Setting Table Maximum Betting Limits
Risk and volume dependence in relation to determining maximum betting limits for independent trial, negative expectation table games. By Andrew MacDonald Senior Executive Casino Operations, Adelaide Casino, 1995 |
Casino Analyser Reference Maximum Bet Volatility |
Introduction | Key Principles | Setting Limits | Mathematics | Effective Maximum Bet Limits | Variable Betting | Volume Dependence | Maximum Loss Point | Variable Bet Distribution | Non High-End Casino Operation | Analysis of Results | Conclusion |
To calculate the bet level that may reasonably be offered requires the following in this case:
1. Ignore fixed expenses and project all costs at a variable percentage based on fixed and variable levels, and represent this in place of a commission percentage. This removes the interdependence that would otherwise link the level of fixed expenses to the period under consideration. 2. Calculate the number of decisions achieved per day based on the average table hours, and decision rates per hour per table. 3. Determine the length of the period which is to be evaluated. 1 day,1 week, 1 month, or 1 year. 4. Estimate the maximum loss that the shareholders will accept for the period under consideration. 5. Calculate the maximum bet on the above basis taking into account bet distributions and market matching. As an example of this the following is provided : |
|||||||||||||||||
Average table operating hours | = 25 tables x 18 hours | ||||||||||||||||
= 450 | |||||||||||||||||
Average decisions per hour per table | = 50 | ||||||||||||||||
Total decisions per day | = 22,500 | ||||||||||||||||
Total decisions per week | = 157,500 | ||||||||||||||||
Win net of tax and variables expenses and fees are
W = (((N.E%)-(1.64. sqrt(N.V.SB))).(1-T%)-(N.C%)).(1- M%) For simplicity if the game considered is totally Baccarat with no opposing bets and, if T = 13.75% Example Consider the scenario where the shareholders were prepared to experience a loss of $350,000 or greater in any one week with a probability of occurrence of one in twenty. Based on market research and past experience an assessment can be made regarding probable bet distributions in this operation. The following three betting scenarios highlight the effect that high limit betting has on the weekly risk profile. Bet Distribution 1 |
|||||||||||||||||
Bet size= x
|
Frequency= f
|
x.f
|
(x)^2.f
|
||||||||||||||
7,000
|
1.5%
|
105.00
|
735,000
|
||||||||||||||
2,500
|
2.5%
|
62.50
|
156,250
|
||||||||||||||
500
|
10.0%
|
50.00
|
25,000
|
||||||||||||||
100
|
26.0%
|
26.00
|
2,600
|
||||||||||||||
50
|
35.0%
|
17.50
|
875
|
||||||||||||||
25
|
25.0%
|
6.25
|
156
|
||||||||||||||
SUM
|
100%
|
267.25
|
919,881
|
||||||||||||||
Standard deviation | = sqrt ( sum( x^2.f) – (sum(x.f))^2) | ||||||||||||||||
= sqrt ( 919,881 – 267^2) | |||||||||||||||||
= 921 | |||||||||||||||||
Equivalent Bet Limit | = Bav ( 1 + (SB / Bav)^2) | ||||||||||||||||
= 267. ( 1 + (921/267)^2) | |||||||||||||||||
= 3,442 | |||||||||||||||||
In other words, instead of behaving like N hands with equal bet size $267, this will behave like 0.0777N hands with equal bet size $3,442. If N = 157,500 which implies that 0.0777N = 12,229 then the win net of tax etc is
W = (( 12,229. 0.0125 – 1.645. sqrt( 0.97.x12,229)) x As the size of the bets are now assumed to be $3,442 instead of $267 this implies a 90% lower probability limit of = -93.7 . 3,442 Bet Distribution 2 |
|||||||||||||||||
Bet size= x
|
Frequency= f
|
x.f
|
(x)^2.f
|
||||||||||||||
5,000
|
2%
|
100.00
|
500,000
|
||||||||||||||
2,500
|
3%
|
75.00
|
187,500
|
||||||||||||||
500
|
9.5%
|
47.50
|
23,750
|
||||||||||||||
100
|
21.5%
|
21.50
|
2,150
|
||||||||||||||
50
|
29.0%
|
14.50
|
725
|
||||||||||||||
25
|
35.0%
|
8.75
|
281
|
||||||||||||||
SUM
|
100%
|
267.25
|
714,344
|
||||||||||||||
Standard deviation | = sqrt ( sum( x^2.f) – (sum(x.f))^2 ) | ||||||||||||||||
= sqrt ( 714,344 – 267^2 ) | |||||||||||||||||
= 802 | |||||||||||||||||
Equivalent Bet Limit | = Bav ( 1 + (SB / Bav)^2) | ||||||||||||||||
= 267. ( 1 + (802/267)^2) | |||||||||||||||||
= 2,673 | |||||||||||||||||
This behaves like 0.10N hands with equal bet size $2,673 and the win net of tax etc is W = (( 15,748.00125 – 1.645. sqrt( 0.97. 15,748). ( 1 – 0.1375) – 15,748.x 0.007).(1 – .135))x 2,673 = – 267,725Bet Distribution 3 |
|||||||||||||||||
Bet size= x
|
Frequency= f
|
x.f
|
(x)^2.f
|
||||||||||||||
10,000
|
1%
|
100.00
|
1,000,000
|
||||||||||||||
5,000
|
2%
|
100.00
|
500,000
|
||||||||||||||
500
|
5%
|
25.00
|
12,500
|
||||||||||||||
100
|
15%
|
15.00
|
1,500
|
||||||||||||||
50
|
32%
|
16.00
|
800
|
||||||||||||||
25
|
45%
|
11.25
|
281
|
||||||||||||||
SUM
|
100%
|
267.25
|
714,344
|
||||||||||||||
Standard deviation | = sqrt ( sum( x^2.f) – (sum(x.f))^2 ) | ||||||||||||||||
= sqrt ( 1,515,081 – 267^2 ) | |||||||||||||||||
= 1202 | |||||||||||||||||
Equivalent Bet Limit | = Bav ( 1 + (SB / Bav)^2) | ||||||||||||||||
= 267. ( 1 + (1202/267)^2) | |||||||||||||||||
= 5,669 | |||||||||||||||||
This bet distribution will behave like 0.047N hands with equal bet size $5,669 and the win net of tax etc is
W = ((( 7,425×00125 – 1.645. sqrt ( 0.97×7,425)). Whilst bet distributions 1 and 2 fall within what is deemed as an acceptable level of risk (in this case a one in twenty chance of losing $350,000) the large variance of the bet sizes in bet distribution 3 cause the risk to increase to above $450,000. Therefore, if a very high degree of confidence could be placed in the assessment of the bet distribution, then it would not be inappropriate to offer a $7,000 table maximum bet in this casino, based on what the shareholders have indicated that they are prepared to lose on a weekly basis. If, however, a player or players come into the casino and bet substantially more than recognised in the market analysis, there is a real danger of experiencing losses at a much greater level as demonstrated by distribution 3. Thus, if that is not acceptable at all, then a much more conservative bet range should be provided for, with maybe a $5,000 table maximum bet being allowed, or if the shareholders were totally intolerant of any greater loss within the probability allowance, a maximum limit of $3,500 could be provided depending on a re-assessment in both cases of the bet distribution and frequency of bets. Once established, this then would form the upper limit available on any of the games with lower limits, stratified if required, to allow egos and social groupings as the market requires. Of course not all games are the same within the casino and different game profiles would exist. However, this would be a reasonable framework for establishing table maximum limits within the operation, as we have used a “base case style” game with low house edge and low game variance. This bet limit would then provide the even money bet maximum for all games with higher pay-off bets being discounted appropriately. Alternatively, it would be possible to build the entire casino profile based on an assessment of the individual games on offer and the markets for each. To establish box or player maximum limits from the calculated table maximum limits merely requires a division by the number of boxes on offer and the rules pertaining to the number of players per box. Otherwise what has been calculated is a maximum bet per table per round or what is commonly referred to as a table differential. |