Forums › English Language Forums › General › General Discussion

Search

Expected value of Forging at Mid amount (warning, math ahead)

8 replies [Last post]
Mon, 08/26/2013 - 19:02
Mohandar's picture
Mohandar

I've been doing some back of the envelope calculation regarding crafting with the forge, and it seems to suggest that choosing the Mid value may actually be the best long term choice in conservation of crystals. Typically, the trend for crystals required is 1X for Low, 2X for Mid, and 3X for High.

Assuming Mid has a 70% chance of succeeding (the most costly ones at least), the expected value (number of crystals needed) looks like this:
0.7*2X + 0.3*0.7*4X + 0.3^2*0.7*6X + 0.3^3*0.7*8X + ...
where each term in this expression represents the probability of success on try number 1, 2, 3 and so forth. Mathematically, it would be:
summation[n, from 0 to infinity, (0.3^n)*0.7*(n+1)*(2X)]
Wolfram Alpha states this summation converges to 2.857X, which is less than the 3X required of the High forge cost.

It can also be expressed more generally as (1-P)^n * P * (n+1) * (2X) where P is the probability of a successful forging; for an 80% success rate, this converges to 2.5X.

These calculations don't take into account the fact that you could win a very valuable Prize Box for using the High amount though. Anyone want to back up my math and assumptions on expected value?

Mon, 08/26/2013 - 20:14
#1
Broxaim's picture
Broxaim
yes you have the maths

yes you have the maths correct, but there was a much quicker way of working it out

2 / 0.7 = 2.85714... which is just working out the expected value
and assuming 30% is the low option you would have 1 / 0.3 = 3.33333
but at 40% it's 1 / 0.4 = 2.5
so yes the mid option does seem like the best one if it is 70%

however I have noticed that it's not always 70% for the mid value

but i just want no hassles and a chance at the forge box.

Mon, 08/26/2013 - 20:25
#2
Juances's picture
Juances

2.857X = you'd only save like 5 crystals at the last level and miss the chances of a box.

Really, the numbers aren't big enough to make a difference for most of the process. Even at 2.5, only the last 3 levels might be worth the risk.

Mon, 08/26/2013 - 23:31
#3
Heroia's picture
Heroia
Why...

My head now hurts. It's still summer, I shouldn't be doing this.

But still I have to credit you on a great formula. Maybe OOO should take a look at it...

Mon, 08/26/2013 - 21:27
#4
Broxaim's picture
Broxaim
@Juances true, you'd only

@Juances

true, you'd only save 5 crystals at last level, maybe 15 crystals total.
and not really worth it

Mon, 08/26/2013 - 21:27
#5
Skyguarder's picture
Skyguarder

I bet this is Grade 12 math. :P

Tue, 08/27/2013 - 02:13
#6
Zeddy's picture
Zeddy

My guildmate did an 80% forge 49 times in a row.

They all failed. All of them.

If the percentage display wasn't a complete and total lie, the mid-option would probably be pretty wise, yes.

Tue, 08/27/2013 - 05:07
#7
Mohandar's picture
Mohandar

Did it actually succeed on the 50th try? I know people have reported that the Mid option actually succeeds sometimes, but 49 times suggests there is a bug even with only one sample. Those odds are statistically improbable. Perhaps it is a bug in the code or the way it fetches random numbers that causes you to fail repeatedly (i.e. doesn't actually reroll if you don't exit the Forge screen)?

Tue, 08/27/2013 - 05:10
#8
Zeddy's picture
Zeddy

You know, I forgot to ask.

Powered by Drupal, an open source content management system