Major bug in the card drawing mechanism

Fallen London
1

Cards with 2 (or more) unlocking conditions turn out to be 2 (or more) times more likely to be drawn. Now, I know that this is an outlandish claim, but I’ve procured hard evidence for it, so please bear with me!
Before anything else, I’d like to direct you to this study done previously on the opportunity deck: https://www.reddit.com/r/fallenlondon/comments/50ctcr/card_draw_probabilities_an_exhaustive_study/
It found some highly improbable anomalies in the deck, and - what’s relevant here - all the affected cards have 2 unlocking conditions, and the biggest offenders (SMHS…) have 3.

(for the sake of full disclosure, I should note a flaw in this study: &quoteach card progressed through all five slots of the lodging before it was played&quot - this had a flattening effect on the graph, as any anomalously common/rare cards would spend more/less time in the hand, thus reducing/increasing their draw rate. So, in reality, the anomalies should be expected to be even bigger than what was found out)

Also of relevance is this post by someone who peeked at the source code of the app: https://www.reddit.com/r/fallenlondon/comments/83aocr/opportunity_deck_research/

They’ve ruled out card mislabeling as unlikely, and also described the drawing mechanism in a way that shows how the bug could be happening: presumably for each condition after the first, a card gets duplicated, and thus gets an additional roll.

Of course, so far this is all circumstantial, which is why I made a series of my own Storynexus experiments with the help of Passionario#5970
You can also verify our results by yourself, in Passionario’s Nexus World: Shenanigans | Storynexus.com

Phase One: the control
First Card: standard frequency, no unlocks
Second Card: standard frequency, no unlocks
Sample: 100 drawings to a 1-card hand (2 permanent blocker cards were used)
Results: 56 first, 44 second
Conclusion: the world works as expected, no inherent anomalies

Phase Two: main experiment
First Card: standard frequency, no unlocks
Second Card: standard frequency, two unlocks
Sample: 300 drawings to a 1-card hand (2 permanent blocker cards were used)
Results: 108 first, 192 second (interesting partial results: 33:67 and 66:134 - right on the money!)
Conclusion: the bug is real

Phase Three: testing for scaling
First Card: standard frequency, no unlocks
Second Card: standard frequency, ten unlocks (yes, 10 - you can only see 4 in the world 'cos 6 are of the &quotat most&quot variety)
Sample: 100 drawings to a 1-card hand (2 permanent blocker cards were used)
Results: 11 first, 89 second
Conclusion: the bug is definitely real AND scales with the number of unlocks

(Also, the Remote Lodgings made things worse for those not using them…)
edited by mp on 1/13/2019

2

Ah yes, this explains many things, such as Slowcake’s showing up commonly despite its listed rarity.

3

Discovering this in the weeks leading up to Sunless Skies’s full release means it will likely be a while before this gets fixed, but I hope it does get addressed eventually.
edited by NotaWalrus on 1/14/2019

4

This probably should be fixed, as it isn’t working as intended, but I’m not so sure it must be. The deck works as is, and it would be an adjustment to get used to cards suddenly having cards show up half as often or less.

5

Mp, you should file a support ticket with Failbetter. They don’t always read the forums and with something of this magnitude it would be good to have an explicit tab on their awareness of the issue

6

I did, even before making this thread!

7

Just a quick update, 'cos I got an initial reply from the support:

[quote=Support]Thanks for the report and for the thoroughness of your investigations.

I will look into this when I have time.[/quote]
So they’ll look into it, but no ETA (we probably shouldn’t expect much 'till Skies are out).
edited by mp on 1/23/2019

8

[color=#C2B280]Thanks, this is persuasive. We’ll implement a fix for this, but it will be some weeks from now – we’re going to need to think a bit about whether to alter any card frequencies at the same time or shortly afterwards, and as a couple of you mentioned, we’re a little preoccupied with a certain other game this week…[/color]

9

[quote=Flyte][color=#C2B280]Thanks, this is persuasive. We’ll implement a fix for this, but it will be some weeks from now – we’re going to need to think a bit about whether to alter any card frequencies at the same time or shortly afterwards, and as a couple of you mentioned, we’re a little preoccupied with a certain other game this week…[/color][/quote]It is nice to go shopping with my Noman and have an evening’s zailing, but that singer can sometimes… well, you know!

In these tests, does not having a quality count towards card draw rate? For example, does the Amanuensis count for 2 x 80% because it requires to have no obscurity?

10

I’ve also received a further reply from the Support about the nature of the bug:

Also,

[quote=Skinnyman]
In these tests, does not having a quality count towards card draw rate? For example, does the Amanuensis count for 2 x 80% because it requires to have no obscurity?[/quote]
Yes, based on the test we did on the Storynexus world, unlocks and locks have the same effect.
edited by mp on 1/28/2019

11

So it’s NOT worth to move to a remote location if:

  • you don’t know the Struggling Artist AND
  • you keep your Nightmares to 0 AND
  • you keep your Bohemians favours to 0

I’m ignoring the weasel card as it doesn’t have any other requirements and it’s at 20% frequency. Should be enough to counter the AP costs to move between the lodgings when needed (if needed).

As many had the luck of meeting the artist, I would recommend to go for a remote location! Entering the Nadir before and right after TtH will surely improve the outcome.

EDIT: As much as I look at retired companions, I think someone may deserve something… rubbery!
edited by Skinnyman on 1/28/2019

12

Not to necro a thread too hard, but have there been any recent updates from FailBetter about this issue?

13

Is it just me or this seems to be fixed?
I discarded the Singer a few times to temporary make space in my deck and it took few quite a while to pop back.

14

Maybe it’s partially fixed? I’ve also noticed devils do not take a shine on me as often as they used to. But in this thread, James says they expect it to be fixed soon (at the very end of the post).
edited by rahv7 on 9/3/2019

15

Sigh. The singer and the Amanuensis are both very useful when Making Waves. Not sure I want them gone. If the struggling artist had his own story that gets rid of him, that would be lovely though.

16

Amanuensis, definitely. FB did state repeatedly they’d be open to looking over the entire deck from a modern balace point of view, with this change very much being one of the factors that would be looked at, so it’s not inconceivable they’d simply make it slightly more frequent to compensate.
The singer, though, is the sole reason I’ve spent the last couple of years living in a temple carved into a stalagmite. She gives less MW than even simply side-converting.

17

There are times where I want to be like: &quotHere’s 1000 Echoes. You’re not poor anymore. Stop pestering me.&quot

That plus I forgot how I was able to be rid of the Sardonic Music-Hall Singer previously, and when I returned to the game I accidentally re-acquired her attention. (RIP my deck. :-p )

18

I’m most sad that "“A commission from anarchists” has become scarcer. I now seem to run a slight net positive in Suspicion over time.

19

I’m with rahv7 here with the impression that shine-afflicted devils are only coming around, say, once a day instead of three times.

20

I am also not getting the artist as often as I used to (I had already gotten rid of the Devil using Fate at the Feast of the Rose). However, I miss the singer. The MW and the presbyterate paraphrases were helpful.