Why is 16% the most common luck number for strategic templates?

What makes the number 16 so special? Why not make it 10% of 20% or something else?

I know it's not as big anymore as no luck is much more common but 16% has always had a large presence in ladder templates and whatnot. Anyone got a good explanation about this or is it just by convention that it's used?

I vaguely remember the idea that analysis attack graph for WR is only accurate within 1% so it could be more like 99.5 or whatever. Dunno if that is true or I am confusing it with something else could easily be the latter.

Rikku is correct. The analyze graphs are an estimate and not 100% accurate.

Fun fact: When I first calculated the highest value that 4v2 would always succeed, I made a mistake and came up with 18%. For a while, this was the value used in strategic games, You can still see this if you dig back into old 1v1 ladder games.

4v2s would succeed almost every time at 18%. But maybe one in 100 games, one attack would have a 4v2 failure somewhere, and that player would rage.

I might be misunderstanding how luck works but I thought 17% luck meant 4v2's failed .02048% of the time, still rarely, but much more common than knyte's figure.

I'm not sure whether I'm right or Knyte is, but as I understand it:

A 4v2 is 4 separate attacks with a 60% success rate.

So long as one of them succeeds you have this:

Expected kill 2.4 Actual dice rolls 1

Difference 1.4 * .17 = .238 2.4-.238=2.162

Therefore always a success, since then WR comes into affect and rolls another dice to decide if itd kill 2 or 3 armies, 16.2% chance to kill 3, 83.8% to kill 2.

If you fail all the rolls though you get:

2.4 * .17 = .408 2.4-.408=1.992

Which I believe means a 0.8% chance to kill 1 army wenever you fail all 4 attacks.

The chance to fail 4 attacks being

.4 * .4 * .4 * .4, thus the total chance being .0256 * .008 = .0002048, or 0.02048%

Huh, I was just using a script made by Math Wolf some time ago that should be doing the same math as Nauz. I'll edit this later with details, but I probably just misused the script.