BGonline.org Forums

More results and conclusions on cube magnifies luck

Posted By: MK
Date: Monday, 2 June 2025, at 1:26 a.m.

This is a continuation of my previous post:

https://www.bgonline.org/forums/webbbs_config.pl?noframes;read=215982

Please read it first if you haven't already.

I'll just retype here the last pair of experiments:

4A = 3-ply vs 0-ply checker skill with 0.060 noise, cubeless.
In 10,000 games weaker player won 14.6% of 17,045 total points.

4B = Added 2-ply cube skill to both players.
In 10,000 games weaker player won 18.2% of 29,870 total points.

Now, the following experiments are all variants of 4B
in attemts to further assess the differences between
the various cube skills added to both players. (BTW,
no noise is added to any level of cube skill plies).

5B = Same as 4B but with "cubeful checker evaluation"
again to double-check if and how much it matters. In
10,000 games weaker player won 18.7% of 30,037 total
points. Tentative conclusion is that the tangentially
added checker skill helps 0-ply player a little more.

6B = Same as 4B but with "3 maximum beavers", to see
whether beavers require increasingly more cube skill
or further magnify luck. In 10,000 games weaker player
won 18.7% of 29,836 total points (0.5% more of almost
exactly the same number of 29,870 total points as in 4B).

7B = Same as 6B above but with only 1-ply cube skill
added to both players, instead of 2-ply. In 10,000 games
weaker player won 18.0% of 32,614 total points. Tentative
conclusion may be that weaker player doesn't benefit from
cube-injected luck as much with a lower cube skill added.

8B = Same as 7B above but with only 0-ply cube skill (i.e.
almost none) added to both players. In 10,000 games weaker
player won 18.5% of 31,078 total points. This contradicts
7B but instead shows that when cube is introduced even at
lowest or no-skill level, it magnifies luck.

To recap, the combined conclusion from all these experiments
(with 60,000 plus 50,000 previous = 110,000 games in total)
is that the weaker players wins between 18.0% and 18.7% of
totals points in cubeful sessions vs 14.6% in cubeless ones.

The luck-magnifying effect of the cube is simply undeniable!

If you think 10,000 games for each variant experiment aren't
eanough, feel free to run your own sessions of as many games
as you want and let us know what results you come up with.

MK