Quote:AhighThat would lend a lot of credibility to comfort level of customers that fair dice are in play.

Casinos already do this for roulette, and that doesn't seem to quell conspiracy theories, it just saves some players having to bring their own paper and pencils.

Quote:CalderCasinos already do this for roulette, and that doesn't seem to quell conspiracy theories, it just saves some players having to bring their own paper and pencils.

Yeah, why didn't I think of that. (slaps forehead). THANKS!

Quote:AhighYeah, why didn't I think of that. (slaps forehead). THANKS!

You are too funny man!

Quote:AhighYeah, why didn't I think of that. (slaps forehead). THANKS!

Next time put down the beer can down first. OUCH !

Your numbers are out of alignment.Quote:AhighChi-squared compared to evenly distributed outcomes comes up with a 0.393078 confidence

Chi-squared compared to the theoretically optimally gaffed dice comes up with a 0.680780865 confidence

I calculated the ideal gaffed dice for Harley's theories to be 9,8,8,8,8,9 ratios. Which scaled up is 129.92, 122.28, 122.28, 122.28, 122.28, 129.92,

129.92 130 0.680780865

122.28 125

122.28 108

122.28 125

122.28 119

129.92 142

124.83 130 0.393078479

124.83 125

124.83 108

124.83 125

124.83 119

124.83 142

Recheck your calcs. You had the 40% value for the fair dice before.

I agree that 9,8,8,8,8,9 ratios look very good to use

bias

2.175 statistic

p value 0.824479809

fair

5.117489987 statistic

p value 0.401710642

That is not what the numbers say.Quote:AhighThe numbers say these dice are 73% more likely to be the theoretically biased dice than they are to be fair dice.

Of course these dice are theoretical at this time and are not known to even exist.

One does not conpare p-values.

None of my stats book shows that.

You want the chi-square test of homogeneity, tests to see if the rows come from the same distribution or appear to come from different distributions

Under the null hypothesis that both sets of data come from the same distribution (homogeneity) and a proper sample.

You can not do that with the same distribution. You get a statistic of 0 and a p-value of 1.

examples here

http://wiener.math.csi.cuny.edu/Statistics/R/simpleR/stat013.html

But the test is still better with a larger sample.

here is what I did in R

bias die vs fair die

I used your 9,8,8,8,8,9 weights

and ran the code a few times.

Look at it.

The first thing that shows at the 750 sample size is how many p-values are below 25%.

But we want more samples

I ran it at 7500

Now look at the p-values.

Now we are talking.

Still a few high p-values, but most are very small pointing to a conclusion the samples at 7500 are NOT from the same distribution.

So I asked for more rolls

I got 12,000

Now we know the null hypothesis can be rejected with a very high degree of confidence.

> die.fair = sample(1:6,749, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,749, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 130 125 130 122 128 114

res.bias 123 116 105 135 125 145

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 7.5929, df = 5, p-value = 0.1801

>

> mean (table(die.fair))

[1] 124.8333

> mean (die.fair)

[1] 3.447263

> mean (die.bias)

[1] 3.611482

> die.fair = sample(1:6,750, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,750, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 120 130 132 131 121 116

res.bias 154 111 122 110 114 139

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 10.2235, df = 5, p-value = 0.06914

>

> mean (table(die.fair))

[1] 125

> mean (die.fair)

[1] 3.468

> mean (die.bias)

[1] 3.448

> die.fair = sample(1:6,750, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,750, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 131 129 142 122 127 99

res.bias 123 126 115 128 126 132

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 7.9861, df = 5, p-value = 0.157

>

> mean (table(die.fair))

[1] 125

> mean (die.fair)

[1] 3.376

> mean (die.bias)

[1] 3.538667

> die.fair = sample(1:6,750, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,750, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 99 132 128 133 135 123

res.bias 130 137 125 103 126 129

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 8.5918, df = 5, p-value = 0.1265

>

> mean (table(die.fair))

[1] 125

> mean (die.fair)

[1] 3.589333

> mean (die.bias)

[1] 3.46

> die.fair = sample(1:6,750, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,750, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 132 123 128 141 126 100

res.bias 137 120 117 134 109 133

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 6.7056, df = 5, p-value = 0.2435

>

> mean (table(die.fair))

[1] 125

> mean (die.fair)

[1] 3.408

> mean (die.bias)

[1] 3.476

> die.fair = sample(1:6,750, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,750, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 109 132 117 126 131 135

res.bias 140 99 113 115 118 165

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 12.8241, df = 5, p-value = 0.02508

>

> mean (table(die.fair))

[1] 125

> mean (die.fair)

[1] 3.590667

> mean (die.bias)

[1] 3.622667

> die.fair = sample(1:6,750, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,750, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 118 122 122 135 130 123

res.bias 139 109 125 143 118 116

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 3.4999, df = 5, p-value = 0.6234

>

> mean (table(die.fair))

[1] 125

> mean (die.fair)

[1] 3.541333

> mean (die.bias)

[1] 3.453333

> die.fair = sample(1:6,750, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,750, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 119 140 128 106 130 127

res.bias 147 117 114 115 123 134

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 6.5636, df = 5, p-value = 0.2552

>

> mean (table(die.fair))

[1] 125

> mean (die.fair)

[1] 3.492

> mean (die.bias)

[1] 3.469333

> die.fair = sample(1:6,7500, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,7500, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 1200 1238 1301 1260 1256 1245

res.bias 1421 1232 1173 1140 1196 1338

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 36.0882, df = 5, p-value = 9.12e-07

>

> mean (table(die.fair))

[1] 1250

> mean (die.fair)

[1] 3.515867

> mean (die.bias)

[1] 3.462933

> die.fair = sample(1:6,7500, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,7500, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 1234 1247 1241 1244 1294 1240

res.bias 1311 1206 1224 1192 1300 1267

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 4.5469, df = 5, p-value = 0.4736

>

> mean (table(die.fair))

[1] 1250

> mean (die.fair)

[1] 3.5116

> mean (die.bias)

[1] 3.502

> die.fair = sample(1:6,7500, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,7500, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 1256 1277 1234 1276 1249 1208

res.bias 1400 1213 1191 1164 1215 1317

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 20.5302, df = 5, p-value = 0.0009934

>

> mean (table(die.fair))

[1] 1250

> mean (die.fair)

[1] 3.4812

> mean (die.bias)

[1] 3.470933

> die.fair = sample(1:6,7500, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,7500, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 1245 1196 1189 1248 1317 1305

res.bias 1405 1229 1233 1154 1164 1315

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 24.0609, df = 5, p-value = 0.0002113

>

> mean (table(die.fair))

[1] 1250

> mean (die.fair)

[1] 3.548133

> mean (die.bias)

[1] 3.451733

> die.fair = sample(1:6,7500, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,7500, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 1202 1223 1257 1319 1292 1207

res.bias 1346 1222 1167 1185 1195 1385

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 34.6581, df = 5, p-value = 1.761e-06

>

> mean (table(die.fair))

[1] 1250

> mean (die.fair)

[1] 3.5196

> mean (die.bias)

[1] 3.5088

> die.fair = sample(1:6,7500, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,7500, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 1223 1193 1274 1229 1307 1274

res.bias 1314 1200 1226 1217 1172 1371

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 15.1741, df = 5, p-value = 0.009644

>

> mean (table(die.fair))

[1] 1250

> mean (die.fair)

[1] 3.5368

> mean (die.bias)

[1] 3.5128

> die.fair = sample(1:6,7500, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,7500, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 1295 1257 1236 1264 1231 1217

res.bias 1387 1163 1199 1108 1238 1405

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 31.1286, df = 5, p-value = 8.835e-06

>

> mean (table(die.fair))

[1] 1250

> mean (die.fair)

[1] 3.470667

> mean (die.bias)

[1] 3.514933

> die.fair = sample(1:6,12000, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,12000, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 2011 1939 2050 2048 1973 1979

res.bias 2205 1901 1958 1881 1878 2177

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 30.2897, df = 5, p-value = 1.293e-05

>

> mean (table(die.fair))

[1] 2000

> mean (die.fair)

[1] 3.4975

> mean (die.bias)

[1] 3.488083

> die.fair = sample(1:6,12000, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,12000, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 2009 2009 1958 2051 1955 2018

res.bias 2182 1937 1880 1941 1913 2147

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 17.5227, df = 5, p-value = 0.003608

>

> mean (table(die.fair))

[1] 2000

> mean (die.fair)

[1] 3.499

> mean (die.bias)

[1] 3.49225

> die.fair = sample(1:6,12000, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,12000, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 1992 1943 2015 1977 2023 2050

res.bias 2184 1900 1897 1956 1857 2206

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 25.8003, df = 5, p-value = 9.756e-05

>

> mean (table(die.fair))

[1] 2000

> mean (die.fair)

[1] 3.5205

> mean (die.bias)

[1] 3.501667

> die.fair = sample(1:6,12000, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,12000, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 2035 1937 2061 1996 1954 2017

res.bias 2114 1920 1943 1911 1986 2126

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 10.0335, df = 5, p-value = 0.07429

>

> mean (table(die.fair))

[1] 2000

> mean (die.fair)

[1] 3.495667

> mean (die.bias)

[1] 3.509417

> die.fair = sample(1:6,12000, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,12000, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 2082 1928 2023 2010 1932 2025

res.bias 2223 1951 1907 1840 1896 2183

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 21.956, df = 5, p-value = 0.0005338

>

> mean (table(die.fair))

[1] 2000

> mean (die.fair)

[1] 3.488083

> mean (die.bias)

[1] 3.482

> die.fair = sample(1:6,12000, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,12000, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 2006 1949 2048 2062 1999 1936

res.bias 2172 1924 1929 1943 1868 2164

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 30.9703, df = 5, p-value = 9.495e-06

>

> mean (table(die.fair))

[1] 2000

> mean (die.fair)

[1] 3.49225

> mean (die.bias)

[1] 3.491917

> die.fair = sample(1:6,12000, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,12000, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 2028 2046 2022 1907 1993 2004

res.bias 2251 1914 1910 1860 1935 2130

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 24.495, df = 5, p-value = 0.0001744

>

> mean (table(die.fair))

[1] 2000

> mean (die.fair)

[1] 3.483583

> mean (die.bias)

[1] 3.475333

> die.fair = sample(1:6,12000, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,12000, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 2001 2049 1930 1938 1986 2096

res.bias 2134 1947 1912 1932 1939 2136

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 7.916, df = 5, p-value = 0.1609

>

> mean (table(die.fair))

[1] 2000

> mean (die.fair)

[1] 3.51225

> mean (die.bias)

[1] 3.50025

> die.fair = sample(1:6,12000, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,12000, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 1959 2013 2014 2019 1945 2050

res.bias 2211 1908 1897 1907 1938 2139

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 26.6393, df = 5, p-value = 6.705e-05

>

> mean (table(die.fair))

[1] 2000

> mean (die.fair)

[1] 3.510667

> mean (die.bias)

[1] 3.489167

> die.fair = sample(1:6,12000, p=c(1,1,1,1,1,1)/6,replace=T)

> die.bias = sample(1:6,12000, p=c(9,8,8,8,8,9)/50,replace=T)

> res.fair = table(die.fair);res.bias = table(die.bias)

> rbind(res.fair,res.bias)

1 2 3 4 5 6

res.fair 2058 2008 2002 2008 1996 1928

res.bias 2184 1860 1966 1921 1903 2166

> chisq.test(rbind(res.fair,res.bias))

Pearson's Chi-squared test

data: rbind(res.fair, res.bias)

X-squared = 27.7126, df = 5, p-value = 4.142e-05

>

> mean (table(die.fair))

[1] 2000

> mean (die.fair)

[1] 3.471667

> mean (die.bias)

[1] 3.49975

>

That is why sample size is so important.

Even at 3750 dice rolls we were getting many p-values that were boring (showing same distribution)

but were still on the lower side.

The Bias sect claims to have over 10,000 dice rolls.

2 dice totals are fine, I doubt they have face frequencies.

Where is that data? What is so secret about it?

Ahigh, get your proper sample size and this can be settled very easily.

My $1 says yes.Quote:CalderNope, a genuine question.

What problem do you intend to solve with a counter?

Wouldn't any noticeable variation from a normal distribution of dice rolls on the counter simply lead to all kinds of pointless complaints?

All players and the DIs would be calling their lawyers, the NGC, the FBI, CNN, Oprah... after the first two quick 6-1 7outs in a row.

Could not blame them.

The reason I ask is because a counter is the first thing mentioned in this thread that would in any way be manifest at a table at which I'm playing: "Oh, look, a dice counter."

I don't sweat biased dice, I don't crunch numbers from prior rolls, whether controlled rolls are possible has little bearing on me given where and how often I play, and I wouldn't watch craps on the internet just as I don't watch poker on TV. If you and others are interested in those aspects of the game for fun or profit, have at it, no one is stopping you.

But the counter you suggest is something I'd actually see at a table (other than players recording rolls, I guess). A tangible change in the environment around the craps table.

Thus my question, what would it really settle? Or rather, might it needlessly cause more distrust and suspicion at a table game that is often already rife with superstition, distrust, and conspiracy theories?

It wouldn't change the way I bet; I'd attribute the rolls displayed to random, unpredictable, and unexploitable variations in normal roll distrubution.

If your suggestion was just and end-of-post throw away, don't sweat it. My question was genuine, no offense intended.

So let me suggest this:

It is not in the interests of a casino to have biased dice just as it isn't in the casino's interest to have a biased roulette wheel. As soon as someone figures out the bias the casinos will lose.

If the games are fair, the casino's edge on the bets is all it needs to guarantee its profits.

If you believe otherwise it makes me think that all of you "biased dice" conspiracy theorists must be taking classes with Rob Singer (bless his soul).

1) They already do it and just don't share it with the players

2) They don't already do it and the only way they know if dice are suspicious is if the table starts dumping or taking more money than it should

It is an ABSOLUTE FACT that the dice become biased towards certain outcomes as a result of normal use. This thread is not about that subject at all, and the stick I counted had just gone into play when I started counting. I did miss the first 40 rolls or so in my counts, but if it wasn't a new stick, it had not seen enough play for me to tell the difference.

I absolutely agree with you that most people trust the casino when it comes to using fair dice. And therein lies the problem. Once you begin asking questions about the dice themselves, it becomes obvious that the trust is not warranted.

I personally believe that casinos do not count die face outcomes, and instead, they just swap out the dice any time that the table starts dumping. The whole idea that a brand new fresh stick of dice could have bias that goes against the average bet on the table or even the expected type of bet for a particular player is an idea that seems ridiculous to most who would be in the position to question the dice in the first place. IE: it's not obvious that you can do this until you run the math through a simulator. And the fact that the only exposure without massive volatility in between the player and the wins is on the field bet is a little unexpected to me.

All that I am suggesting is that this _SHOULD_ be a big enough problem for the casinos to at least keep counts of the faces. In other words, what harm would come of keeping those counts to decide if the dice are fair or not? Instead of the table dumping, they _SHOULD_ be concerned any time the results look suspicious on a chi-squared test of expected results assuming there have been enough results to be concerned. In other words, instead of just changing the stick of dice when they stand to lose more than expected, they should chance the stick of dice any time anyone stands to lose more than expected.

I honestly think that most people will simply not accept the theory of casinos making more money by using different dice as plausible enough to worry about it, and this practice will be something that people like Harley and others who are paying attention are aware of, and everyone else just keeps thinking that they got unusually unlucky.

One of the personal aspects of this dilemma is out good friend, TeddyS. Teddy was about the only player I ever met who could keep up with my marathon craps playing habits, and he played at several of these sweat joints with me that are suspected of using these cheap dice. Teddy did all of his homework and followed the Wiz's advice on keeping his house edge low to limit the cost of the game over the long run, and he was playing for the long run just like me, and he got destroyed.

I had already adapted my play techniques to be more tolerant of a higher number of sevens, but Teddy was betting like the Wizard would say the smart way to bet would be.

In summary, if you deny that a problem exists, you are right. I am thinking in the context of after people are aware of this problem, which currently has not been proven to exist, it merely looks like there is enough evidence to continue digging into it. But if we get to a point where there is no doubt that this problem exists, casinos should be frank in admitting this is a way that they can legally make more money by deceiving the player into thinking the dice are fair when they are not, and they should be willing to take additional steps to advertise that their dice are fair in a way that cannot be faked: show the outcomes.

We are a LONG way from the average joe being aware of this problem because it is not yet known to even be a problem. But if it is in fact real, I will keep digging until I am assured that it's not a problem or until we have a story on this from an authority capable of following up on holding the casinos responsible for being fair to their customers.

Quote:AlanMendelsonI'm just curious now why there is so much concern about fair dice in casinos? Why is no one asking if the roulette wheels are biased?

So let me suggest this:

It is not in the interests of a casino to have biased dice just as it isn't in the casino's interest to have a biased roulette wheel. As soon as someone figures out the bias the casinos will lose.

If the games are fair, the casino's edge on the bets is all it needs to guarantee its profits.

If you believe otherwise it makes me think that all of you "biased dice" conspiracy theorists must be taking classes with Rob Singer (bless his soul).

The assumptions have been that the dice have been fair. The Wizard very openly chastised me directly to my face for even DOING face counts. He walked up as I was counting faces for the second time in my life. The Fiesta Rancho was the first time I had ever done it. Point being the reason why there is now concern is because there is evidence that looks suspicious to me personally, and because I am concerned.

Roulette is a totally different game with an edge per event ten to 100 times as high as craps per event (the roll compared to the spin). Craps is advertised as having a low edge and as such is expected to be a more fair game warranting larger bets than roulette.

IE: In layman's terms roulette is for dummies and/or nobody expects to be able to grind a $5 bet on red for very long at 5% or higher per spin compared to a $5 passline at less than 1/10th the cost per roll than roulette's cost per spin.

People choose to bet larger amounts, ESPECIALLY ON THE ODDS BETS, under the faith and assumption that the dice being used are not compromised to make those free odds bets fair.

Once someone with more authority than me steps in to suggest where this is going, I will happily step aside. But as far as I can tell, there is nobody willing to subject themselves to being associated with the dice conspiracy theorists who cares to follow up on it. Until then, I will keep digging even if it turns into a colossal waste of time.

Don't forget I already spent a huge amount of time on this last year when I was merely focused on the balance issue. I don't currently think that following up on balance is even worthwhile because proving one way or another on balance, you could still have another effect causing bias (material elasticity or corner wear for example).

That being said, I think it's suspicious that there is zero demand for the balance I had created: because casinos don't care. They mic 'em and they put 'em in play, and if they dump, they pull 'em. That's it as far as what they are held to be responsible to do here in Nevada.

Quote:AhighTeddy did all of his homework and followed the Wiz's advice on keeping his house edge low to limit the cost of the game over the long run, and he was playing for the long run just like me, and he got destroyed.

Bear in mind Teddy was critical of himself during his bad run , IIRC, in some aspects; in any case, I'd like to hear from Teddy on that.

Teddy and the Wizard himself both know, and in the Wizard's case, openly give me a hard time for not betting in a way that minimizes the house edge.

I don't know how much the Wizard plays craps himself, but it's quite possible that I have more play experience on the game of craps than the Wizard. I'm certain I have more play experience than Teddy who was following very strictly the Wizard's betting strategies.

I play this game a lot. And I rarely see locals who cruise around to these places with 10x and 20x odds that bet max odds. Teddy built up quite a reputation here locally pretty quickly among the dealers as being a rare bird. And all he was doing was following the Wizard's betting strategies.

Let me say that another way: very few people bet max odds on 10x and 20x tables and play as frequently as Teddy was playing. I know a FEW max odds betters who bet like Teddy and the do and fewer on the don't. But they don't play as often as I do or as often as others who have other strategies.

So Teddys got destroyed partially because of bad dice and not the FACT he did NOT do all his homework on how he played?Quote:AhighTeddy was about the only player I ever met who could keep up with my marathon craps playing habits, and he played at several of these sweat joints with me that are suspected of using these cheap dice.

Teddy did all of his homework and followed the Wiz's advice on keeping his house edge low to limit the cost of the game over the long run, and he was playing for the long run just like me, and he got destroyed.

He over bet his bankroll.

That has already been shown.

myth busted

https://wizardofvegas.com/forum/gambling/craps/12558-really-blew-out-my-bankroll/9/#post208658

My comment here:

so a EV + 2*SD bankroll (for a 1 in 20 session bust rate)

$3336

Way Underfunded at $1000

Those that push the lowest house edge and max odds, lots of variance,

also forget to mention how important the proper bankroll is and what kind of bankroll should be used.

The Wizard nowhere shows beting 10X odds what is a proper bankroll for a certain risk of ruin playing Craps

Teddys had a 50% risk of ruin the moment he walked into the casino

and when he made larger bets even when winning, that went even higher.

https://wizardofvegas.com/forum/gambling/craps/12558-really-blew-out-my-bankroll/14/#post208941

https://wizardofvegas.com/forum/gambling/craps/12558-really-blew-out-my-bankroll/11/#post208764

https://wizardofvegas.com/forum/gambling/craps/12558-really-blew-out-my-bankroll/10/#post208678

Nice post, but...

The reason why it "does not work" is the players fault. almost 100%

Too small a bankroll! Yep, you got it right.

They are underfunded. BIG TIME and it ain't no funny...

Most do not show proper bankroll needed to play different methods in Craps.

Math or simulation data.

The Wizard does not.

Not even a simple RoR table available.

Only a very few gamblers will ever play with the proper bankroll for their selected bet sizes and length of session play.

The others will just bitch and moan and blame all the mathss gurus for their misfortune, that they mostly brought onto themselves

I absolutely think there should to tote boards similar to roulette on every craps table. It should show the last 20 or so numbers rolled, horns and hardways, length of roll, and other optional info like face percentages, expected/actual ratio of numbers rolled. It will never happen because the pit will say it is too much work to enter two numbers every time (even though they do stuff like that for the All Tall bets, etc.) But the first casino to add this will get a ton of business, I guarantee it. It is already a feature on the electronic bubble craps games.Quote:AhighFrankly, I think that there should be six counters on each craps table that shows the occurrence of each face for the stick that is currently in play. That would lend a lot of credibility to comfort level of customers that fair dice are in play.

The point is that he did his homework to keep his edge low and he firmly believed the dice were 100% fair, and there is some question in my mind whether this was the case or not.

I don't have anything more to say about my craps losses. It's pretty self-evident what happened, and I think others have explained it pretty well (e.g., 7craps). The fact is I didn't appreciate the variance curve well enough. A $10,000 bankroll was nowhere near enough for the way I was playing, even on a 99.82% payback game. (Most video poker players would love to play with that percentage.)Quote:AhighLet's not argue whether or not Teddy was responsible for what happened. That is not the point.

The point is that he did his homework to keep his edge low and he firmly believed the dice were 100% fair, and there is some question in my mind whether this was the case or not.

A 10k bankroll for a 4 hour session would be fine.Quote:teddysThe fact is I didn't appreciate the variance curve well enough. A $10,000 bankroll was nowhere near enough for the way I was playing, even on a 99.82% payback game. (Most video poker players would love to play with that percentage.)

You were complaining you had long session losing streaks.

What gambling writers (Wizard included) mentions

how to calculate the starting Craps bankroll requirements for a certain bet size, session length of play, number of bets and type of variance bets made?

It ain't ALL Teddys fault.

Answer: there are none, only a bunch of "rules of thumbs" I like that

(there are many that show how to calculate your expected loss. That is all that matters to most gambling writers. Sad)

If there are it is a well kept secret.

Just google: bankroll needed to play craps

You will see a link to something like this

Proper Bankroll for Craps

http://scoblete.casinocitytimes.com/article/proper-bankroll-for-craps-31134

"If you are going to spread at $45, then you should bring $450.

That's 10 times your spread. Having that kind of money backing you will do several things."

Ah, easy money

One size fits all

Just a reminder, the topic is:

"BAD DICE: the saga CONTINUES"

You may want to put the criticizing section of your brain into neutral for while.

Quote:teddysIt will never happen because the pit will say it is too much work to enter two numbers every time (even though they do stuff like that for the All Tall bets, etc.)

The dealer manually enters every roll in Rapid Craps, and there usually are no problems (unless the wrong numbers are entered).

I do not live in Vegas, maybe some day (I lived in Reno) but you do.Quote:Ahighand there is some question in my mind whether this was the case or not.

You have biased dice at home too.

Take some time without telling anyone and get your sample size to 10k or more and you will have your answer.

(It will take me quite some time to get my sample size high from casino throws)

Track known biased dice and known fair dice.

You could do this easily. Your kids can even roll the dice for you. make it an event that is fun.

Does not have to be done in one day.

Rome was not built in one day.

When some one claims the casinos are knowingly cheating their craps players

it would be sweet to show some data backing up that claim up instead of a few super small samples and a few videos showing more 1s and 6s have just rolled.

But for some that IS proof that would stand up in their court of law.

They have to or the terminals do not know what bets win or lose.Quote:tuppThe dealer manually enters every roll in Rapid Craps, and there usually are no problems (unless the wrong numbers are entered).

Fact still remains.

More data is needed to prove that casinos are knowingly using biased dice

and then change the Nevada laws on dice specs

as the proof continues to build up from more data collected.

Quote:

SanchoPanza

That sounds like a flat-out waffle setting up a perfect escape valve, to mix a couple of metaphors. It will be fascinating to see any explanation of how it can be possible to penalize the do's while not rewarding the don'ts. Unless all the dice do is come up boxcars -- a hell of a lot more than 1 in 36.

This has nothing to do with you SanchoPanza, but I've heard the same argument for so long, its just funny when anybody brings it up. I hear it from the guys that are losing on the tables when someone says something about bad dice that plays everday and can't figure out why they are now losing so much. These are the same guys that have been playing craps for years and haven't changed they betting strategy. Most of them have falling by the wayside and are no longer playing like they used to play. They are the ones that blame it on bad luck.

You got to love someone that thinks about only the math of the game they can tell you why the only way to win at craps is by betting on the 6's and 8's as they are the best bets on the table, after all every math book out there says the samething. You don't need to be to intelligent to win at craps if you go by what the math boys tell you to , just bet the 6's and 8's and you just might be a winner, if you happen to get lucky.

But wait a minute; why is it that I keep seeing all those players following that strategy go down in flames. Could it be that when you are playing craps the math of the game doesn't mean a thing, surely you do need to know it, but it doesn't do you one bit of good if the 6's and 8's are not being rolled!

I know that just about every book out there will tell you that the very best bet on a craps table is the “Pass-Line bet with full odds.”

Well then, why aren't the players that listen to that foolishness winning every time they go to the tables?

I have laugh my butt off every time I've heard your type of arguments about playing the don't ever since the bad dice came to light. Guys like you need to spend the time and money to figure this out for yourself, or go stick your head in the sand, like everybody else wants to do. It doesn't matter to the guys that have spent the time figuring out that there is a problem. Lose your money at they craps table and always blame it on your bad luck, or the fact that you didn't have a big enough bankroll to over come the variance.

Ahigh gave a good explanation with his graphs, but some people will never see beyond what they think they know. Their brains get in their way!

Quote:

AlanMendelson

I put that question to the Nevada gaming authorities and I wrote about it some time ago in an article on my website. Here's the quote:

There's a Zellhoefer of the Encorcement Division told me that "Nevada operates differently" than New Jersey and Nevada does not publish specific rules or regulations for the game of craps. "We established minimum controls," she told me, and the casinos have to meet those minimums and they can exceed them, "and then we say okay."

So Alan that answer satisfied you, it doesn't take to much to appease you or anybody else that plays craps here in NV!

You don't care that there are no rules in place that are there to protect you when you are playing craps?

We know that New Jersey, Colorado, and yes Pennsylvania all have rules to protect their players but NV has none, maybe that is why Pennsylvania is now second in the nation for a gambling destination. They have rules that give the players a fair shake for their money! Isn't that what we all want, isn't that why they have legal casinos, and why we aren't playing at the underground craps tables that someone is running out of a back room?

Let's take a look at Pennsylvania's rules for craps:

http://www.pacode.com/secure/data/058/chapter623a/058_0623a.pdf

http://gamingcontrolboard.pa.gov/files/regulations/Final_Regulations_Master.pdf

§ 603a.12. Dice; physical characteristics.

(a) Except as otherwise provided in subsections (b) and (c), each die used in the play of table games must:

(1) Be formed in the shape of a perfect cube and of a size no smaller than 0.750 inch on each side nor any larger than 0.775 inch on each side, with a tolerance of +/- 0.005.

(2) Be transparent and made exclusively of cellulose except for the spots, name or logo of the certificate holder and serial number or letters contained thereon.

(3) Have the surface of each of its sides perfectly flat and the spots contained in each side flush with the area surrounding them.

(4) Have all edges and corners perfectly square and forming 90° angles.

(5) Have the texture and finish of each side exactly identical to the texture and finish of all other sides.

(6) Have its weight equally distributed throughout the cube with no side of the cube heavier or lighter than any other side of the cube.

(7) Have the six sides bearing white circular spots from one to six respectively with the diameter of each spot equal to the diameter of every other spot on the die.

(8) Have spots arranged so that:

(i) The side containing one spot is directly opposite the side containing six spots.

(ii) The side containing two spots is directly opposite the side containing five spots.

(iii) The side containing three spots is directly opposite the side containing four spots. 393

(9) Each spot shall be placed on the die by drilling into the surface of the cube and filling the drilled out portion with a compound which is equal in weight to the weight of the cellulose drilled out and which forms a permanent bond with the cellulose cube and extends into the cube exactly the same distance as every other spot extends into the cube to an accuracy tolerance of 0.0004 inch.

(10) Have imprinted or impressed thereon a serial number or letters and the name or logo of the certificate holder in whose licensed facility the die is being used.

(b) Dice used in the table games of Pai Gow and Pai Gow Poker must comply with the requirements of subsection (a) except as follows:

(1) Each die must be formed in the shape of a perfect cube and of a size no smaller than 0.637 inch on each side nor any larger than 0.643 inch on each side.

(2) Instead of the name or logo of the certificate holder, a certificate holder may, with the approval of the Board’s Executive Director in accordance with § 601a.10(a) (relating to approval of table game layouts, signage and equipment), have an identifying mark imprinted or impressed on each die.

(3) The spots on each die do not have to be equal in diameter.

(c) Dice used in the table game of Sic Bo must comply with subsection (a) except each die may be formed in the shape of a cube 0.625 inch on each side with ball edge corners.

d. Dice may not be utilized in a licensed facility unless a detailed schematic depicting the actual size, color of the dice as well as the location of serial numbers, letters or logos has been submitted to the Bureau of Gaming Operations and approved in accordance with § 601a.10(a).

Isn't it funny that the one state that legalized gambling first couldn't come up with rules for playing craps in all these years.

There I did the work, that all you guys that are saying prove it should be doing. I gave you the rules that other states are using. Now surely this doesn't prove that there are bad dice, but it does prove that there are rules to protect players in different state around the country from bad dice! Instead of looking to get the rules changed most guys would rather say, you guys prove that the dice are bad, so we can take advantage of what you found!

Sorry but I don't think so, do the work yourself!

We could always say prove that the dice are 100% fair, I'm sure one of you math guys could use a math project like that, it would give you something to do beside wanting a free ride on somebody else is hard work. Lets make it for 20,000 rolls of the dice on real tables, but please don't do it in Pennsylvania New Jersey or Colorado!

It should be done right here in Vegas where there are no rules as to what they use on the tables!

Now I know what everybody wants to say , so I will say it for you. If you know the dice are bad then why are you playing craps?

Very simple, I did my homework and know how to get around them and if I can't I leave the table.

Quote:7crapsWhen some one claims the casinos are knowingly cheating their craps players

it would be sweet to show some data backing up that claim up instead of a few super small samples and a few videos showing more 1s and 6s have just rolled.

What is super small is the number of samples of possibly bad dice you have provided.

I spent multiple hours seven days ago putting together 749 samples from an actual casino.

I know you're not necessarily trying to sound critical, but put yourself in my shoes.

I went and spent a couple of hours talking to Harley the day before that about what to look for too.

And I drove there and back both days (about 40 miles of driving).

Not good enough, huh?

Interesting perspective that I only have a few super small samples and a few videos, because I'm not sure who else is doing a better job following up on this.

It all takes time to do all this stuff.

Add to that losing money making max bets on the passline before these dice are sent to the table, and then add in your comments.

It's interesting, your removed, un-vested, yet seemingly superior position of knowing how to better do what I am doing.

Add to all of this that I was the guy saying none of this existing a year ago (IE: this goes against my claims from before).

There is no two ways about it: what's left is that this is a lot of work to do, and really no significant reward for me personally.

And now we have Alan saying I have "too many conspiracy theories" and I'm sure many other people who are not open to the possibility of this being real are also thinking this negatively affects my credibility.

But all this happened just like I described it. Sure it's possible that this was all random, though. Just like the 9 hardways in 10 throws, or meeting the Wizard "randomly" hours after telling him after I wanted to meet him, it could just be one of those "what are the odds?" things.

Just random things always happen and make me think they are not random because I'm not smart. I'm sure that's the most likely explanation. So laugh it up guys!

Quote:7crapsThey have to or the terminals do not know what bets win or lose.

Really? I thought that the dealers were manually entering the rolls merely to exercise their fingers.

Quote:7crapsFact still remains. More data is needed to prove that casinos are knowingly using biased dice and then change the Nevada laws on dice specs as the proof continues to build up from more data collected.

Not sure why you are making this statement following my quote.

By the way, I have never made a claim that casinos are knowingly (or unknowingly) using unbalanced dice.

Quote:7crapsget your sample size to 10k or more....Track known biased dice and known fair dice.

When some one claims the casinos are knowingly cheating their craps players

it would be sweet to show some data backing up that claim up instead of a few super small samples and a few videos showing more 1s and 6s have just rolled.

.

Now we're talking.

Good point 7craps. Let's see how "biased dice" (excluding deliberately weighted dice) compare to "unbiased dice" after 10,000 rolls.

Quote:superrickQuote:

AlanMendelson

I put that question to the Nevada gaming authorities and I wrote about it some time ago in an article on my website. Here's the quote:

There's a Zellhoefer of the Encorcement Division told me that "Nevada operates differently" than New Jersey and Nevada does not publish specific rules or regulations for the game of craps. "We established minimum controls," she told me, and the casinos have to meet those minimums and they can exceed them, "and then we say okay."

So Alan that answer satisfied you, it doesn't take to much to appease you or anybody else that plays craps here in NV!

You don't care that there are no rules in place that are there to protect you when you are playing craps?

Well, I guess you would want such rules in Nevada as those in other states if you thought the casinos in Nevada were using rigged or bias dice.

I have never suspected rigged or bias dice at any casino I have played at.

I have seen shooters have monster rolls, and shooters throw nothing but point-7s using the same dice from the same bowl.

Just how much regulation do you need? Do you need written rules in Nevada that says the dealers must make the correct payout? Do you need a rule that defines what the field bet pays? Do you need a rule that says a shooter must throw the dice below the eye level of the dealers? Do you need a rule that says a shooter has no more than ten seconds to throw the dice after the stickman pushes the dice to him?

There is a reason why the "other states" have comprehensive rules that Nevada doesn't have: bureaucracy. And politics. Someone probably had the idea that if everything was written down in black and white that it would be easier to pass gaming in the state, and someone probably had the idea that if everything was written down in black and white that it would create a political job here and there.

Do you really need a rule that says:

(10) Have imprinted or impressed thereon a serial number or letters and the name or logo of the certificate holder in whose licensed facility the die is being used.

Nevada casinos do it as way to protect themselves from loaded dice being introduced into the game. I guess in other states they have to tell the casinos how to protect themselves.

I wish you guys who fear biased dice were video poker players. You could party with Rob Singer.

I posted earlier I will NOT be providing any data to the world, I keep it in my team's knowledge only from what I collect.Quote:AhighWhat is super small is the number of samples of possibly bad dice you have provided.

And since I do not live in Vegas, I would have to collect data from the short times I visit Vegas.

exactly.Quote:AhighIt all takes time to do all this stuff.

And access to actual Nevada casinos. I am only 200 miles away from Primm and their new layouts with NO padding suck!

I do not play there unless I am tired.

But your 140 dice roll video and those dice that started you down this road still exist.Quote:AhighThere is no two ways about it:

what's left is that this is a lot of work to do,

and really no significant reward for me personally.

And many think that session was 100% proof of bias and their view of biased dice used knowingly by Las Vegas casinos.

There are even videos stating that as a FACT from your show results.

Fascinating.

You have those dice.

Have you thrown a few more 140 roll sessions with those magical 6-1 7 producing dice?

That would show what is up or down with that one session.

But most do not even understand what to expect from any random variable over a finite time frame.

Here is the link to my comments on that 140 roll session, I am sure it was lost in the many pages of that thread.

https://wizardofvegas.com/forum/gambling/craps/12821-ahigh-show-tuesday-at-7-30pm-pacific-time/48/#post220567

"You see, ya can't please everyone, so ya got to please yourself"Quote:AhighAnd now we have Alan saying I have "too many conspiracy theories"

and I'm sure many other people who are not open to the possibility of this being real

are also thinking this negatively affects my credibility.

The claim still is "Las Vegas Casinos are systematically using Biased Dice in an effort to cheat their customers and increase profits."

One must either agree with it or disagree. No grey area here.

IMO, There are no perfect casino dice any where. They would cost way too much to make.

And they would not remain perfect even after the first few throws.

Every die has some bias from perfection.

how much difference does the difference really make??

You have the bias dice that started this all.

How much do they weigh vs. some fair dice you have? Why are they biased?

AlanM keeps asking and he get no good answers.

Roll more dice and see.

Do not roll more dice and do not see and believe

(to a certain confidence level)

6,1 7s roll more because the 5 and 6 faces are heavier.

Continued Good Luck in your efforts and more positive credibility

- Keyser

34 35 29 29 26 37

I wasn't there very long. There were 4 twelves AFTER I started recording in the first 25 rolls. There were several twelves before I started recording too.

Wenda asked at one point, "are you going to play or are you just going to..." she didn't finish her sentence.

I responded by saying, "do you want me to leave?"

She said nothing.

Carl was there too. I think they knew what I was doing but I didn't tell them. Not that many samples though.

31.667 34 0.704320276

31.667 35

31.667 29

31.667 29

31.667 26

31.667 37

Quote:AhighI collected these numbers at Gold Coast today

34 35 29 29 26 37

I wasn't there very long. There were 4 twelves AFTER I started recording in the first 25 rolls. There were several twelves before I started recording too.

Wenda asked at one point, "are you going to play or are you just going to..." she didn't finish her sentence.

I responded by saying, "do you want me to leave?"

She said nothing.

Carl was there too. I think they knew what I was doing but I didn't tell them. Not that many samples though.

31.667 34 0.704320276

31.667 35

31.667 29

31.667 29

31.667 26

31.667 37

How many times did the casino switch the dice out during your "research"? That info would help those of us trying to determine casino bias and fear from players who may or may not have an advantage over them due to their dice skills.

Quote:BozHow many times did the casino switch the dice out during your "research"? That info would help those of us trying to determine casino bias and fear from players who may or may not have an advantage over them due to their dice skills.

None. Any other rhetorical questions? This was hardly enough samples to make any conclusions at all. But there were a lot of 12's.

It's just more data from another suspect casino, is all. Nothing conclusive.

I would also like to say thank you to Boz for joining in the conversation. Nice to meet you and thank you so much for your wonderfully insightful and intelligent posts!

At least I made a profit during this research trip. But I wasn't playing during the majority of the time I was collecting samples.

By the way, Boz, do you happen to play craps?

Quote:KeyserI can't see any value in adding the two dice together if you're attempting to prove that a bias exists within the dice. All you're doing is washing out and contaminating what little data you're collecting.

These are the face counts. There are two faces per each roll resolved. The list represents the number of ace, two, three, four, five, and six faces. So if boxcars roll, I count two six faces. If aces rolls, I count two ace faces, etc.

Since the shooter can select any two dice, the only thing that matters is the count of each of the faces in terms of establishing any theoretical bias from the stick.

Quote:AhighI collected these numbers at Gold Coast today

34 35 29 29 26 37

I don't understand what these numbers are? What is a 29 for example??

Quote:AlanMendelsonI don't understand what these numbers are? What is a 29 for example??

Sorry.

34 - 1 face count

35 - 2 face count

29 - 3 face count

29 - 4 face count

26 - 5 face count

37 - 6 face count

Six possible outcomes for a die .. six faces .. these are the face counts. I put these weights into my simulation to generate random data with these weights. It doesn't mean anything until you have enough samples, though, and this was only about an hour's worth of data.

What did you count? Is a 3-4 a seven? Is a 2-9 an eleven?

I added up all the sides and it kept coming up 28?

I don't understand either Ahigh,

Petro

But this does mean something.Quote:Ahigh34 - 1 face count

35 - 2 face count

29 - 3 face count

29 - 4 face count

26 - 5 face count

37 - 6 face count

Six possible outcomes for a die .. six faces .. these are the face counts.

It doesn't mean anything until you have enough samples, though, and this was only about an hour's worth of data.

Your cell counts are at least 5 in 80% of them so you can test this distribution against your null hypothesis.

For a video on this see (chi-squared test)

http://www.youtube.com/watch?v=WXPBoFDqNVk

null hypothesis = the distribution comes from a fair die

Say you are comparing your observed data to a fair die

(1,1,1,1,1,1) (ev: 31.67 from your data)

alternate hypothesis = the distribution could not come from a fair die

let us give a .01 significance value (1 in 100 samples could still be from a fair die)

2.884210526 statistic

5 degrees of freedom

p-value = 0.717831519

71.78% of these 190 samples would have a p-value as skewed or worse.

It would have to be less than 1% for us to reject the null hypothesis

So we accept the null H

Next...

null hypothesis = the distribution comes from a certain biased die

alternate hypothesis = the distribution could not come from a certain biased die

let us give a .01 significance value (1 in 100 could be from a fair die)

Say you are comparing your observed data to a biased die

(9,8,8,8,8,9) 18% vs 16% (ev: 34.2, 30.4, 30.4, 30.4, 30.4, 34.2)

1.692 statistic

5 degrees of freedom

p-value = 0.889874494

88.99% of these 190 samples would have a p-value as skewed or worse.

It would have to be less than 1% for us to reject the null hypothesis

So we accept the null H

results show that the distribution could have come from a fair die or a certain biased die.

We do not know which one or which one is more likely.

The sample size needs to increase.

And Ahigh is right.

A larger sample is needed.

He could have also gathered the totals from both dice and run the same tests also.

But only counting faces, we have no total data.

Nice work

This is some pretty strenuous work, that you have lined up for yourself... I hope you can go back to having fun afterwards.. I equate it to working at a slaughter house for 15 hours and then fixing yourself a steak..

Quote:TheWolf713This data cannot be used... Unless he has the data of every die in the pack... The shooter just picks up two dice.. And unless they are the same two everytime, the data wont accurately depict if there is a bias or not.

This is some pretty strenuous work, that you have lined up for yourself... I hope you can go back to having fun afterwards.. I equate it to working at a slaughter house for 15 hours and then fixing yourself a steak..

If they are of the same stick, then quite possibly the same deformity would influence the same bias, no?

The data can be used if all the dice from the same stick have the same bias.Quote:TheWolf713This data cannot be used... Unless he has the data of every die in the pack... The shooter just picks up two dice.. And unless they are the same two everytime, the data wont accurately depict if there is a bias or not.

I think that is what the bias team claims because

all the dice in a stick all have the same heavy white material for the pips (unless they are using some heavy metal flakes in it)

Quote:DeMangoIf they are of the same stick, then quite possibly the same deformity would influence the same bias, no?

I would say yes and no...

Quote:7crapsThe data can be used if all the dice from the same stick have the same bias.

I think that is what the bias team claims because

all the dice in a stick all have the same heavy white material for the pips (unless they are using some heavy metal flakes in it)

Ok so they are concerned with the batches versus the individual die itself.. Gotcha

Quote:TheWolf713Ok so they are concerned with the batches versus the individual die itself.. Gotcha

The selection of the dice and those details are all part of the equation.

I'm looking at overall bias. I have no idea of the details at this point, and just collecting data to see if there is enough to warrant continuing to dig.

I have no idea, for example, if some dice are biased to the six and some to the aces and it depends on what dice the shooter picks as to whether you're going to see more aces, more boxcars, or more 6-1 seven-outs. Or if there is some magical way to bias each cube to come up with sixes and ones more often than 2, 3, 4, and 5.

But really, it doesn't even matter all that much how, it's just the overall bias for all the cubes that resolve. And fortunately that's a lot easier to keep track of anyway. And it's nearly impossible to keep all six dice straight (yes, they use SIX dice at Gold Coast, interesting, huh?)

Quote:KeyserDo you guys have any real dice test studies of at least a few thousand trials, instead of just a few hundred? I would be interested in seeing something more statistically relevant if you have it.

- Keyser

What a silly question ?Tr ials of a few 1000 rolls might actually prove something !

Don't expect that to happen anytime soon.

She's a smart cookie, and I think they know plenty about what I am doing because I told them as much several weeks ago. Basically that I thought all the dice conspiracy theories were bullshit, but if there was any validity to it, they were right in the middle of it because just looking at their dice they appear to me to be suspicious looking dice!

Even back then they didn't say anything like "yeah that's freaking hilarious" or "you should do a count and see for yourself" or anything else. It just got really quiet. (!!!!)

I don't even try to operate under the radar. I even told them as I was counting and out loud, "WOW THAT'S 4 BOXCARS in 24 ROLLS. THAT IS UNUSUAL!"