Nice analysis as usual MathProf.

If you use the Schlesinger suggestion for "unit", then 100 unit losses are not that unusal for me. If, on the other hand, you get into a game with a $1 minimum and $200 maximum and consider $1 your unit---the loss can be in the thousands! How do you define "unit"? Maybe your average bet size?

In the days of Ken Uston, a "unit" was just the minimum bet at the tables where you played, or the minimum bet you would consider making, consistent with your overall cover strategy. This is what revereman apparantly means when he says "unit."

Today we have Optimal Betting Theory (OBT), and the way it works is you input a maximum to minimum bet ratio such as 4 to 1, or 12 to 1, or whatever else you want, and then OBT tells you exactly what all your bets must be for a given Risk of Ruin to a given Bankroll.

For example, OBT could tell you to bet $12.37, at TC less than or equal to +1, $17.97 at +2, $23.35 at +3, $39.44 at +4 and $49.48 at +5 and above. $12.37 would then be called your "unit" because it corresponds with the 1 in your 4 to 1 bet ratio. (Note $12.37 x 4 = $49.48) This is just an example, and the numbers will be wildly different, depending on the game, bankroll, desired RoR, etc.

But this is what math heads typically mean when they say "unit" nowadays, because "Hey! It's a cool theory!" -- even when it has no direct bearing on bets actually played (which tend to be multiples of $5, for obvious reasons). Neither type of unit relates very well to RoR or standard deviation without a lot more information.

ETF

A unit should be defined as your minimum bet which in most cases will be also the table minimum, anything else is an abuse of language.Calling $12.37 a unit sounds utterly ridiculous. Not only won't you find a casino who would sell you chips with a value of $12.37, you cannot even bet such an amount.In most casinos, the bets have to be multiples of the table minimum.
For purely theoretical considerations you might want to define unit in the way you described but this has no bearing on actual casino play.
To answer the question of the original poster, I have already had losing sessins of 2400 units (unit= 5 Euros).Expressed in a different way it sounds far less dramatic :20 max bets (3*200).

Francis Salmon

"A unit should be defined as your minimum bet which in most cases will be also the table minimum, anything else is an abuse of language."

So BJRM abuses the language every time you do a One Second sim? All the sims in Chapter 10 of BJA3 are abusive!? You'd better alert Don S!

"Calling $12.37 a unit sounds utterly ridiculous. Not only won't you find a casino who would sell you chips with a value of $12.37, you cannot even bet such an amount.In most casinos, the bets have to be multiples of the table minimum."

Wow, that's certainly a new one on me. Do you perhaps mean that bets must be multiples of $1?

"For purely theoretical considerations you might want to define unit in the way you described but this has no bearing on actual casino play."

It does if you want to use Optimal Betting Theory, which is what most professionals are doing nowadays. (Psst: There's also this new thing called rounding. Eg. You could change the $12.37 to $10, and increase your EV while decreasing RoR. Spread would no longer be 4:1 though. Or you could bet $10 half the time and $15 half the time.)

I support your right to use the definition of "unit" to which you've grown accustomed, but you really need to get up to speed on betting theory.

Harumph!
ETF

I thought that posting the link would be enough, and that people would be able to read what I had wrote before. I guess not.

This was first paragraph in the thread :

I am not sure what you mean by �unit�. Different people measure them differently. In BJA Don suggested dividing what we sometimes call a Kelly Equivalent Bank into 400 units, and I think this is a fairly typical unit size.

For the remaining posts in that thread, I used the term �unit� to mean that.

Francis� post illustrates the problems with other definitions. Some guy posts �I had a great session last night and won 1200 units� and it sounds the like the greatest win ever, and then you discover that they mean they won 10 max bets.

I would prefer to discuss these issues in terms of KEB�s.

"and that people would be able to read what I had wrote before."

Hmm. We're starting to go beyond spelling and typos here. :-)

Don

Why get so excited for such a futile subject?

Do you perhaps mean that bets must be multiples of $1?

No,I meant what I said. The world isn't limited to the USA. In most casinos I know, you must bet by multiples of the table minimum.The reason is simple: They wouldn't have the necessary "units" to pay out a Blackjack for an odd amount.
I think it's unwise for a casino to let people bet in increments of 1$ (unless 1$ is table min),not only for the possible loss of money because some dealers are uncapable to properly calculate the payout for a natural with say $47, but first and foremost because of the loss of time that goes with it.It's not only the additional time it takes for the payouts but the necessary change operations will be grossly inflated (think also of splits and doubles) and the simple action of placing a bet takes much longer.
For the same reason it's unwise for an AP to place such odd bets.The marginal win he might have by using theoretically optimal betting units will be more than eaten up by the waste of time.

You could change the $12.37 to $10, and increase your EV while decreasing RoR. Spread would no longer be 4:1 though. Or you could bet $10 half the time and $15 half the time.

I agree with the first sentence but then the unit has regained it's traditional meaning.I don't get the idea of your second sentence. Why would an AP want to ever bet more than necessary in neg. situations?
Take it easy.

Francis Salmon