MJ: Question for Don and Norm--MIT Blackjack Team post-trip procedures

[MJ]

During the 90s when the MIT Blackjack Teams reigned supreme, the spotters were required to keep meticulous records of what transpired every time the BP was called into a shoe.

In particular, the spotters were responsible for recording the following 4 variables for each and every round the BP played:

1) Running count
2) # Players at table
3) How many decks remaining (nearest 1/4)
4) Location of cut card

At the conclusion of the trip, the spotters would submit the player sheets to the team manager who would then feed the aforementioned data into a custom simulator (along with the unit size) which would then spit out the CE, EV, and SD for each round. BPs and Spotters were compensated based upon a fraction of the CE which was generated, irrespective of whether the team won or lost for the trip.

With the current simulators out on the market, we tend to just estimate EV, SD, and CE based upon a Time factor. For example, if a player backcounts for 2 hours where 100 rounds are dealt per hour, assuming the % of hands played is 20%, then the simulator assumes he played 40 rounds and calculates EV accordingly. But, that is just an ESTIMATE of EV based upon what SHOULD have happened. What if I only played 20 rounds over the course of 2 hours because I witnessed unusually low counts?

I guess what I am getting trying to ask is whenever a player conducts a post-trip evaluation using some of the popular commercial software, rather then simulate what was SUPPOSE to happen, why not simulate what actually DID happen?

Do either of you see any advantage to using MITs method of post-trip evaluation? To me, it seems like what MIT is doing will give a more accurate estimate of the value of the trip. Additionally, it also has the benefit of not having to estimate how many rounds are dealt per hour. On the downside, it must be a real nuisance for the spotter to have to remember and record all that information.

On a side note, I believe CVData does permit the user to calculate EV based upon hands played using the Customize button. But, I don't think SD can be adjusted for Hands Played, even though the option is there.

Thanks,
MJ

[Norm Wattenberger]

CVData does not 'assume' a set hands per hour. You set it. It has always been my contention that the best estimate of your hands/hour is based on your experience - not a calculation.

[Don Schlesinger]

> In particular, the spotters were responsible for
> recording the following 4 variables for each and every
> round the BP played:

> 1) Running count
> 2) # Players at table
> 3) How many decks remaining (nearest 1/4)
> 4) Location of cut card

Awful lot of work for little reason.

> At the conclusion of the trip, the spotters would
> submit the player sheets to the team manager who would
> then feed the aforementioned data into a custom
> simulator (along with the unit size) which would then
> spit out the CE, EV, and SD for each round. BPs and
> Spotters were compensated based upon a fraction of the
> CE which was generated, irrespective of whether the
> team won or lost for the trip.

You already know my thoughts on this.

> With the current simulators out on the market, we tend
> to just estimate EV, SD, and CE based upon a Time
> factor. For example, if a player backcounts for 2
> hours where 100 rounds are dealt per hour, assuming
> the % of hands played is 20%, then the simulator
> assumes he played 40 rounds and calculates EV
> accordingly. But, that is just an ESTIMATE of EV based
> upon what SHOULD have happened.

I suppose.

> What if I only played
> 20 rounds over the course of 2 hours because I
> witnessed unusually low counts?

The point being? What if you played 60 rounds, because there were tons of high counts? It evens out in the long run, which is why you use the simulated result and not the actual result.

> I guess what I am getting trying to ask is whenever a
> player conducts a post-trip evaluation using some of
> the popular commercial software, rather then simulate
> what was SUPPOSE (sic) to happen, why not simulate what
> actually DID happen?

Because then, it's not a simulation! Since you already know exactly what happened, why not just use the actual results themselves? Either you simulate what will happen on average, or you report exactly what actually took place. "Simulating what actually DID happen" makes no sense to me.

> Do either of you see any advantage to using MITs
> method of post-trip evaluation?

Yes, see above. That is, if you believe in this method of compensation to begin with, which, of course, I don't.

> To me, it seems like
> what MIT is doing will give a more accurate estimate
> of the value of the trip.

It's not an estimate, if you already know what happened! If you're going to use the actual events at the table, why stop at estimating an e.v.? Why not just use what happened? The whole thing sounds pretty stupid to me.

> Additionally, it also has
> the benefit of not having to estimate how many rounds
> are dealt per hour. On the downside, it must be a real
> nuisance for the spotter to have to remember and
> record all that information.

That's what he got paid to do!

> On a side note, I believe CVData does permit the user
> to calculate EV based upon hands played using the
> Customize button. But, I don't think SD can be
> adjusted for Hands Played, even though the option is
> there.

It's trivial. I don't see the problem.

Don

[AutomaticMonkey]

> ...In particular, the spotters were responsible for
> recording the following 4 variables for each and every
> round the BP played:

> 1) Running count
> 2) # Players at table
> 3) How many decks remaining (nearest 1/4)
> 4) Location of cut card

> At the conclusion of the trip, the spotters would
> submit the player sheets to the team manager who would
> then feed the aforementioned data into a custom
> simulator (along with the unit size) which would then
> spit out the CE, EV, and SD for each round. BPs and
> Spotters were compensated based upon a fraction of the
> CE which was generated, irrespective of whether the
> team won or lost for the trip....

Sounds to me like a recipe for getting screwed.

1) It forces the spotters to play a game of chance against one another. Supposedly when you're a paid spotter the idea is to make sure money, not take your chances. What happens if you happen to not get good counts? A team should know what the EV is before entering the casino and if pay is agreed to be a function of EV, then pay based on EV, and not assign this busywork to players who are supposed to be doing important things like counting cards.

2) It allows the team to screw the spotters by simply "hiring" more spotters. The can double the number of spotters but not the BP's. This allows the BP's to take only the cream of the calls, and being spotters only get paid if the BP plays, the BP's will be making more money while the spotters make less.

3) It gives the spotters an incentive to screw the team by calling in a BP when the pay will be highest for the spotter, with or without there actually being an advantage. Why would they care, especially if they believe they are being screwed by 2)?

[MJ]

> The point being? What if you played 60 rounds, because
> there were tons of high counts? It evens out in the
> long run, which is why you use the simulated result
> and not the actual result.

I concede that it would even out in the long run. But the problem is that using a simulated result may not give the best indication for how much the actual play is worth. Not all tables have equal penetration, # of players, etc.

The drawback to using the simulated result is that it is essentially a "one size fits all" approach. It does not take into account DYNAMIC conditions, which vary from table to table.

Conventional simulators take a generic approach to simulation. They account for a fixed # players at the table, a fixed # decks cut off, an average bet based upon a TC frequency distribution, an average # decks remaining, etc.

MIT figured out the precise value of their play post-trip, based upon a CASE BY CASE basis which provides the EXACT conditions under which the rounds were played.

If you were trying to determine the THEORETICAL value of your play after a weekend trip of wonging, would you prefer to use a generic approach or a case by case approach, which examines precisely what occurred for each round?

> Because then, it's not a simulation! Since you already
> know exactly what happened, why not just use the
> actual results themselves?

The point is to determine the value of your play. The actual results themselves do not tell us what the CE, EV, and SD are for each round. The only way to determine that information is to plug in the variables for each round and then simulate each round on a case by case basis.

> will happen on average, or you report exactly what
> actually took place. "Simulating what actually
> DID happen" makes no sense to me.

See above. They reported what took place and then ran simulations on it for compensations purposes as well as to see how far off the mark their actual results were.

Let us look at two examples:

Scenario A:
# Players: 4
Cut card: 1.50 decks cut-off
RC: +12
# Decks remaining: 2

Scenario B:
# Players: 2
Cut card: 1.0 deck cut-off
RC: +7
# Decks remaining: 1.25

Do you think the above scenarios have the same EV, SD, and CE for a BP that enters mid-shoe? Do you think that a "one size fits all" simulation would produce an accurate estimate for these stats?

At best, one might invoke the ComboSim feature on CVData to take into account varying # players and penetration levels. But no where can the user specify the RC and # decks remaining to simulate and thereby determine the value of any particular round.

MJ

[Don Schlesinger]

You're acting as if I don't understand the concept. You really don't have to explain it to me. You asked what I thought of it, and I told you.

I am unalterably opposed, for team purposes, to paying someone more because, tonight, he got "lucky" and two players got up from the table and left him alone. Or, because the count got good tonight, while it got bad for a teammate. In fact, I despise the concept! All it does is pit teammate against teammate and foster resentment that someone got paid more for circumstances that were probably completely beyond his control.

Please don't write back to try to convince me of what a genial idea this was on the part of MIT. It isn't. Neither is paying people salaries with investors' money, if the team is losing.

Do me a favor: don't get me started on any of this, please.

Don

[GMan]

> Spotters were compensated based upon a fraction of the
> CE which was generated, irrespective of whether the team won or lost for the trip.

25% of generated CE

> Sounds to me like a recipe for getting screwed.

> 1) It forces the spotters to play a game of chance
> against one another.

Nope, total trip CE was being split among all spotters and BPs.However, it was forcing everybody to shop for good conditions !

> 2) It allows the team to screw the spotters by simply
> "hiring" more spotters. The can double the
> number of spotters but not the BP's. This allows the
> BP's to take only the cream of the calls, and being
> spotters only get paid if the BP plays, the BP's will
> be making more money while the spotters make less.

All players were investors?You think these guys were stupid enough to screw themselves ?
This make you point no.3 obselete?