suspicious skewness and kurtosis dependency on number of MC samples

[mcol]

This code

data(ExampleData.DeValues)
set.seed(1)
rbindlist(calc_Statistics(ExampleData.DeValues$BT998, n.MCM=100), idcol="type")
rbindlist(calc_Statistics(ExampleData.DeValues$BT998, n.MCM=1000), idcol="type")
rbindlist(calc_Statistics(ExampleData.DeValues$BT998, n.MCM=10000), idcol="type")

produces the following output:

> rbindlist(calc_Statistics(ExampleData.DeValues$BT998, n.MCM=100), idcol="type")
         type     n     mean   median   sd.abs    sd.rel   se.abs   se.rel   skewness   kurtosis
       <char> <int>    <num>    <num>    <num>     <num>    <num>    <num>      <num>      <num>
1:   weighted    25 2896.036 2884.460 240.2228  8.294885 48.04457 1.658977   1.342018   4.387913
2: unweighted    25 2950.818 2884.460 281.6433  9.544584 56.32866 1.908917   1.342018   4.387913
3:        MCM    25 2949.896 2885.441 296.4963 10.051073 59.29925 2.010215 126.121735 470.492370
> rbindlist(calc_Statistics(ExampleData.DeValues$BT998, n.MCM=1000), idcol="type")
         type     n     mean   median   sd.abs   sd.rel   se.abs   se.rel    skewness    kurtosis
       <char> <int>    <num>    <num>    <num>    <num>    <num>    <num>       <num>       <num>
1:   weighted    25 2896.036 2884.460 240.2228 8.294885 48.04457 1.658977    1.342018    4.387913
2: unweighted    25 2950.818 2884.460 281.6433 9.544584 56.32866 1.908917    1.342018    4.387913
3:        MCM    25 2950.958 2885.677 294.7178 9.987188 58.94355 1.997438 1284.733078 4742.367336
> rbindlist(calc_Statistics(ExampleData.DeValues$BT998, n.MCM=10000), idcol="type")
         type     n     mean  median   sd.abs   sd.rel   se.abs   se.rel     skewness     kurtosis
       <char> <int>    <num>   <num>    <num>    <num>    <num>    <num>        <num>        <num>
1:   weighted    25 2896.036 2884.46 240.2228 8.294885 48.04457 1.658977     1.342018     4.387913
2: unweighted    25 2950.818 2884.46 281.6433 9.544584 56.32866 1.908917     1.342018     4.387913
3:        MCM    25 2950.848 2885.22 294.8593 9.992357 58.97185 1.998471 12750.243313 47030.735947

Note that skewness and kurtosis in the MCM case grow with the number of iterations, while the other values remain aligned to the weighted/unweighted figures.

[RLumSK]

Thanks @mcol I've merged #123; note to @coffeemuggler