Card Game DB
Sign In
Create Account
Welcome to Card Game DB
Register now to gain access to all of our features. Once registered and logged in, you will be able to create topics, post replies to existing threads, give reputation to your fellow members, get your own private messenger, post status updates, manage your profile and so much more. If you already have an account, login here - otherwise create an account for free today!
Register now to gain access to all of our features. Once registered and logged in, you will be able to create topics, post replies to existing threads, give reputation to your fellow members, get your own private messenger, post status updates, manage your profile and so much more. If you already have an account, login here - otherwise create an account for free today!
1
Turing Test
Nov 25 2013 06:05 AM |
Eldil
in Android: Netrunner
Android: Netrunner Turing Test Eldil
A week ago, on a whim, I designed a deck with a single purpose: to put a Monolith and a full suite of icebreakers into play as quickly as possible. After I’d built the deck, I started to test and tune it by running it solo, just to see how many turns it would take me to install Monolith. I’d do that a couple times, record how many turns it took, and then tinker with the deck.The deck I wound up with wasn’t what I would call competitive, but the process was illuminating. I saw, first hand, what cards were working, what cards were clogging my hand, what got me closer to my goal, and what just got in the way.
There have been several attempts at “Chess Problems,†for Netrunner, and although some of them have been quite clever, I’ve never found them very satisfying. This is because chess problems present a predetermined board, hand, and deck, which eliminates the most vital and interesting strategy element of Netrunner: deck construction.
After several attempts at a “Monolith speed drill,†it occurred to me that this might be a fun challenge for others, and might lead to helpful analysis and discussion. So I’m presenting it to you as the first “Turing Test†(since every deck is like a program you run, and the Turing Test is designed to evaluate the intelligence of a program, and the phrase “Turing Test†sounds cool and cyberpunkesque).
If this is something that people find fun and enjoyable, and scratches that “Chess Problem†itch, then I’d like to do a series of them. Each Turing Test will present a simple goal in static, predetermined conditions, and the challenge will be to build, tune, and test a deck that accomplishes that goal in a minimum amount of turns.
Turing Test 1: MAD (Monolith Assembly Drill)
“Drill Sergeant: Why did you put that weapon together so quickly, Gump?
Gump: You told me to, Drill Sergeant?â€
Goal: Install a Monolith and three icebreakers, one of each type (fracter, killer, decoder) in as few turns as possible.
Setup: The three central servers are each protected by a rezzed Ice Wall. There is one empty remote server protected by a rezzed Enigma. The corp has five credits. Successful runs on HQ, R&D, or Archives do not result in access, though runners can still use Account Siphon and any other cards that trigger off a successful run. At the end of a runner’s turn, the following occurs.
* Any unrezzed ICE is automatically rezzed again at no cost to the corp (sorry, Xanadu, Cortez Chip and Rook)
* Any trashed ICE remains trashed, and no new ICE replaces it.
* The corp gains credits until it has five credits.
* If the runner has a tag, he or she loses all credits (as if a Closed Accounts had been played)
Then the runner’s next turn immediately begins.
Deck Construction Limitations: None
How to participate: Build a deck, run it three times, and each time record how many turns and clicks it took to accomplish the goal. Then post your results as a comment, with your Deck List, Average Performance Rating, Peak Performance Rating, Rundown, and Extra Credits:
* Average Performance Rating (APR) is the average number of turns across your three runs (clicks are ignored in APR, only turns are considered)
* Peak Performance Rating (PPR) is the number of turns it took on your best run, followed by the number of clicks spent on your final turn.
* Rundown is where you present your analysis of the deck: how it worked, how it didn’t work, what you learned.
* Extra Credits is a place to list any special achievements or conditions you met: for the MAD challenge, building your deck with either an Anarch or Criminal ID should be noted in the Extra Credits. Or, if you want added difficulty within Shaper, try a MAD Opus drill, where the goal is to have Monolith, a full icebreaker suite, AND a Magnum Opus installed.
Below is my submission for this Turing Test. I went in a different direction than my first attempt; I don’t think my approach was optimal, and I’m hoping that some of you deck Whizzards can show me some much better builds.
Identity:
Kate “Mac†McCaffrey: Digital Tinker (Core)
Total Cards: (45)
Event (18)
Easy Mark (Core #19) x3 *
Sure Gamble (Core #50) x3
Stimhack (Core #4) x3 *
Diesel (Core #34) x3
Quality Time (Humanity's Shadow #87) x3
Infiltration (Core #49) x3
Hardware (3)
Monolith (Creation and Control #36) x3
Program (9)
Snowball (Trace Amount #27) x3
Gordian Blade (Core #43) x3
Creeper (Humanity's Shadow #89) x3
Resource (15)
Personal Workshop (Cyber Exodus #49) x3
Aesop's Pawnshop (Core #47) x3
Wyldside (Core #16) x3 ***
Same Old Thing (Creation and Control #54) x3
All-nighter (A Study in Static #67) x3
Influence Values Totals:
* Anarch: 12
* Criminal: 3
* Shaper: 60
APR (Average Performance Rating): 6.3
PPR (Peak Performance Rating): 6 turns, 2 clicks
Rundown: My goal was to use Stimhack + Personal Workshop to effectively discount Monolith by 9+ credits, but I was honestly a little disappointed in this deck’s performance profile. In two of the tests, I was hoping to break five or even four turns, but there were too many pieces to pull together, and one of the pieces would stubbornly refuse to come into my grip. Economy was never a problem—Sure Gambles, Easy Marks, Infiltration (played for 2 credits), and Aesop’s usually had me at 10+ credits within three turns, which is all I need for the Stimhack combo. But either the Stimhack or the Personal Workshop would be elusive, and I would be stuck with a hand of 4 cards I wasn’t willing to discard (Gordian Blade, Creeper, Snowball, Monolith). Additionally, Kate’s ID ability never came into play. I’m hoping that one of you—or several of you—can build a deck that will run laps around mine.
Extra Credits: None
- Midian and CommissarFeesh like this
17 Comments
Chaos Theory: Wunderkind (Cyber Exodus)
Total Cards: (40)
Event (20)
Diesel (Core #34) x3
Freelance Coding Contract (Creation and Control #33) x3
Modded (Core #35) x3
Sure Gamble (Core #50) x3
Quality Time (Humanity's Shadow #87) x3
Easy Mark (Core #19) x3 â–
Account Siphon (Core #18) x2 â– â– â– â–
Hardware (3)
Monolith (Creation and Control #36) x3
Program (14)
Gordian Blade (Core #43) x3
Snowball (Trace Amount #27) x3
Creeper (Humanity's Shadow #89) x3
Magnum Opus (Core #44) x2
Battering Ram (Core #42) x1
Peacock (What Lies Ahead #6) x1 â– â–
Ninja (Core #27) x1 â– â–
Resource (3)
Armitage Codebusting (Core #53) x3
Influence Values Totals -
Anarch: 0
Criminal: 15
Shaper: 51
APR (Average Performance Rating): 3.3
PPR (Peak Performance Rating): 3 turns, 2 clicks
Rundown: I believe Chaos Theory was the key here, only having 40 cards in the deck really seemed to allow the right cards to come up. If I was able to get at least one Sure Gamble right away I was very fortunate, however Armitage Codebusting seemed to be the most reliable way to get a good burst of cash. Freelance Coding Contract also accelerated the deck quite a bit as it was able to clear all the extra programs for money I grabbed using Diesel and Quality Time. Drawing Account Siphon (while it would be great vs an actual player) was a dead draw whenever I grabbed it. Lastly because I had Creeper/Battering Ram in the deck it would cause me to have to acquire an extra credit before being able to lay down the whole combo which became irritating. Overall it was a blast trying to figure out this Turing Test and I had a lot of fun doing it! Nice job creating this!
Extra Credits: The one Game it took the full 4 turns I ended up in great shape afterwards with 5 cards in hand (Magnum Opus, 2x Armitage Codebusting, Easy Mark, Sure Gamble) Plus 2 credits left over!
So, should I work on a different enough solution, or post this one?
Total Cards: 45
- Event (26)
- Account Siphon (Core) x3
- Easy Mark (Core) x3
- Emergency Shutdown (Cyber Exodus) x1
- Hostage (Opening Moves) x1
- Inside Job (Core) x3
- Special Order (Core) x3
- Diesel (Core) x3
- Quality Time (Humanity's Shadow) x3
- Infiltration (Core) x3
- Sure Gamble (Core) x3
- Hardware (2)
- Monolith (Creation and Control) x2
- Program (10)
- Aurora (Core) x3
- Faerie (Future Proof) x1
- Ninja (Core) x3
- Peacock (What Lies Ahead) x3
- Resource (7)
- Bank Job (Core) x3
- Mr. Li (Future Proof) x3
- Same Old Thing (Creation and Control) x1
APR (Average Performance Rating): 3,0PPR (Peak Performance Rating): 3 turns, 2 cliks...one click short of 2 turns but than the early tags would have deleted my $
Rundown: Andromeda is awesome to manage the challenge with just 2 Monolith. There are a lot of cards that are "bad" for the deck like hostage ore Same Old Thing but without Influence left all alternatives seem to do worse. I usualy played Aurora to break the walls for account siphon but in the last game inside job and bank job did the trick. Mister Li is very good for searching if burning through quality time dosn't help. I was a little lucky with those 3 games but the cash and the cards flow in like crazy and there is even time to special order the last pice of the puzzle.
Extra Credits: Criminal. Did not use Account Siphon in game 3
Awesome challenge, though I would have loved to use data dealer xD
Identity:
Chaos Theory: Wunderkind (Cyber Exodus)
Total Cards: (40)
Event: (9)
Diesel (Core) x3
Quality Time (Humanity's Shadow) x3
Freelance Coding Contract (Creation and Control) x3
Hardware: (2)
Monolith (Creation and Control) x2
Program: (29)
Battering Ram (Core) x3
Creeper (Humanity's Shadow) x3
Cyber-Cypher (Creation and Control) x3
Dagger (Creation and Control) x3
Gordian Blade (Core) x3
Inti (Creation and Control) x3
Pipeline (Core) x3
ZU.13 Key Master (What Lies Ahead) x3
Snowball (Trace Amount) x3
Corroder (Core) x1 â– â–
Mimic (Core) x1 â–
Resource: (0)
Influence Values Totals -
Anarch: 3
Criminal: 0
Shaper: 69
APR 3,,1
PPR 2 turns, 4th click
Quality Time, diesel and freelance coding contract can generate a ridiculous amount of money amazingly fast if your deck is pretty much all program. As a bonus, all that draw makes it more likely you'll hit a monolith.
Nice idea with the massive program spam!
@CommissarFeesh: I've already designed the next Turing Test, and it will not require C&C (in fact, my example deck for that Test is built just from just the Core Set).
@FreezerBurner: My initial Monolith deck was a Chaos Theory/Coding Contract deck, but it never performed anywhere near as well as yours. Nice work and thanks for sharing!
@PeekaySK: Your decklist and solutions can be as different as you want them to be. Sometimes it can be really helpful to see the performance differences between similar decklists, and sometimes it can be fun to see how well you can do with something radically different. But, if you've already got a solution tested, definitely post it.
@Christars: I appreciate your thoughtfulness, but you should never worry about "going against the spirit" of a Turing Test. I'll post the goals and the restrictions, and anything else is fair game. If you can install a Inti/Zu/Faerie/Monolith with a significantly better performance than other breakers, do it. The spirit of Turing Test is the spirit of lawless hackers and ruthless corporations; figure out a way to do it better and faster than anyone else, and let the ones on the sidelines argue about whether it was good form.
@Moloch: Daaaaaamn. I wouldn't have expected a Criminal deck to do so well. I'm glad someone decided to exploit the weakly defended remote with Bank Job. I agree Data Dealer could even accelerate it even more. And imagine if your opponent's surprise if you slapped down a deck that runs both Monolith AND Data Dealer.
@sechen: Wow; you got monolith installed on the second click of your fourth turn? That's... terrifying.
Sure Gamble x 3, Modded, Monolith, Corroder, Yog, Ninja, (ninth card)
Odds against it are stupidly high.
OK. First of all, a disclaimer: I'm thinking about this more as a "build a viable Monolith deck" challenge more than a controlled test. Hence, my goal wasn't to build something that's super-tuned to perform exceptionally well in the specifically prescribed situation (and makes generally suboptimal choices for actual play in order to boost its performance in the stated test case), but rather I decided to build a deck I wouldn't mind playing on my league night, and then run it through the test.
I've played some OCTGN games with it, haven't put it together in flesh yet though, so I don't have the APR and PPR for it yet. Will add those later (how do I put it... my job is pretty cool, but I still can't play around with actual proper cards on my desk
Now, the list:
"Big Girls Don't Play With Dolls"
Identity:
Chaos Theory: Wunderkind (Cyber Exodus)
Total Cards: (40)
Event (17)
3x Diesel (Core #34)
3x Quality Time (Humanity's Shadow #87)
3x Sure Gamble (Core #50)
3x Freelance Coding Contract (Creation and Control #33)
2x Account Siphon (Core #18) â– â– â– â–
3x Levy AR Lab Access (Creation and Control #35)
Hardware (6)
2x Monolith (Creation and Control #36)
2x Plascrete Carapace (What Lies Ahead #9)
2x R&D Interface (Future Proof #107)
Program (17)
3x Cloak (Creation and Control #41)
3x Dagger (Creation and Control #42)
3x Corroder (Core #7) â– â–
3x Gordian Blade (Core #43)
3x Magnum Opus (Core #44)
1x Net Shield (Core #45)
1x Mimic (Core #11) â–
Resource (0)
Influence Values Totals -
Anarch: 7
Criminal: 8
Shaper: 59
APR: <coming soon>
PPR: <coming soon>
Rundown:
What's Monolith's biggest benefit? The 3 extra actions it gives you for installing programs, and the 3 extra memory to use for it. I'm not fixating on using programs that actually make use of the discount - the logic here being that if the discount makes me use programs that are suboptimal in the long run (see: Battering Ram), it actually costs me more money than it saves. It's like Scud said a while ago - the real question with Monolith isn't how to get it into play (runner economies can do some crazy ****), but how you use it once it is in play.
Onto the breaker selection: Cloak & Dagger is awesome and very efficient, its only two problems are "actions-intensive" and "memory-intensive". Good thing we're building a monolith deck
Core engine: Massive card draw is the way to go if I want to actually get my pieces together (Cloak & Dagger sorta requires it, really). That, combined with Monolith's prevention ability means Freelance Coding Contract. THAT in turn means Levy AR Lab Access - initial drafts of the deck only had 2 copies, but I decided to drop a third Monolith (was stuck with 2 earlyMonoliths too often) to add a third Levy, to keep going longer with FCC - FCC becomes incredibly awesome on the second run through the deck, when all you have in are programs, one spare Monolith and card draw events
Finishing touches: Net Shield is a must vs. all those Jinteki decks that started popping up lately (Monolith doesn't change the click economy of net damage, just reduces the impact if you have a card in hand you need to protect), so is 2x Carapace. 2x RnD interface is needed since I might not be able to get into RnD more than once per turn, no matter my credits ('cause of Cloak & Dagger). At this point, I noticed I have 6 influence left, so I dropped two Mimics and added two Siphons - because hey, why not. It might make me more volatile (if I really need that Mimic and can't get it), but it'll also speed me up considerably. I normally wouldn't play a Runner deck without Infiltration, but Monolith's brain-damage prevention makes it possible (tags I don't care about, so Ghost Branch is out too).
Extra Credits (a.k.a. "Why my deck is different from FreezerBurner's"):
For starters, I feel my breaker suite is much better suited to actual play overall(his breakers are better tuned to this specific Turing Test). Also, I'm including things that you *just can't live without*, like Carapace (pun intended).
Levy AR Lab Access is something that I feel is essential here - if you don't play it, you'll deck yourself, guaranteed (and pretty fast, too). Second and third passes through a deck like this will make Freelance Coding Contract just godly. As an added benefit, it also makes you super-resilient to program trashing.
Initially, I also considered going the Armitage route (to get money up faster), but then decided to stuff in more programs instead - Armitage is a dead card once Magnum Opus hits the table, and is pretty anti-synergistic with Levy here. I'll probably take a huge speed hit in the test for it, but so be it - like I said, I want to create an actually competitive deck her.
----------------------------------------------------------------
There, that's about it for me. Hope I didn't blabber on about things that are too obvious, and I'll add the test numbers when I get around to throwing the deck together
----------------------------------------------------------------
edit:
Update: So I've run my CT deck through the test. The results were... better than what I expected.
APR: 3 turns, 3 clicks
PPR: 2nd turn, 2nd click
Notes: In two of those three games I did the Siphon play to speed myself up through the finishing line. In one that wasn't needed... and surprisingly enough, it was the one with PPR. That game, I just got a God Hand - Freelance Coding Contract, Quality Time, Sure Gamble and 2 programs. After I gambled and QTed, I got a Diesel and some more programs. Playing the Diesel netted me a Monolith and enough programs to be able to dump the full Contract and keep the trio I'd need for the Monolith play. Turn 2 was just the second Sure Gamble (that I drew in that card-drawing frenzy on turn 1) and slap down the Monolith. DONE.
I've since run the test about 10 more times, to make sure I wasn't getting a statistic anomaly, and my PPR is still somewhere around 3.75, so it seems to do reasonably well. Granted, a couple of those times I sorta cheated by only having Dagger and no Cloaks when I reached my goal, but I didn't find it making such a huge difference (not like the test is terribly realistic anyway
Extra Credits: I've played quite a few regular/serious games with this at this point, and I have to say I like how it holds up. The whole "be smart, recycle!" aspect makes me super-resilient to program trashing (played a guy who trashed a 2-pointer in order to nuke my Corroder with Archer... I replaced it next turn
Overall though, I'm very happy both with the test results and the deck's performance in actual play.
Identity:
Andromeda: Dispossessed Ristie (Humanity's Shadow)
Total Cards: (45)
Event: (18)
Sure Gamble (Core) x3
Easy Mark (Core) x3
Levy AR Lab Access (Creation and Control) x1 â– â– â–
Freelance Coding Contract (Creation and Control) x3 â–
Forged Activation Orders (Core) x2
Quality Time (Humanity's Shadow) x3 â–
Account Siphon (Core) x3
Hardware: (2)
Monolith (Creation and Control) x2 â– â– â–
Program: (25)
Aurora (Core) x3
Crescentus (A Study in Static) x3
Femme Fatale (Core) x3
Ninja (Core) x3
Sneakdoor Beta (Core) x2
Gorman Drip v1 (Opening Moves) x3
Snitch (Cyber Exodus) x3
Pheromones (Humanity's Shadow) x2
Peacock (What Lies Ahead) x3
Resource: (0)
Influence Values Totals -
Anarch: 0
Criminal: 59
Shaper: 15
I thought of this deck but haven't tried it out yet. Will update with progress in few days(hopefully)
APR: 3 turns, 3 clicks
PPR: 2nd turn, 2nd click
Notes: In two of those three games I did the Siphon play to speed myself up through the finishing line. In one that wasn't needed... and surprisingly enough, it was the one with PPR. That game, I just got a God Hand - Freelance Coding Contract, Quality Time, Sure Gamble and 2 programs. After I gambled and QTed, I got a Diesel and some more programs. Playing the Diesel netted me a Monolith and enough programs to be able to dump the full Contract and keep the trio I'd need for the Monolith play. Turn 2 was just the second Sure Gamble (that I drew in that card-drawing frenzy on turn 1) and slap down the Monolith. DONE.
I've since run the test about 10 more times, to make sure I wasn't getting a statistic anomaly, and my PPR is still somewhere around 3.75, so it seems to do reasonably well. Granted, a couple of those times I sorta cheated by only having Dagger and no Cloaks when I reached my goal, but I didn't find it making such a huge difference (not like the test is terribly realistic anyway
Extra Credits: I've played quite a few regular/serious games with this at this point, and I have to say I like how it holds up. The whole "be smart, recycle!" aspect makes me super-resilient to program trashing (played a guy who trashed a 2-pointer in order to nuke my Corroder with Archer... I replaced it next turn
Overall though, I'm very happy both with the test results and the deck's performance in actual play.
(putting this edit into the original post as well, for completeness' sake)
Thanks for this challenge. I don't think my Chaos Theory deck is significantly different from the one above but I will share it anyway (http://netrunnerdb.c...-monolith-t1000)