors.

> Another question with respect to this topic is
> regarding penetration for the simulation. Most players
> select one level of penetration which they feel is a
> close approximation to their game, and then use a bet
> schedule based upon that level of pen. But, what
> happens if the estimate of pen is off by 1/4 deck?
> Now, the player would be using a suboptimal bet ramp
> in their game. Is this anything to be concerned about?

> At the end of the day, I guess one should realize that
> simulations are just estimates of real world play. I
> mean there are just sooo many variables that can be
> slightly different than that which was simulated when
> one sits down at a BJ table.

> MJ

3. ## Don Schlesinger: Re: Actually

> CVData does not make these assumptions if you don't
> want them made. You can set it to vary both the number
> of players and the penetration during a sim. You can
> also combine multiple sims with different parameters.
> It will merge the thousands of stats and recalculate
> DI, std. dev., SCORE, etc.

That's a wonderful feature, but, as MJ points out, when all is said and done, no one can know exactly what situation will develop during play, and I doubt that it matters very much.

You set a sim for four players and 75% pen. During your session, someone arrives, and your hourly win rate drops. Then two people leave, and your win rate increases. Next shoe, a new dealer cuts a little worse. Later, someone else cuts even better than 75%.

Bottom line: I wouldn't drive myself crazy over this. It all evens out.

Don

4. ## MJ: Re: Actually

> You set a sim for four players and 75% pen. During
> rate drops. Then two people leave, and your win rate
> increases. Next shoe, a new dealer cuts a little
> worse. Later, someone else cuts even better than 75%.

When that extra player arrives in the middle of the shoe, does the WR/hand now decrease for the counter or is it still the same as that prior to the ploppy's arrival? Clearly, WR/Hr will fluctuate as more/fewer rounds are dealt, but that is not what I am seeking clarification on.

One thing I noticed on CVCX/play-all is that the SCORES are pretty much identical across the board regardless of how many players are at the table for any given counting system. But again, this assumes no players enter/depart mid-shoe.

Is it better to be conservative or aggressive in the estimate of penetration level (assuming CVCX)? If we are conservative, then that would mean we under bet. If we are aggressive, that would mean we over bet. I am inclined to go with conservative, which means that we would probably be playing to a lower ROR and lower SCORE.

Actually, SCORE would be lower regardless if you estimate pen conservatively or liberally, due to the fact that it would be suboptimal as the pen level was not exactly that used in the casino, which means the bet schedule is not quite optimal in real world play. But, as you stated, it probably does not make a difference. However, if the estimate of pen is off by half a deck, then that probably will make some difference.

5. ## Don Schlesinger: Re: Actually

> When that extra player arrives in the middle of the
> shoe, does the WR/hand now decrease for the counter or
> is it still the same as that prior to the ploppy's
> arrival? Clearly, WR/Hr will fluctuate as more/fewer
> rounds are dealt, but that is not what I am seeking
> clarification on.

The win rate doesn't change enough to discuss. I mentioned this in BJA3, p. 205, second full paragraph.

> One thing I noticed on CVCX/play-all is that the
> SCORES are pretty much identical across the board
> regardless of how many players are at the table for
> any given counting system. But again, this assumes no
> players enter/depart mid-shoe.

If you really meant to write SCORE, above, and not "win rate," then of course they will be the same, because SCORE assumes four players at the table, and, no matter how many players YOU decide the to put there, the SCORE isn't going to change.

> Is it better to be conservative or aggressive in the
> estimate of penetration level (assuming CVCX)? If we
> are conservative, then that would mean we under bet.
> If we are aggressive, that would mean we over bet. I
> am inclined to go with conservative, which means that
> we would probably be playing to a lower ROR and lower
> SCORE.

To each his own.

> Actually, SCORE would be lower regardless if you
> estimate pen conservatively or liberally, due to the
> fact that it would be suboptimal as the pen level was
> not exactly that used in the casino, which means the
> bet schedule is not quite optimal in real world play.
> But, as you stated, it probably does not make a
> difference.

Right.

> However, if the estimate of pen is off by
> half a deck, then that probably will make some
> difference.

Half a deck is a lot.

Don

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•