Grinding for the Un-Maxed Player

Fallen London
1

I have done some grinding with a character who is not yet maxed but when I looked around this forum I noticed that all of the comprehensive calculations and analysis where for capped players. I would guess that many players have a Un-Maxed Player (maybe an alt) so this seems a little strange.

In this post I will look at one small situation: grinding echos through casing. The character in question grinds casing using their gang of hoodlums or well-planned villainy. Then the casing is converted to Antique Mysteries and then sold.

I wrote a calculator that finds the average echos per action of a character (capped or uncapped) who is grinding echos through casing. Here is the link: http://benjaminhelkey.weebly.com/grinding-echos-through-casing.html. (I don’t know if/how I can post the calculator to this site)

Example:

Has anyone else thought about grinding with an Un-Maxed Character?
[li][/li][li]
edited by Roas on 5/18/2014

2

Serious grinding tends to make an unmaxed player into a maxed player pretty fast. People just assume you are using an action that you have 100% chance of success for grinding, that is you pick the most profitable thing available to you that you can do 100% of the time.

I actually would not recommend grinding with a Gang of Hoodlums if you have not maxed Shadowy yet, since you get Shadowy CP from every skill check, and for the Gang you will have one check per 5 actions. Much better to do the Box and get CP from every action (although more clicking that way unfortunately).

3

I don’t think that Casing is the most efficient money making grind. As you mention a Gang I presume you are at POSI level. So now that Collections of Curiosity can be traded to the Tomb Colonists the War of Assassins can be profitable and there are options that use different stats so you can work on getting stat capped at the same time.

4

i rarely read about anyone using casing to grind echos, it looks more useful for grinding connections or specific resources.

i’m of the opinion that the university invesitigations/flit trick and hunter’s keep(especially with the rare success reward) are still the most profitable in terms of echos/action , not including options that are not easily repeatable.

for grinds that don’t really need a skill/luck check, you could raise your pygmalion high enough in mahogany hall, so your easiest pygmalion challenge would become straight-forward. gives a half decent 1.07 echos/action AFTER you raise pygmalion high enough.

alternatively, you could buy cheap diamonds with cryptic secrets in the sidestreet for about 1.12 echos/action profit?? i don’t remember the numbers. it also has a small chance of giving an extra higher quality diamond.

5

140p/action, actually. E4.00 of cryptic clues for E5.40 of diamonds, plus a rare magnificent diamond. On its own, it’s one of the more profitable grinds, though it does start to lose its appeal once you add in the 3-4 actions spent getting those clues.
ETA: I just ran the numbers. Assuming you get the clues by fortune-telling, which should be reliable enough for non-maxed players, the whole grind does come to about E1.12/action. Of course, this is also ignoring the chance of a magnificent diamond.

I’ll second the University/Flit exploit; I think I calculated it as up to E1.80 per action, depending on luck and how valuable sudden insights are. I’m also having good luck with the Affair of the Box despite not being maxed out, Hunter’s Keep ditto.
edited by Raymond R Price on 5/22/2014

6

Even at 140 ppa, the empire adornments diamond bag is a bad deal considering there’s no skill gain. You might as well get 140 ppa through the Box and skill up shadowy.

7

I’d appreciate a more in-depth explanation of this. I’ve often seen the University/Flit way mentioned as providing 177-180 ppa, but I can only reach those number, as you say, &quotdepending on luck&quot, which is not a useful calculation for something you repeat multiple times. You need to either have 100% success rate or to account for the probability of failure.

You can raise Investigating in the university to 5 in 5 actions (earning ~50 Cryptic Clues each).
But then you only get 70% success in the Flit to convert to 5.4 echoes worth of items.
If you succeed, then you’ll indeed would make a total of ~10.4 echoes in 6 actions = ~173 ppa.

But at 30% of the times you’ll fail.

If you try make the challenge 100% successful, you’ll need to raise Investigating to 8, which takes 7 more actions (again, in each earning ~50 Cryptic Clues). So you’ll end up with a total of ~17.4 echoes for 13 actions: ~133 ppa.

Is there a more efficient way to go about this as a grinding method with an expected return of 170-180 ppa?

8

[quote=dov]
Is there a more efficient way to go about this as a grinding method with an expected return of 170-180 ppa?[/quote]

The most efficient way to go about similar challenges is usually to do them as soon as possible, even though it means lower than 100% success rate. The linear success rate increase WRT quadratic CP needed increase just tends to work like that in this game (and the concrete numbers set for the challenges).
In this case, failure costs 5 cp investigating, so you usually need 2 actions worth of university (at ~50 clues gained each, too!) to get back to where you can attempt again.
Chance of success at Nth attempt: P[N] = 0.3**(N-1) * 0.7
Profit when succeeding at Nth attempt: x[N] = 540 + 5*100 + (N-1)*200=840+200N
Action cost when succeeding at Nth attempt: a[N] = 5+(N-1)*3+1=3+3N

Assuming that anything past M attempts is a fail (you just give up, because the potential profit from 540 pence/(3+3M actions) multiplied by P[M] is so small that it doesn’t meaningfully affect the expected profit estimate), you have
E(ppa)=(sum(P[N]*x[N], N=1…M) + AA)/(sum(P[N]*a[N], N=1…M) + BB) // I am taking a huge pool of actions here, and distributing various instances of “succeeded at N-th attempt” between these actions proportionately, scaling to the simplest pool size that came to mind (I could also do "out of 1000000 actions, how many are part of success-at-1? how many are part of success-at-2?, etc …, then divide by 1000000)
and, of course, the real E(ppa)=limit(above, for M->infinity)

BB= sum(P[N]a[M], N=(M+1)…inf represents all actions lost to “failed M attempts, reset to zero and tried again”
with the corresponding profit from university grind being
AA=sum(P[N](500+200*(M-1)), …), because each fail except the last is followed by two actions grinding.
The trick here is that M doesn’t depend on N and Q=sum(P[N], N=(M+1)…inf) = 1-sum(P[N], N=1…M) is easily computable, so:
BB=Qa[M], AA=Q(200M+300)

I put this in a simple script to print values for a few low values of M, and see that the estimate quickly raises from 146 to around 155 ppa, and stays there up to M=990 (e(ppa)=154.5098039…). Higher M causes the script to fail on a recursion depth excess.
That’s not bad, but it isn’t broken. Affair of the box on the side that gives correspondence plaques is 157.6923…ppa, so … no reason to keep university open infinitely here.
(it’s actually slightly better, because not all failures cost 2 actions, but that doesn’t matter until attempt #5, which means only 0.81% of all “runs” are improved by it, and such runs are also already so long that success doesn’t improve your ppa much above the baseline from just grinding investigation; it will not be good enough to overcome AotB)

9

Actually, my action costs are 1 off (you only grind 4 actions before going to flit; also affects x[N]), and you have 1 extra cp from that, so the “not all failures cost 2 actions” matters at attempt #4 already (2.7% of all runs! Still won’t improve ppa by more than 0.02717/14(subbstituted from later calculation: 163.16)~~5.3ppa, and that’s an upper bound, not a good calculation) … recalculating … now it looks as 163.181818… ppa
That may be worth writing home about, but not worth throwing away the opportunity to breed a bifurcated owl (the other way to get 4 BRD pet costs over 20k echo … and is fate locked too).

10

I have posted these numbers elsewhere, but why not here as well? End result is 178 ppa. I stopped counting when the low probability yielded so low ppa that it wouldn’t be accounted for due to rounding.

Interview the Department of _______ staff (100% at Watchful 177 apparently): Investigating… +4 CP, 53 x Cryptic Clue (1.06 Echoes)
Raid a Message-drop: 70% chance of 270 x Cryptic Clue, 30% risk of Suspicion +1 CP

0.7 x (4 x 1.06 + 5.4) / (4 + 1) = 1.3496 Echoes/Action
0.21 x (5 x 1.06 + 5.4) / (5 + 2) = 0.321 Echoes/Action
0.063 x (7 x 1.06 + 5.4) / (7 + 3) = 0.080766 Echoes/Action
0.0189 x (8 x 1.06 + 5.4) / (8 + 4) = 0.021861 Echoes/Action
0.00567 x (9 x 1.06 + 5.4) / (9 + 5) = 0.0060507 Echoes/Action
Total: 1.7792777 Echoes/Action ~178 ppa

Depending on what you usually do you may want to account for getting rid of Suspicion from all those 30% failures.

I enjoyed an even better 198 ppa by ‘dealing with mirror-smugglers’ in the third coil of the Labyrinth of Tigers (100% at Dangerous 180, apparently), and then using The Hunt is On! to kill the goat demon in the Flit. But after &quotA night for escapes&quot my making progress in the Labyrinth of Tigers hit 13, and I was ‘under orders’ and &quotThe third coil&quot storylet disappeared. So no more The Hunt is On! +3 CP together with 108 x Rostygold for me. Not unless that storylet is somehow accessible again when I return from the fourth coil. But that would be unprecedented. Right now I wish I hadn’t progressed so soon.[li]
edited by Gillsing on 5/22/2014

11

[quote=Gillsing]
I enjoyed an even better 198 ppa by ‘dealing with mirror-smugglers’ in the third coil of the Labyrinth of Tigers (100% at Dangerous 180, apparently), and then using The Hunt is On! to kill the goat demon in the Flit. But after &quotA night for escapes&quot my making progress in the Labyrinth of Tigers hit 13, and I was ‘under orders’ and &quotThe third coil&quot storylet disappeared. So no more The Hunt is On! +3 CP together with 108 x Rostygold for me. Not unless that storylet is somehow accessible again when I return from the fourth coil. But that would be unprecedented. Right now I wish I hadn’t progressed so soon.[/quote]
[li]

Is the goat demon in the Flit the best use for The Hunt Is On… ?

[/li]

12

[quote=Valerio De Camillis]Is the goat demon in the Flit the best use for The Hunt Is On… ?[/quote]Yeah, I ‘only’ got to 193 ppa for killing the rat-brigands. The goat demon gives Wounds +2 CP on failure instead of just +1 CP, but given the amount of actions it takes for each cycle, that difference ends up being rather inconsequential. (The Wounds/Action down there is total wounds, not the difference between the goat demon and the rat brigands. Since the total wounds ended up being inconsequential I didn’t even bother doing that calculation for the rats.)

Kill the Rat-Brigands: The Hunt is On! 9
15 x 108 Rostygold + 810 Primordial Shrieks or The Hunt is On! -9 CP and +1 Wounds

0.7 x (15 x 1.08 + 16.2) / (15 + 1) = 1.4175 Echoes/Action
0.21 x (18 x 1.08 + 16.2) / (18 + 2) = 0.37422 Echoes/Action
0.063 x (21 x 1.08 + 16.2) / (21 + 3) = 0.10206 Echoes/Action
0.0189 x (24 x 1.08 + 16.2) / (24 + 4) = 0.028431 Echoes/Action
0.00567 x (27 x 1.08 + 16.2) (27 + 5) = 0.008037225 Echoes/Action
Total: 1.930248225 Echoes/Action

Kill the Goat Demon: The Hunt is On! 11
22 x 108 Rostygold + 1188 Primordial Shrieks or The Hunt is On! -11 CP and +2 Wounds

0.7 x (22 x 1.08 + 23,76) / (22 + 1) = ~1.446260 Echoes/Action
0.21 x (26 x 1.08 + 23,76) / (26 + 2) = 0.3888 Echoes/Action +0.015 Wounds/Action
0.063 x (30 x 1.08 + 23,76) / (30 + 3) = ~0.10721 Echoes/Action ~0.007636 Wounds/Action
0.0189 x (33 x 1.08 + 23,76) / (33 + 4) = ~0.03034 Echoes/Action ~0.00306 Wounds/Action
0.00567 x (37 x 1.08 + 23,76) (37 + 5) = 0.0086022 Echoes/Action ~0.00108 Wounds/Action
Total: 1.9812 Echoes/Action (~0.02678 Wounds/Action)

13

Great, thanks for the analysis! Looks like i’ll be in that coil of the labyrinth for a long time. :)

14

xKiv and Gillsing,

Thanks a lot for the detailed analysis and explanation. This is exactly the math I was looking for :-)

15

@Gillsing - I am pretty sure your math is incorrect - you are computing the average ppa of an entire run, irrespective of how many actions it takes. Exxagerated example of what’s wrong with that:
Let’s say a year has 365 days. Let’s split that in two “runs” - one 364 days long with a gain of 728 “coins” (two per day), the other 1 day long with a gain of 1 coin. This repeats every year, so they each have equal probability of happening: 50%.
By your “maths”, your average coins per day = 0.52+0.51 = 1.5.
But it’s actually (7280.5+10.5)/(3640.5+10.5)=(728+1)/(365)=1.99.

In the case of uni-flit, you want
( – total echoes gained over X actions
0.7 x (4 x 1.06 + 5.4) +
0.21 x (5 x 1.06 + 5.4) +
0.063 x (7 x 1.06 + 5.4) +
0.0189 x (8 x 1.06 + 5.4) +
0.00567 x (9 x 1.06 + 5.4)
)
/
( – and here’s how you get X, which is expected number of actions taken for some unknown amount of (successful) attempts
0.7 x (4 + 1) +
0.21 x (5 + 2) +
0.063 x (7 + 3) +
0.0189 x (8 + 4) +
0.00567 x (9 + 5)
)
= 1.71848839… echo/action (compare to your result of 1.7792777)

Except you should also account for unsuccessful attempts, which means also calculating:
A1 = # of actions spent on “unsuccessful” attempts = (1-(0.7+0.21+0.063+0.0189+0.00567))(9+5)
A2 = # of those that were spent in uni = (1-(0.7+0.21+0.063+0.0189+0.00567))(9)
P1 = how many echoes you got from those = A2*1.06
And then you add P1 to nominator and A1 to denominator
= 1.712549…

And then the only thing you know that the correct answer is somewhere between those two numbers (the first one is too high because it ignores some failures, and the second is too low because it ignores that those failures don’t set you back all the way to zero). Which is cool, because the difference between them isn’t even an entire ppa.

16

[quote=xKiv]@Gillsing - I am pretty sure your math is incorrect …[/quote]Yeah, that seems to be the case. Thanks for the correction!

17

I just hope mine isn’t wrong too.
(and you went for more correct uni ppa, and for more correct number-of-actions-for-success-at-Nth-flit)

18

Damn, I stumbled across this article just a few minutes after I did &quotA night for escapes&quot
edited by Bones on 7/1/2016