To work out the value of a thief at each level, I:
(1) Calculated his chance of success at each different hijink (10% - 35%)
(2) Calculated his chance of being caught at each different hijink (5% - 25%)
(3) Calculated his wages
(4) Calculated the syndicate’s fee if he succeeded in each hijink ($95 - $1,300)
(5) Calculated the syndicate’s expected return from success (% from step 1 x value from step 4), per hijink
(6) Calculated the syndicate’s damages if a ruffian is captured in a hijink*
(7) Calculated the syndicate’s expected loss from capture (% from step 2 x value from step 6), per hijink
(8) Calculated the syndicates return in each hijink as (expected return from success) - (wages + expected loss from damages).
(9) Disregarded any hijinks that would have a return of 0, and averaged the return of all other hijinks
*Note on 6: This in turn cascades down into another set of subtables for each type of hijink which modles out the expected damages at each level of conviction or acquittal, for each time of crime, then sumps of the likelihood of each level of conviction x expected damages, for each type of hijink. For punishments that caused maiming, I calculated the damages as equal to the cost of magical curing. I then used estimates of attorney and bribe expenditure at each level. This resulted in a net expected cost if caught for each crime, ranging from $51.03 for drunkeness to $4,174.52 for treason.
I did this for every combination of hijinks at every level. This took a long time. How did this work out in practice?
At 0th level, the result was 14gp expected return, 6.89gp expected loss, and 6gp wage, for a total of (14 - 12.89) 1.11gp, which I rounded to 1gp.
At 1st level, the result was 33gp expected return, 3.45gp expected loss, 25gp wage, for a total of (33 - 28.45) 4.55gp, which I rounded to 5gp.
It was then time to use this data to apply it against the thieve’s guilds themselves. I first calculated the number of thieves of each level likely to be present in each class of city. I then inputted the revenue (from above) for each thief of each level and was able to work out a total value for the thieves’ guild of each city.
I then assumed that the thieves’ guilds were organized into syndicates with 5 underbosses of each tier at each level, and adjusted earnings accordingly. The result was the charts in the book.
An astute accountant might say “a thieves’ guild that ran itself specifically to make money by doing only those hijinks which had just the net highest value at every level, rather than a mix of ALL profitable hijinks, would make more than this!”
But in the time frame ACKS covers, there were no astute accountants. There was no excel, and no double-entry book-keeping, and hardly any understanding of budgeting. To assume that a criminal enterprise would be run at optimum, excel-based efficiency seemed, to me, to be ludicrous. So I didn’t do that.
A ruthless PC thief might say “As boss, I don’t think its worth worth paying for attorney’s fees or bribes for my failures, and I certainly would’t worry about treating their wounds!”
I would respond that this is certainly the PC’s prerogative. But in actual play, such PCs have quickly run into a shortage of thieves; moreover, if one tracks XP (as we did in my playtest), one realizes that abandoning low-level thieves to their fate soon leads to no mid-level thieves advancing. So, in actual play, PCs tended to do everything they could to keep their thieves alive and out of jail. So, based on that, I assumed NPC bosses would do the same.
Hope that makes sense!