Luck is one thing ... this is just bad

[GielenDJ]

Last game i played 23 thows .. 7 doubles. A 5.46% rate.
These rates are more often then not in one of both players. It determins the win in alot of the games.

Sorry, but if you call this a good algorithm, you never played backgammon with real dice!

FIX

I guess your random seed algorithm just fails!

[Kevinbga]

Maybe we should create an official topic where players get a free pass to rant about bad luck.
It would be a safe space for players to vent without getting eye-rolled.

[GielenDJ]

I know that this is a comon topic. But if you look at the stats ... look at the numbers. It's just not adding up anymore!

Yet another game 18 - 5 doubles.

In all cases i have in mind alot 6's.

Why bother showing stats, if you cant at least mention the weird behaviours against the odds!

[Waikirikiri]

Last game i played 23 thows .. 7 doubles. A 5.46% rate.
These rates are more often then not in one of both players. It determins the win in alot of the games.

Sorry, but if you call this a good algorithm, you never played backgammon with real dice!

FIX

I guess your random seed algorithm just fails!

Your sample size, 23 throws, is far too small for a valid statistical analysis.
What you have is a sequence of throws that don't make for a valid analysis.
Come back when you have the results for 1,000 or better 10,000 throws.

Meantime play or not play it is your choice.

[Gulchen]

Python 3.8.0 (default), on linux:

Code: Select all

>>> from math import comb
>>> histogram = [(5**(23-d))*comb(23,d) for d in range(24)]
>>> sum(histogram) == 6**23
True
>>> threshold = histogram[7]
>>> histogram.count(threshold)
1
>>> (sum(entry for entry in histogram if entry <= threshold))/(6**23)
0.09001239489514781
>>> histogram = [(5**(18-d))*comb(18,d) for d in range(19)]
>>> sum(histogram) == 6**18
True
>>> threshold = histogram[5]
>>> histogram.count(threshold)
1
>>> (sum(entry for entry in histogram if entry <= threshold))/(6**18)
0.20580943877599434
>>> 

[Romain672]

Last game i played 23 thows .. 7 doubles. A 5.46% rate.

And yet, it has a 7.5% chance to happen: https://docs.google.com/spreadsheets/d/ ... 1846324187 .
So it should happen once every 13 games.

Yet another game 18 - 5 doubles.

This one has a 16.8%. It should happen in average once every 6 games.


So those games are not likely, but they should happen very regularly.

[GielenDJ]

correct, but not almost everygame.

Soit, doesnt matter. It seems normal, so as suggested, play or not to play

[Remkar]

correct, but not almost everygame.

Soit, doesnt matter. It seems normal, so as suggested, play or not to play

But your lifetime stats show an average of:
23.65 rolls per game
4.07 doubles per game

So you're overall rolling slightly more doubles than you'd expect. So... it doesn't seem likely to be happening "almost every game."

[Gulchen]

From your 10 most recent games, I get

Code: Select all

15,70	high
07,37	high*
10,46	high
10,41	high
05,47	 low
08,48	equal
06,34	high*
07,50	 low
10,45	high
06,54	 low

, ​ where high* indicates that one fewer double would've made it low instead.

[GielenDJ]

Take also the opponents when you take those of me