Heist Math

In another thread there were claims that heists can break 1.80 EPA, and I’m trying to confirm that.

Here’s what it takes to go on a heist: (no big rat; see later)
5 actions with Gang of Hoodlums to get 18 CP of casing
3 actions with The Decoy to gain 9 CP of casing (and a bundle worth ~20 pence)
1 action to choose a heist
2 to purchase 2x Inside Information
1 action to start the heist
3 actions with ideal card draws to get 5 Burglar’s Progress
1 action to finish heist
Some amount for reducing suspicion later (let’s ignore for now)

Total: 16 actions. That would require a 28.80 echo reward to be 1.8 EPA. (16 x 1.8 = 28.80)

There are a few optimizations we might be able to do, though. The Big Rat can reduce the number of actions to get our necessary casing by 5, at the cost of 7.20 echos (and being a horrible bastard who murders poor rats).
New requirement: 11 x 1.8 + 7.20 = 27 echo reward needed for 1.8 EPA. Oh dear.

Well how about forgoing inside information? This will increase risk though and even ideal cards will only give 1 burglar’s progress per action (barring a handful of luck checks that occasionally give 2). With no big rat, 5 less actions at the start, 2 more in the actual heist. With a big rat, 3 less actions at the start and 2.40 echoes saved, and 2 more actions in the actual heist
New requirement (big rat used): 10 x 1.8 + 4.8 = 22.8. Well, that one’s possibly reachable, but we’re taking much more risk! We won’t actually get ideal cards all the time.

Regardless let’s look at rewards now:
A Sealed Archive gives 20 echoes on a regular success, 62.5 on a rare. The chance of the rare was nerfed sometime last year, and no one seems to have solid numbers on what the new chance is. I’m going to assume 10%, though it could be lower. This averages to 24.25 echoes.

Mr Baseborn’s Papers gives 27 echoes on a success with no rare success possible, however this will require 2 more actions. Probably not worth it unless the rare success on Sealed Archive is actually 5% or less.

Of note is the Snaffle Documents option from the A Clean Well Lighted Place card. It gives 4.90 echoes at a 50% chance, so a 2.45 EPA action. However, it’s only safe to play if you already are near the end of your heist with no lost cat-like tread, and doesn’t show up all the time. I’m not sure it bumps up average EPA all that significantly, given the large number of overall actions in this grind.

All in all, 1.8 EPA seems dubious since I don’t think the cards will comply often enough and a 10% rare success seems too optimistic. Am I missing anything?
edited by Kaijyuu on 2/23/2017
edited by Kaijyuu on 2/23/2017

  1. There’s no risk w/o inside information, sometimes even 1 will overkill the heist due to luck and good cards.
  2. Snaffle Documents is 5.5 and offers 1 progress, the increase from them is ~4.3%
  3. Even on a bad run with 4C I used 7 cards(no cat), but didn’t end in jail
  4. Convert TLS and convert CC from the success

Sorry for the short reply, but I’m very busy!

You can use the &quotSteal Paintings for the Topsy King&quot to get 3 CP of Casing for one action. That plus Gang of Hoodlums gets you enough casing to buy one Inside Information and start the heist. You’ll spend one more action getting burglar’s progress up inside the heist, but overall your action investment reduces to 14 (Though Big Rat is probably still better). The chance of rare success on the sealed archive is still a big question mark, though. If it’s as high as 15%, 1.8 epa is very doable. If it’s as low as 5%, then as you say, Mr. Baseborn’s Paper becomes the superior option and recalculation is in order.

It’s also worth noting that optimized heist running assumes a 5 card hand, and with that probabilities become very much in your favor. With 5 cards and 1 Inside Information, I never find myself risking failure, which means I can almost always try for Snaffling when it shows up (at least until I fail the luck check).
edited by Kaigen on 2/23/2017

About snaffling, I do it pretty aggressively. I snaffle about 55% of my heists at least once, netting an average of 1.35E per heist taking into account the actions to shed suspicion for failures.

I will note as well, however, that I’ve rechecked my EPA calculations and found that I was undercounting suspicion shed actions by a little bit. My empirical EPA over my last 150 heists is 1.76 and not 1.8 as I reported in the other thread.

You blokes might want to consider the opportunity cost of having to draw cards as Heist cards too. Also, don’t forget that the way to end a Heist is to get that damn &quotAbundant&quot frequency card - which sometimes just refuses to come up because RNGeezus told it not to. These two issues are the main problems with Heists.

I won’t recommend Heists as a way to get profit. I would recommend if you want to accumulate Searing Enigmas, but for general profit? No.
edited by Rostygold on 2/25/2017

I have to disagree with both of those. If you have exceptional friendship (and thus a 10 card deck), you can go into the heist with 5 cards, which will double to 10. Inside the heist, I average 4 actions and 9 draws, which means if I wait 20 minutes to get my deck up to 3 before using the prize card, I then have 10 cards upon exit, netting 3 card draws (enter heist with 5 in deck, generate 2, end up with 10) over what I would have had otherwise.

As far as drawing the prize card goes, you can dramatically increase your chances of getting it by using multiple cards in a row without drawing and then drawing to a full hand with 5 progress. For example, I save my inside information until the end and will generally use a 1 progress card and a card that benefits from inside information to get all the way up to 5 and then draw 2 cards. It makes a significant difference in how often you need extra draws. You can also mitigate the badness of failing to draw the prize card by using the &quotmoment of safety&quot card (more catlike tread means less suspicion when you leave the heist) and/or getting further progress so you can use the 7 progress option (slightly more profitable on average, even with TLS conversions taken into account).

It is disappointing that it comes in slightly less profitable than just farming souls as a spirifer despite involving much more skill and preparation. I think the SE drop rate could safely be tuned up a bit without being overpowered.
edited by Wiwo on 2/25/2017

I am the only one who think heists as a good way to have a correct epa while getting a good amount of making waves ? ( 1 progress = 2 MW )
I am also an exceptionnal friend, so as explained by wiwo, it might impact my perception of its rentability.

Or there is a better epa grind with the same amount or more making waves ?

The War of Assasins and getting a collection of curiosities(you need the Bandages and rust card at least once).

Unless I am higly mistaken, the war of assassin doesn’t give any making waves ?

The real question here is what the new chance of a searing enigma is.

From long experience, having used Heists to gain about 20 of the 49 Seating Egnimas needed for an Impossible Theorem, I think it’s about 10%.

There have been several threads before on Heists and rare success estimates but I’ll link to my work on calculating out EPA in a previous thread here:
http://community.failbettergames.com/topic22401-best-ways-of-earning-echoes.aspx?Page=1#post146292

Edit: concerning getting over 1.8 EPA: if, and a big big if here, you manage exactly 13 actions every heist (which is the min, I’m pretty sure, from how I calculated it - 9 pre-heist, inside information (2), knifers (2, 50℅ luck check), any 1 progress option, 1 to complete ) you will get 1.87 EPA assuming the 10℅ chance of a Searing Engima, without accounting for Snaffling Documents at all. I’m not even trying to deal with the suspicion gain, however.
edited by Parelle on 2/26/2017

My log frog last year, starting in september - 183 heists, 2745 actions spent, 4397 echoes net gain (both actions and net gain include cost of cleaning my suspicion using ablution solutions), 11 searing enigmas out of it (6% rate). I never took the higher option (baseborn’s papers) during these heists, explicitly because I was interested in the enigma rate.
That’s “between 2.56% and 9.45% at 95% confidence level”, btw.

If I take Parelle’s numbers to mean “20 successes in about 200 attempts”, then we have combined 31 successes in 383 attempts (8.09%), and “between 6.04% and 11.75% at 95% confidence level”.

There’s another player on the http://fallenlondon.wikia.com/wiki/...a_Sealed_Archive page claiming 8 successes in 152 attempts. Now the combined total is 39 in 535 (7.289%) and “between 5.086% and 9.49% at 95% confidence level”.

I believe we need a lot more data. Tightening the range to 1% at this confidence level will require at least 1900 samples (if it’s 5%; more if it’s larger) …

Ah. Don’t count my numbers in for mathematical fact! That was my best guess after previously seeing the 25% mark thrown about.

But here’s some fairly concrete results from last summer:
http://community.failbettergames.com/topic22283-heists-are-fixednerfed.aspx

Explain how are you calculating your suspicion decrease, if you don’t mind?
edited by Parelle on 2/26/2017

We should probably just ask failbetter what the new searing enigma chance is on heists. I’d like to also ask why they thought it necessary to nerf the good grinds but that might be rude.

They are tweaking things all the time.

Sure, some grinds got nerfed. But it’s easy to overlook that some things were also made much easier (just consider how gaining Notability used to be at first.)

So I should only use &quot5 successes out of 35 tries&quot for your data?
(everything else on that thread seems to be either &quotX tries until first enigma&quot which is biased data; or the calculator sheet which was contaminated by old data, and has no reported samples for the new version)

That’s then 11+5+8 out of 183+35+152, 6.486%, and &quotbetween 3.976% and 8.99% at 95% confidence&quot.

I wrote down how many solutions I actually used. Each use is +1 actions and -1.5 echo.

Be advised that a new row has been added to the Calculations Spreadsheet. Do please add your own results to the “Multiple successes” tab. I’ve put in the last couple of weeks’ worth of my data, but we should need a full 220 trials or more to differentiate between 5% and 10% with 95% confidence.

OK. I have now added my 172+11. And only mine.

Which means at this point we can say that the data is highly suggestive of a drop rate of roughly 5%.

I’ve been trying out some heists myself, getting to 7 progress and snatching mr baseborn’s papers. So far it looks like it’s slightly less profitable than my standard grind… however, it gives 20 cp shadowy at the end of every heist. Since Shadowy is easy to atrophy, doing a few each week will allow me to continue weekly trips to the nadir, thereby bumping the whole thing up to profitability.

EDIT: So I just finished an awful run where I ended up with 1 cat like tread, 6 progress, 4 risky cards and the prize achieved card. Decided to cut my losses and just take the sealed archive… and got a Searing Enigma.

Damn capricious RNG.
edited by Kaijyuu on 3/3/2017