Home | About | Tournament Winning Decklists | Forums

Fallacies In Netrunner And How To Avoid Them by Xenasis


#1

Originally published at: http://stimhack.com/fallacies-in-netrunner-and-how-to-avoid-them-by-xenasis/

Andrew “Xenasis” Hynes returns for his second StimHack article, outlining some common Netrunner fallacies and why they are misleading.


#2

As always, it’s greatly appreciated if you could go to reddit and upvote stimhack for visibility!


#3

Great article. One minor nitpick is the following isn’t necessarily true:

Saying that 7 shuffles is adequate to randomize a [52 card] deck is based on a pretty loose definition of “random,” where you’re willing to allow for a given event (ie. a specific sequence of cards) to be up to 33% more likely than under a uniform (ie. `truly’ random) distribution, which IMO is a bit much. Now, to be fair, the fact that netrunner decks generally run duplicates in theory cuts down the necessary number of shuffles, as does the smaller deck size, but if every card in our deck is unique, 10-12 shuffles gets us to within about 1% of a uniform distribution.

So basically, the recommendation of at least 10 riffle/mash shuffles is a good one, but, like most things to do with statistics and probability, the math is actually not quite as cut and dry as most people realize :wink:


#4

Yeah, (as far as I understand it) I’m pretty sure best amount you can get without wasting your time lies at about 13 riffles for a deck of cards, but 7 is about the amount where it makes a minimal amount of difference compared to the difference the first 7 made.

It’d be unreasonable to expect people to shuffle 13 times each Keyhole run, but yeah, I should have been more clear. ‘perfect’ is wrong, “near-perfect” is more accurate.


#5

Nice article about one of my favorite topics.

The “don’t trash assets because i have let them stay for the last few turns”-thing happens to me all the time. I have played so many games where at one point, I look at the board, think “It would have been correct to trash that asset a while back. Well, now it is too late anyways” and don’t touch it for the rest of the game knowing that it would be correct to go trash it even now.


#6

Great article!

I’d always heard the opposite side of the Gambler’s Fallacy, the whole thing with “streaks”. If the runner has a 40% chance of hitting an agenda on HQ, while it’s true that the chance each access stays the same, relying on them continuing to not hit one is ill-advised as well.

But yeah, really great, and the kind of thing that’s not intuitive to most folks (and even to those who grasp the ideas, is good to have the reminder).


#7

I did the human randomization thing and after 100 tries was 51-49. Although humans are better at mind games than machines, that definitely boosted my confidence.


#8

0/10 would not read again


#9

Yup I agree. Honestly, riffling your deck a couple times then cutting after each Keyhole is probably fine, considering all but 2 of your cards should be properly randomized to start with.

Even harder than trashing partially-emptied assets for me is walking away from an NAPD. Figuring out when you’re actually able (in a “this is a +EV play” sort of sense) to steal one of those is one of the hardest decisions in any game I play, just because the irrational sunk-cost-calculator part of my brain goes into overdrive.


#10

I usually riffle shuffle 5-6 pile shuffle once then riffle shuffle 5-6 times at the beginning of the game. The pile shuffle in the middle is to guarantee I haven’t had any cards sticking together which would make my riffle shuffle not random (in a deterministic way).


#11

Small thing, hovering over the card names shows the full card art. Did FFG change their policy or something?


#12

No, @SneakySly just made the website better.


#13

Small thing on your shuffling part of your great article

Perfectly “Mash” shuffling 8 times will results in the cards being in the same order. Meaning shuffling 10 times isn’t that great either. Although this is near impossible with sleeved cards(I tried, can get it maybe 4/10 times).


#14

Not that it ever matters in practice, but as some trivia, this number changes depending on your deck size. It’s 8 for 52 cards, 58 for 60 cards, 12 for 45 cards, and 21 for 49 cards. (And 30 for 77 cards, hello syntax)

It follows this sequence: http://oeis.org/A002326

[quote=“OEIS”]In other words, least m>0 such that 2n+1 divides 2^m-1.
Number of riffle shuffles of 2n+2 cards required to return a deck to initial state. [/quote]

(Actually I didn’t know that in advance, I just made a quick program to generate the sequence and then asked OEIS if it recognized it. Pretty cool huh?)

In practice when I shuffle I intersperse two or three overhand shuffles with a lot of smash shuffles, because otherwise a card that starts at the very top or bottom could stay there or move only a small distance away.


#15

tl;dr

Always consider all scenarios when deckbuilding
Every deck can be better
Mill is good
Riffle shuffle at least 7 times, octgn shuffler is ok
There is never a case that you “must” do something
It’s never too late to trash something

Good article


#16

Great article.

Now I understand how to build better deck.

And why Memstrip is not a good card @mediohxcore. :wink:


#17

Solid article. My only nitpick is with sunk costs – something like the toll on an unexpected booth becomes very relevant if you anticipate that you’ll need to complete this run at some point in the future anyway. E.g. if there’s a Sundew back there, and you can still afford to trash it after breaking the Tollbooth, you should probably continue even if it means some tempo loss (clicking for credits) because a) you don’t want to let it sit on the table for several turns, and b) coming back later will cost an additional 3 credits (minimum).


#18

That was a good read indeed. Sure wished it was written last year’s January when I first started. Would’ve saved a TON of the constant failing of deck building, and losts due to player error. I guess that’s why netrunner is great, skill developing is the true reward of trial and error :smiley:


#19

But even then, you’re considering the future cost, not the sunk cost. The first 3c is gone forever and shouldn’t matter when you decide.


#20

@mediohxcore this is off topic, but since we don’t have a proper Memstrips thread…