User:ParadiseAdi

Hi, since I investigated the game quite thoroughly, I publish my findings here. The admins are free to incorporate the content into the corresponding main pages.

Profitability/Statistics
Currently only published as comment.

Cheating
A fresh (unplayed) savegame is available. Download (Password: aN45dAfgN4)

Money: 200'000'000

Piastres: 1000

Staff: 1000

Energy: 10000

Starting Level 39 (Optional)

I can't explain, how to get the file onto the phone, but it has to be rooted. Rename the desired file to profile0.sav and copy it to /data/data/com.seventeenbullets.android.island/files. If you want, you can also backup your old profile. Make sure, there is no other profileX.sav in the folder.

I will soon publish an "unlimited" version which can be installed on rooted and non-rooted phones in addition to the original game with all (yes, also the time-limited ones) ;-) buildings unlocked. You can do with it whatever you want, but it may ruin the fun…

Algorithm
I found the algorithm by reverse-engineering the code. So this is the definitive answer:


 * isBroken is invoked after every collection (click on yellow/green buck) of the money
 * repair_counter counts collections until the building is broken which resets the counter
 * randomInt(a,b) returns an integer x: a <= x < b

Mathematically, the average probability should be 6.339 %, but when simulating the code above, the value becomes stable at 7.236 % which is pretty close to the observations. I guess, this is because of the implementation with a counter and new randoms with every iteration.

Note concerning the repair formula
The formula on the Profitability page is:

RealIncome = (14 * NominalIncome - RepairCosts) / (14 * IncomeHours + RepairHours)
 * for P[Building is broken] = 1/14

which is not correct. The formula would imply p = 1/15. Explanation: Consider two event types: repair (the repair symbol appears) and collect (the green(!) buck appears). 1 of 14 events is the repair event and the other 13 events are collect events. Therefore the formula needs the factor 13 instead of 14.

It is clear when you use the inverse(?) probability:

RealIncome = ( (1-p)*NominalIncome - p*RepairCosts ) / ( (1-p)*IncomeHours + p*RepairHours)

For those who still don't believe, consider a simple example:

NominalIncome = 1000

IncomeHours = 5

RepairCosts = 500

RepairHours = 4

The real income can be directly calculated with the formula above:

( (1-1/4)*1000 - 1/4*500 ) / ( (1-1/4)*5 + 1/4*4 )

now multiply the whole formula by 4 and you get:

(3*1000 - 500) / (3*5 + 4)

Voilà. Still don't believe me? Ask someone else. ;-)

How to correct the values in this Wiki: Just use p = 0.07236 and the formula.