Home | About | Tournament Winning Decklists | Forums

Hitting the complete design space limit of the game


Well, the majority of BGG users considers Netrunner (Weight 3.36) to be more complex than Poker (Weight 2.47). And I’m inclined to agree. It’s actually kind of obvious: Netrunner has significantly longer rules, a lot more different cards with keywords, factions, deckbuilding, etc. and it’s also an asymmetric game.

Your anecdote about the Poker AI is quite meaningless since no comparable effort has been made to create an AI for Netrunner.
Personally, I’ve been more impressed that an AI has beaten top Go players. That had been considered impossible for decades. Now, the sky’s the limit and given the progress being made with AIs that are able to learn and play any game, it’s probably only a matter of time until we’re able to create a general purpose AI that will easily beat the Turing test.


The top 16 bracket at least has information about the people, if not the decks.

CTM/Maxx vs. CTM/Val


Can someone explain to me what “complete design space limit” means? I’ve played a lot of games and I hear the term all the time, but I still can’t really get my head around it. What other games have reached this “design space limit” and what did that mean for them?


To me, and others can feel free to correct me, it means “doing everything you can with the game given its core mechanics”.

In Netrunner we have a system where the Runner can run servers to get agendas and the Corp attempts to prevent this from happening and score their own.

We’ve explored the “design space” for the Corp by releasing cards that allow the Corp to score/protect agendas in different ways. There are cards that help the corp lockdown servers, tag/kill runners, spam assets, fast advance agendas, and more. To me further “exploring the design space” of the Corp is releasing cards that open up new strategies for the Corp or cards that allow the Corp to go about previously established strategies in different ways.

Reaching the “design space limit” for the Corp is when, given Netrunner’s core mechanics, we can’t really release cards that do things we haven’t seen done by cards in previous packs (same general argument goes for the Runner I just found it easier to focus the explanation on the Corp).

Hope this was clear/helps!


I’m not sure a recreational message board like bgg is exactly an authoritative source of information as poker isn’t really even a board game per se.

A longer ruler set etc only seems more complex because a lot of what makes poker difficult is non-intuitive. I think the intricacies of just reading basic body language alone have filled more books than I know of.

Poker being too complex for an AI to be able to easily master is extremely relevant. It shows that even one of the worlds best super computers found the game too complex to be able to break down. And that when it did it showed that GTO plays were so difficult to appraise that basically no human had ever come that close to executing them in terms of exact bet sizing.

If no human has ever played a perfect session of poker and humans routinely play perfect games of Netrunner, that is my case.

I assume we are talking about NL Hold’em here but if we are including variations like PLO or hi/lo versions, the game theory gets even more complex.

Also, and I concede this is more anecdotal, but if Netrunner were harder than poker, I think we would see at least some players transition, considering poker can make you wildly wealthy. In 2003 during the poker boom, the highest level backgammon players in the world transitioned to poker and had a huge edge because they were accustomed to thinking at a higher level. Many had huge success.

I don’t see what Go even has to do with that argument.


Umm. I guess you don’t know about the site because it’s about all kinds of ‘tabletop’ games. The weight scores are determined by consensus. You can see how many votes have been given for either game. I’ve found the weight scores to be extremely accurate as soon as about a hundred votes have been made.

You may not have heard, but AIs are already better at interpreting body language than humans. They’re experts in detecting lies (good luck trying to bluff them!) and are already used in crime investigation and job interviews.

It doesn’t matter because you cannot compare it to an AI playing Netrunner, because there is none.

How would you know? How do you even define ‘perfect’ in the context of Poker or Netrunner. I could claim I’ve seen dozens of perfect Poker games and you’d have no way to prove me wrong.

You seem to assume Netrunner is played because the players hope to get wealthy by doing it?! That’s crazy talk.

Apparently you don’t know anything about Go either. I’ve given it as an example of a game with a very simple ruleset but incredibly deep gameplay.

What was interesting about the AI playing against the Go Masters is that the Go Masters were often unable to determine the AI’s strategy. It played in a way that made no sense to them, yet it managed to win almost every game.
This clearly shows that despite of decades of research and practice by millions of people we still haven’t discovered much about the game.


Not to really get into it or anything, but it’s not surprising to see players go from Magic to Poker professionally but not from Netrunner to Poker, given that Magic tournaments have cash payouts and Netrunner tournaments don’t. People who want to play Magic for money will eventually switch to playing something else for more money. People who want to play for money probably wouldn’t be spending their time on competitive Netrunner in the first place.


what is the connection of the Poker discussion to the design space limit? I can’t see it!


Pizzatime discussed it during their free stream at Worlds.


My hot-take (that’s actually really chill)…

Netrunner players are the surfers of Card Game World. Sure, there are a group that compete at the very high levels (same as competitive surfers), but a majority are happy to hone their craft and do what they want to do with the game. Maybe go to a meet up, play some hacks, perhaps someone will bring their acoustic guitar and play it by the digital bonfire.

I dunno…I think this analogy is slipping away from me.

What were we talking about again?


Yeah I have trouble believing that perfect netrunner is being played on a regular basis if at all, that’s a pretty bold claim.

I actually have a passing interest in a lot of this machine learning stuff and I’ve been looking into ways to apply current research to a Netrunner AI. I’m really curious to see how AI can handle probablistic games that have a deckbuilding component, let alone ones as complex as netrunner.

If in poker you have to figure out a distribution on betting for a certain number of states of cards, individual/pot pools, and aggressiveness, Netrunner’s analogue is figuring out a distribution on drawing, running, and any paid action over orders of magnitude more game states and possibilites of deck matchups. Not to mention the whole history of the game being more relevant.


@PizzatimePlayer an AI can’t build a deck and play perfectly a game of Netrunner, I don’t see your point.

You also can play “perfectly” games of poker, like ANR, if you remove your opponent bluff and both player luck from the equation.

The amount of hidden elements is way different between the two games, when in ANR you don’t even know your opponent’s alphabet to begin with. That is a reason why games like Starcraft aren’t yet solvable by AI too.

And even there : Poker cards, or Starcraft game element doesn’t change each month like when ANR was a LCG.

Also, “solvable by AI” doesn’t tell a game is easier than another : I guess Go is harder than Poker. Poker doesn’t take a life to master : remove time limits in Go and youngster get kicks from old dudes.

People saying that is a guy skilled enough to earn money with Poker and places in finale in most card games he participates, including Magic, ANR, and NR. I think he knows his subject better than me.


The connection is though to be “the game is deeper than poker, so there is design space left”, to summarize bluntly, then “if a computer AI solve poker, ANR is not deep because I claim I solved it”, or something.

I don’t know. People saying ANR is solved at some point yet have to demonstrate how an evolving game can be solved.


I’m not sure if there is design space left in poker. They seemed to have stopped releasing new cards a good while ago.


Maybe we should start a fan organisation dedicated to continuing it! :smiley:


But then, nobody could say design space is closed ? :stuck_out_tongue: