Dark Forest is a massively multiplayer (MMO) real-time strategy (RTS) game in closed beta. If you like strategy games, you should play it. Anyone who is a fan of games like Sins of a Solar Empire, Factorio, or Civilization should definitely play it. It’s not a game with global appeal like Super Mario Odyssey, but it doesn’t need to be nor try to be.

Personally, it's the best new game I’ve played in 2022. When I got off the waitlist I played it for 10 hours straight -- that’s why I’m writing this post. It’s the most exciting strategy game I’ve played since StarCraft 2. It’s the most exciting competitive multiplayer game I’ve played since Valorant. And it’s renewed my optimism for the potential of web3 games. Why? It gets many things right, but it gets a few things very, very right:

• It is a refreshing take on a stale genre

• It has a competitive ladder with a fresh take on ranking

• It’s plugin system & open source codebase brings modding back to games

• It shakes up assumptions around the infrastructure costs of online multiplayer games

• It has a thoughtful team that closely engages with the community

Intrigued? Great, read on! But there is one minor detail you might care to know… Dark Forest runs on “The Blockchain”.

Blockchains and cryptocurrencies are a divisive topic in recent times, but especially so in the gaming community. Initial promises of a cross-platform utopia with transferable skins and items quickly gave way to a wave of “crypto games” with terrible monetization schemes and high profile hacks. These, along with other narratives around the most popular blockchains using energy-guzzling Proof of Work algorithms and NFTs destroying the environment, have poisoned the well for any game touting blockchain-adjacent features.

The negative reaction from the gaming community to companies adding blockchain features, like Discord teasing an Ethereum integration, makes sense! The cut runs deeper and older than crypto gaming’s recent spotlight – it triggers all-too-fresh memories of the gaming industry's parlous relationship with new types of monetization. The sting of in-game ads, base game content behind paid DLC, and hellish microtransaction models is still felt in the community. And while some developers, like the SW: Battlefront 2 devs, had a change of heart, it’s been a rocky journey. Misguided attempts to win over players notwithstanding, the community still finds itself with games like The Sims 4 offering $800 of paid DLC.

The good news is that Dark Forest is not trying to sell you anything, nor is it an abusively monetized game. It is a great game that just happens to use new technology in cryptography and blockchains to their advantage, rather than shoehorn them in.

It also isn’t funded through venture capital, instead relying on community and non-profit funding sources. So don’t worry Jack, no a16z or VCs on their GitCoin page or funding page! Raising venture capital is fine (I’ve done it!), but in this case the developers truly have put the game and the community first.

You can skip this part if you just want to read about Dark Forest; it just contextualizes my motivation.

I’m a huge gamer (hello, fellow kids!). I’ve religiously grinded the online competitive ladders of StarCraft 2, League of Legends, and Valorant. You can stat check my Steam profile if you want to know more about my tastes.

I’m also a developer and a founder in the web3 space. I think I’m in a unique position to speak about the web3 gaming experience. I know a lot more needs to be delivered before reasonable skeptics can see past crypto’s current failings. I’m an optimist that believes eventually cryptography and web3 tech can re-empower users on web platforms, and I’m trying to build that reality.

And if you don’t believe me, again, check my gaming habits! I promise Axie Infinity is nowhere to be found in my wallet. I did spend a month as a child catching all 150 Pokemon in the original games with a link cable, 2 Gameboys, and the holy trinity of Red / Blue / Yellow versions. I’ve got no on-chain proof for that, though… only PTSD of the Safari Zone….

Dark Forest is a browser-based 2.5D (2D plane, 3D assets) game that takes place in “rounds”, with each round lasting a few days (the current round is underway and will last 10 days). If the following in-game image makes your eyes bleed, the gameplay might still make you stay!

Generally, a few months pass between rounds, which players use to plan for the next round. The current round in February preceded by one last year in October, and game mechanics are added between rounds. While a round is underway, players try to earn the highest score on the leaderboard, viewable in real-time on the homepage. (Yeah, I’m not doing too hot. And yeah, the CEO of Figma won the first round). The top ranked 63 players at the end of a round get a reward from the devs, but otherwise you just play for fun.

The core game loop is:

• All players spawn with a home planet and vision over a small area in their corner of the universe, with fog of war (i.e. no vision) covering the map

• Planets have energy that regenerates over time

• You send energy from your planets to other planets to conquer them

• Some planets have “silver” and “artifacts” which give you points that determine your rank

• You break the fog of war with an “explorer” whose speed is dictated by the compute power of the real life, physical computer you’re playing on

If it all sounded great until that last point, hear me out. No, this isn’t some shadow-realm version of Boktai: The Sun is in Your Hand where you destroy the environment to play. Nor is the game pay-to-win. Nor is it using your browser to mine Monero.

First, AFAICT the game’s miner seems to have fairly limited scaling capability, with a Raspberry Pi giving reasonable performance (1-2k hashes/sec is roughly equivalent to my browser using a Ryzen 3600 XT & 2 cores). On the backend the game runs on the Gnosis Chain (an Ethereum sidechain — don’t worry if this means nothing to you for now), which is Proof of Stake, not Proof of Work. Shadow-realm Hideo Kojima still has a game to make.

Second, yes – faster computers have an advantage. It’s not as big as you’d think (running on 16 cores vs 2 cores is ~2.5x better, and 16 vs 4 is ~1.7x, so diminishing returns kick in). The bigger advantage comes from using a “remote hasher”, which is a separate server that does hashing. But in the early game, you’re mostly limited by your strategy, not the fog of war. In the mid-game, players need to direct their explorer’s search regardless of how much compute they have. If I were to make an analogy to other competitive online games, having a large amount of compute would be like playing a First-Person Shooter a la CS:GO or Valorant with a 240 Hz monitor and a low-latency connection. It makes a difference, but anyone can upgrade their gear to be at their competitive best. You can still have fun and be competitive without top gear.

Third, if you’re afraid the game might be using your computer for nefarious purposes, it’s all open source (both the code and the live smart contracts and frontend). If you don’t have the skill to audit it, then yes, you’ll have to trust a friend or someone else to do it for you. When I quickly read through it, it looked good, promise. You won’t be sending any excellent privacy coins to anyone.

(That last part was a joke, please don’t buy or mine Monero if you’re trying to make a quick buck.)

There are other mechanics like upgrading planets, a planet leveling system (you progressively conquer larger planets as you expand), and “artifacts” that boost your planets’ creation of energy and other resources. These feel natural and are staples of any good strategy game, but the tech tree is less developed than the richness of a Civilization entry, and there isn’t customization on the level of Stellaris.

What Dark Forest lacks in more traditional strategy game features it makes up for in competitive Player vs Player (PvP) design, though.

Dark Forest’s ladder and competitive rewards are sort of similar to other online games like League of Legends. There’s a publicly viewable leaderboard and end-of-season ranked rewards. What’s unique about Dark Forest’s leaderboard is that the game doesn’t use an “Elo system” or feature any skill-based matchmaking at all, which is a departure from the most popular competitive games today. Instead, all players compete with all other players within a round. This may not seem like it sets up the average player for a good time, but we’ll touch on how this doesn’t lead to new players getting Goomba stomped by more skilled ones.

Players earn points by interacting with the game universe – either discovering “artifacts” on conquered planets, or exchanging “silver” for points. Players interact with other players’ planets using the same moves as they would an unconquered planet. The only difference is that players can beef up their planets’ defense so it requires more energy to capture. So rather than a “Player vs. Environment” (PvE) MMO experience, the resources themselves are a battleground for “Player vs. Player” (PvP) gameplay.

Note that the exact competitive mechanics change between rounds, for e.g. there has also been a “race to the center” round, which keeps the game fresh.

This is where the titular mechanic of “a dark forest” comes in. Since there are finite resources and you expand to the edges of your fog of war, you might think that you must “tread without sound” while exploring the universe. I.e. you should avoid other players. Hey, you’d be in good company, sci-fi author Liu Cixin thought that, and he inspired the game!

Surely you wouldn’t want to discover a player more advanced than you, with more planets and energy, and have them conquer you. And on the flipside, if you encountered a less progressed player you’d crush them under your thumb!

It turns out, it’s not that simple.

The multiplayer metagame is where the strength of Dark Forest’s game design emerges. The current round of the game offers a few interesting levers to reign in more progressed players and generate more sophisticated strategies than “more planets = more energy = me strong!!”. (Note again that these mechanics and the win conditions change from round to round. There’s a good write-up of a previous round’s strategy).

First, engaging with similarly progressed players is encouraged through a “junk” cap on how many planets you have conquered. This is similar to “population” in StarCraft, except, sorry, you can’t construct additional pylons here! Your junk limit is fixed at the start of the round, and each planet comes with junk if it’s unoccupied. I.e., players are incentivized to take already-conquered planets since they’re free of junk. Otherwise players must abandon old, smaller planets to free their junk limit and expand. Since smaller conquered planets typically have their silver and artifacts removed, this mainly pushes similarly progressed players to war over conquered-but-unmined planets.

Second, actions take time to perform. No, not because there’s a lot of clicking (there is, but more on automation with the plugin system in a bit). Once you decide to send a resource – energy, silver, artifacts, or ships – to another planet, actually sending that resource takes real world time proportional to the distance it must travel. In the early game, this is on the order of a few dozen seconds. In the mid-game, this turns into tens of minutes! There’s no worse feeling than knowing you’ve executed a poor game plan while there’s still 20 minutes before your stuff has arrived. The high latency between actions and outcomes rewards sophisticated planners and not the ones with the highest actions per minute APM (sorry MarineKing, this one ain’t for you).

Third, there is a “Broadcast” mechanic that allows a player to reveal the location of a planet through the fog of war, whether owned or unowned by them, every few hours. This leads to emergent meta-strategies around large resources. Need a bit of help to make sure a nearby expanding neighbor has to worry about multiple encroaching players? If you’re ranked 64 and 65, why not join forces if you stumble upon a top-63 player?

Fourth, taking an action costs real-world money. Don’t worry, I won’t be shilling a shitcoin and I’ll explain what the payment stuff is like next. The important thing to know is that choosing the right action matters, because you’re wasting money otherwise.

Yes, you have to pay to play Dark Forest. But you pay in a USD stablecoin (xDai) so the price doesn’t fluctuate and you aren’t pumping some trending coin on Twitter. For the current round one action (e.g. sending energy to a planet) costs $0.002 (2/10ths of a cent). I paid ~$0.50 to play for ten hours but it’ll depend on your gameplay. All players received $0.15 at the start of the round from the devs. I put $10 into my wallet to play when that ran out.

The money you pay doesn’t go back to the devs – it’s a “gas fee” you pay to save your action to the network (in this case, an Ethereum sidechain called the Gnosis Chain, formerly the xDai Chain).

If you enjoy the game, you can buy in-game hats to support the devs. Yes, hats, like the kind on your head or in Team Fortress 2. I bought a Fez hat for one of my planets for $1

Was it worth it? Hey – I report, you decide.

“Ugh… it’s crypto, just tell me how to make money playing this…” is probably what you’re thinking. Again, I’m not hiding anything – you really don’t make money playing Dark Forest. You really do have to enjoy it and find it worth paying to play.

Dark Forest is not play-to-earn (P2E) like other crypto games (e.g. Axie Infinity) are. The top ranked 63 players at the end of a round do get an NFT prize. Since it’s an NFT, it can be sold like any other NFT on the Ethereum network. It’s my understanding that in some previous early releases, the top players got a cash prize (paid in USD-equivalent crypto tokens), but this was a small amount ($1,000 pool).

Before you claim “NFT trading? that’s P2E!” I would ask you to think hard about what P2E means. The NFTs you earn after a round of Dark Forest are akin to an end-of-season reward on a competitive game’s ranked ladder. They’re a personal achievement. Is League of Legends P2E because you can sell a Challenger-tier (the highest competitive ranking) account for a few hundred dollars? While most players with a challenger account would be proud of their rank and never think to resell, some people really do make a living flipping high ranked accounts. E.g. a top ranked Korean League of Legends player, “Dopa”, has been banned from competitive play in Riot’s professional league for selling his high-ranked accounts.

Personally, I think a variation of P2E is inevitable in online games, like the freemium model before it. Dedicated player communities generate a lot of value for a game, whether through streaming, professional play, community art & music, guides, or other content. It’s up to publishers to figure out how best to craft the experience. Even Riot bootstrapped their professional scene by paying $25k/year salaries to players early on! That’s P2E! Just for a small subset of very skilled players.

If you still think Dark Forest and all games with secondary economies are P2E then sure, you can snip this section, add it to web3isgoinggreat.com and dunk me on your social media platform of choice! I’m making an argument of degree (the play-to-earn moniker depends on how a game is monetized and the player experience; just as a game with real-money items is not necessarily pay-to-win). You’re making an argument of kind (players making money off of a game makes it P2E). We fundamentally disagree on the definition of that phrase and you can evaluate whether Dark Forest is still worth playing on other terms.

Plugins are probably the single biggest feature I’m excited to spend “playtime” on in Dark Forest. Plugins let you write short Javascript programs to manipulate the game. You can find a list of them here, and they range anywhere from “show an outline around my planets” to “automatically conquer all planets fitting these criteria around a center point”.

By my estimation, most of the top players in Dark Forest have automated the bulk of their gameplay via plugins or writing programs that interact with the game’s backend directly. After playing for 10 hours, I only had 300,005 points, whereas other players had 2500x as much, despite only playing for 4 more days. Rewards aren’t linear with game time but they don’t scale that aggressively AFAICT.

You might be thinking “wait, if it can automate game actions… and this is browser based… isn’t that like cheating? And how does the game stop people from cheating? Isn’t the blockchain public? Can’t shadowy super coders data mine where I am?”.

This is the coolest part about how Dark Forest is built. How can the game 1) run in a browser, 2) on a public blockchain, 3) allow modding, but 4) have no cheating? (Ok, there are bugs - that’s not the same).

It’s because the player actions are encoded with something called “Zero Knowledge Proofs”. Yes, I can see your eyes glaze over – don’t worry, I’ve saved the nerdy stuff for the next section on blockchain & cryptography. The point is, even though your actions are recorded on a public blockchain, they’re hidden with a kind of cryptographically-secured game engine. So while you can automate your own actions, you can’t do more than “click faster” with automation -- the game’s backend can verify that the moves you submit must be legal and protects itself from denial of service attacks. And even with automation, you still need a guiding strategy to effectively compete with other players.

The other benefit of the game’s backend and state being public on-chain is that other games can plug into the Dark Forest ecosystem. Without having to construct a formal agreement with the Dark Forest team! This is equivalent to an open API, similar to Riot’s API for it’s titles, and is often referred to as “composability” in web3 parlance. If you’ve ever played an online game where your Twitch Prime subscription unlocks items or skins, a la Epic’s Fortnite or Riot’s Valorant, this effectively allows for that same experience but for any developer, without API rate limits, and unconstrained to the limited features the publisher exposes.

And this goes beyond extending the game’s capabilities -- since users are on-chain addresses, player guilds (in MMOs, guild are groups of allied players) can be formulated as Decentralized Autonomous Organizations (DAOs). Guilds can then re-appropriate tools to vote for on decisions! No more spamming in-game guild chat or impersonating players on private guild forums! Will Robinson has a full series on his experience as part of a Dark Forest guild.

This brings us to a more general point about mods. With all of the cool stuff you can build using mods, why is it so rare to find a rich mod framework for any new games, much less a multiplayer game these days? Well, for multiplayer games it makes sense, developers don’t want yet another vector of cheating or other more subtle competitive advantages. Riot has a thorough third party application policy for their games. Hell, they’ve been criticized for installing “a rootkit” on players’ computers to ensure Valorant has an impeccable anti-cheat system (it works well btw, no complaints from me).

But in giving up mods, I think the gaming community has lost out on an immeasurably large amount of rich new gaming experiences. Seriously, this isn’t hyperbole – it’s a hard counterfactual to prove, but the history of modding makes a strong case. Sure, Skyrim mods are amazing and add cool stuff. But the most revelatory story is the birth of League of Legends, DOTA, and the MOBA genre.

DOTA and League of Legends came from a series of mods of Warcraft 3, notably DOTA: All Stars by a developer named Guinsoo. The mod was eventually passed on to another, pseudonymous developer named IceFrog. Guinsoo eventually went on to work at Riot Games on League of Legends and IceFrog went to work at Valve on Dota 2. Today Riot’s League of Legends boasts 180 million monthly players and more than a billion in revenue. It’s at the vanguard of the esports scene with hundreds of professional players globally. And it all started from a mod and some pseudonymous devs circa 2005!

The ironic, unfortunate thing is that Dota 2 and League of Legends don’t have mods. Players are still creative – the popular “All Random All Mid (ARAM)” mode in League of Legends started as an unofficial custom game format in the community before getting productized. What new phenoms would we have if these games turned themselves into true platforms?

All of the prior stuff may have you excited, but still wondering “...so why is this built on that Big Bad Blockchain thing again? I’d like my earthly oceans unboiled, thank you.” I’ll go deeper here. Fair warning this is going to get nerdy.

First, the game runs on the Gnosis Chain which is Proof of Stake. No boiling oceans here! If you don’t know what that means, instead of crunching numbers (Proof of Work), the network is secured by people dumping a million bucks onto a computer in the form of tokens and saying “I won’t lie, fingers crossed winky face!~ Oh but if I do then you get to take all my money!~”. The network is Ethereum-compatible so when Ethereum swaps over to Proof of Stake, the game may move off of this sidechain. It doesn’t really matter to players either way.

Second, the game’s designed to run on basically no developer-owned servers. The game logic runs only on players’ machines and the Gnosis blockchain. Turns out the devs ain’t paying the infrastructure bills because they don’t have any. They do have some nominal bills for their website I’m sure, but the actual backend logic that you’d normally have to run in a data center isn’t costing them anything. $0.

Concurrent interacting players is something MMOs normally struggle with, because it’s a hard problem (you have to resolve O(N^2) player interactions for N players). Developers use tricks to get around this like having small zones, making some actions unimportant like player collision, and splitting playerbases into different instances. Dark Forest is no different -- blockchains don’t have high throughput. The game’s use of “Zero Knowledge Proofs” (ZKPs) affords them a clever workaround. We’ll touch on these more in the next sections, but the main problem here is that Dark Forest must support many concurrent players within a round, but they can’t validate every little interaction a player makes like a real-time MMO a la World of Warcraft does.

Server costs are real, ask any developer that supports a few thousand players! Even a simple Minecraft server with 20 players can you a few thousand in initial setup costs and a few dozen in recurring costs. And this will scale up, depending on how many players you want to support. Dark Forest elegantly skirts recurring infrastructure costs via it’s game design and using a blockchain for it’s backend.

In a communication with the Dark Forest dev team, they told me they spent more on Github than they did on any cloud costs (e.g. their blog/static site)! And this won’t suddenly start going up when they scale to more than a few hundred players!

If you’re a security-minded developer there’ve probably been some alarm bells going off in your head. Maybe akin to, “the game accepts actions client-side… in a web browser… this guy is fucking clown. Why have I read this far?”. Hold on. To explain how the game accepts rich actions executed in a player’s browser without cheating, we need to touch on Zero Knowledge Proofs (ZKPs).

Every action a player takes in Dark Forest produces a ZKP (a string of bits that follow some cryptographic properties). The ZKP is submitted to the game’s backend, where it’s cryptographically verified to have followed the game’s logic. This allows for richer actions to be taken client-side  without having a ton of chatter between front- and backends of the game. Here’s how it works in practice:

First, a ZKP lets another computer verify that a specific kind of computation was performed. In particular, when constructing a ZKP, a computation must match a pre-specified boolean logic circuit. For example, you can produce a ZKP of an addition of two binary numbers. This is the “proof” part of a ZKP. This lets each player run rich logic for their gameplay client-side, instead of sending individual inputs or commands to a central server that validates each command separately.

In the case of Dark Forest, game logic includes things like checking whether sending energy to another planet is valid by ensuring the target planet’s location falls within a certain valid radius. The developers wrote a blog post that goes into more detail about their ZKP circuits.

Second, the ZKP does not reveal any other information about the computation. Seriously, nothing else. Not the inputs. Not the outputs. For example, you can prove you’ve calculated w=(x+y)*z without telling me what w, x, y, z or the answer are. This is the “zero knowledge” part of a ZKP. The backend verifies the client’s ZKP to ensure it was valid.

Yes, this is mind-melting.

The combination of a decentralized database (here, a public blockchain that stores the game state but that can’t process a ton of real-time simple inputs) and decentralized compute (ZKPs created in players’ web browsers) lets the game run in a performant manner, without cheating, and with mods.

That’s the magic behind how Dark Forest works.

In case this isn’t quite clicking & you’re a curious developer, I recommend this repo of Battleship done via ZKPs by Kunal Modi, a colleague from South Park Commons. He also has a write-up.

I had another paragraph here explaining ZKPs with an analogy of a computer crunching numbers and how only you, as the chef of those crunchy numbers, could appreciate them. I axed it. No crunchy numbers today.

Sigh… math seems cool until you have to learn it well enough to teach it…

Does this sound too magical? A little too good to be true? Wouldn’t ZKPs solve a few other privacy problems? Maybe they could put an end to those annoying cookie banner ads? Why don’t all multiplayer games use them to prevent cheating? In fact, shouldn’t these things be used everywhere? At least for a few other useful things? Maybe proving that Berenstein Bears is really spelled Berenstain Bears?

The truth is, they’re legitimately new technology. Not new as in “Gee willikers, that internet thing is new!”. New as in the compilers and software to create ZKP circuits are 2 years old at the time of writing. Sorry, #TeamBerenstAin – the past is the past but the future can still be saved.

The price of casting a spell with this arcane, dark magick technology is… computations take a really really long time. Like, seconds to do basic stuff like arithmetic, even on today’s hardware. If you can write code, try to go through a SnarkJS or Circom tutorial or run a sample app. I instrumented & ran this repo of Battleship done via ZKPs so you can see for yourself.

ZKPs are slow. Devs needs to be clever about their usage. Imagine if adding a few numbers took a few seconds. We’d be back like 100 years in terms of compute power? Yikes.

Yet that’s the price you pay if you want a massively multiplayer game that can run on player machines without cheating, at no marginal infrastructure cost to the developers.

It’s good news! Culture isn’t stuck and the internet isn’t dead. Yaaaay! Ok so it’s slow, but we can make new hardware and software to make it fast. And things will get better, not just for games, but for all types of privacy-preserving applications. No one needs to shill Monero any more, woot woo (kidding, no one knows what Monero is any more)!

Remember my planet with the fez hat? Dark Forest using hats to monetize is a proof-point for how the development team is made of passionate gamers. Yes, it’s an in-joke that references another game that only some will understand. Even if you don’t understand the joke, it’s a signal that the team is not a fan of “pay-to-win” mechanics. This is despite the frothy market for such mechanics in existing games, crypto or not (enjoy paying for another DLC season pass for a $60 AAA game or buying your first Axie! Ok, Axie prices seem to have gone down, but I’m still not playing.).

You can also read their blog to find plenty of community-oriented educational content. Not just on game design decisions! They’re also bringing the cutting edge, dark magic whose mere utterance turns an anonymous Twitter anime profile-pic goon into an S-tier technologi– ok I’ll stop. They’re making Zero Knowledge Proofs more accessible to the broader software development community. Yes, you can cynically say “that's just content marketing for recruiting!” But it’s still great educational content! Plenty of tech companies have useful write-ups on their blogs. I benefitted greatly from them as a young engineer in Silicon Valley! And again, the team’s funded via donations and grants – check their funding page or their Gitcoin. No implicit growth mandate! Yaaaay regenerative economics.

Aside from blog content, there was an IRL meet-up at ETHDenver (which in dramatic irony I didn’t attend since 1) I just received my beta key, but 2) I had to fly to ETHDenver!). The team is active on their Discord. The pace of development and beta access shows they’d rather craft a great game for a few hundred players and make sure it can scale up before showing it to everyone.

There’s a famous Shigeru Miyamoto quote about how  “a delayed game is eventually good, but a rushed game is forever bad”. You may believe the quote is outdated since we’re in a world of online-connected games. Weekly or monthly patches that adjust gameplay are common. Hell, Riot practically re-released League of Legends 3 times with all of the visual and gameplay updates they shipped over the past 10 years! Surely a development team can launch early and fix a game based on player feedback – that’s the Lean Startup way!

(Here’s the above ice cream recipe, btw.)

There are subtleties and limits to the iterative approach of building products, games included, that make it hard for teams to get right. Development teams (even at mature, successful companies) can have egos, limited budgets, or a weak community muscle – it’s true! Anyone who has played both CS:GO (by Valve) and Valorant (by Riot) will tell you that a team’s specific approach to tweaking their game matters, and not all are equally skilled at interacting with their player communities. In a competitive game it’s particularly important. It’s the difference between a stale, solved game that only the most dedicated players can enjoy, and an interesting game that continuously appeals to all skill levels with a flourishing grassroots esports scene. (For anyone unfamiliar, Riot is better at this, don’t @ me).

The game is in closed beta so yes, there are issues.

If everything above sounded super nerdy, even for gamers, you might think Dark Forest is more like Eve Online than League of Legends. A cool part of internet gaming subculture? Sure. But not for the masses, just like blockchains.

During a previous round the Gnosis Chain was congested and couldn’t handle traffic, so players started running their own blockchain infrastructure nodes! Would the average Eve player do that? Maybe. The average League player? Eh, they’d sooner be caught running it down mid than running a piece of infra.

But there are also legitimate problems with the game as-is. The frame rate isn’t great. The community wiki isn’t fully up-to-date. The number of players and invite key list is small so you can’t talk to your friends about it. The new player experience is confusing (I personally re-onboarded 2 other new players at South Park Commons who were discouraged after onboarding friction). Half the plugins I tried didn’t work. Technically the way they store the wallet you use to pay gas fees isn’t super secure (it uses localstorage – if that is gibberish to you just know it has pitfalls but isn’t the end of the world).

But the kernel of the game and the direction of the team is enough for me to get invested (no no, not with a shitcoin, with my time) as a player. I trust the team to carry the game forward and listen to the player community.

More importantly, they’ve given me renewed optimism on the promise of web3 – that even if it’s not for you today, it will come to you in due time. Not because you can make a quick buck, but because the new experiences created will be worth it!

And if you want to build the next Dark Forest, hit me up at omar@exaverse.io :)

Acknowledgements

I’d like to thank Cameron Tuckerman-Lee & Stan Kirdey from South Park Commons and @gubsheep & Ivan Chub (@chubivan) from the Dark Forest team for helpful feedback on this post. I’d also like to thank Dan Romero (@dwr) for introducing me to the team.

Dark Forest is a massively multiplayer (MMO) real-time strategy (RTS) game in closed beta. If you like strategy games, you should play it. Anyone who is a fan of games like Sins of a Solar Empire, Factorio, or Civilization should definitely play it. It’s not a game with global appeal like Super Mario Odyssey, but it doesn’t need to be nor try to be.

Personally, it's the best new game I’ve played in 2022. When I got off the waitlist I played it for 10 hours straight -- that’s why I’m writing this post. It’s the most exciting strategy game I’ve played since StarCraft 2. It’s the most exciting competitive multiplayer game I’ve played since Valorant. And it’s renewed my optimism for the potential of web3 games. Why? It gets many things right, but it gets a few things very, very right:

• It is a refreshing take on a stale genre

• It has a competitive ladder with a fresh take on ranking

• It’s plugin system & open source codebase brings modding back to games

• It shakes up assumptions around the infrastructure costs of online multiplayer games

• It has a thoughtful team that closely engages with the community

Intrigued? Great, read on! But there is one minor detail you might care to know… Dark Forest runs on “The Blockchain”.

Blockchains and cryptocurrencies are a divisive topic in recent times, but especially so in the gaming community. Initial promises of a cross-platform utopia with transferable skins and items quickly gave way to a wave of “crypto games” with terrible monetization schemes and high profile hacks. These, along with other narratives around the most popular blockchains using energy-guzzling Proof of Work algorithms and NFTs destroying the environment, have poisoned the well for any game touting blockchain-adjacent features.

The negative reaction from the gaming community to companies adding blockchain features, like Discord teasing an Ethereum integration, makes sense! The cut runs deeper and older than crypto gaming’s recent spotlight – it triggers all-too-fresh memories of the gaming industry's parlous relationship with new types of monetization. The sting of in-game ads, base game content behind paid DLC, and hellish microtransaction models is still felt in the community. And while some developers, like the SW: Battlefront 2 devs, had a change of heart, it’s been a rocky journey. Misguided attempts to win over players notwithstanding, the community still finds itself with games like The Sims 4 offering $800 of paid DLC.

The good news is that Dark Forest is not trying to sell you anything, nor is it an abusively monetized game. It is a great game that just happens to use new technology in cryptography and blockchains to their advantage, rather than shoehorn them in.

It also isn’t funded through venture capital, instead relying on community and non-profit funding sources. So don’t worry Jack, no a16z or VCs on their GitCoin page or funding page! Raising venture capital is fine (I’ve done it!), but in this case the developers truly have put the game and the community first.

You can skip this part if you just want to read about Dark Forest; it just contextualizes my motivation.

I’m a huge gamer (hello, fellow kids!). I’ve religiously grinded the online competitive ladders of StarCraft 2, League of Legends, and Valorant. You can stat check my Steam profile if you want to know more about my tastes.

I’m also a developer and a founder in the web3 space. I think I’m in a unique position to speak about the web3 gaming experience. I know a lot more needs to be delivered before reasonable skeptics can see past crypto’s current failings. I’m an optimist that believes eventually cryptography and web3 tech can re-empower users on web platforms, and I’m trying to build that reality.

And if you don’t believe me, again, check my gaming habits! I promise Axie Infinity is nowhere to be found in my wallet. I did spend a month as a child catching all 150 Pokemon in the original games with a link cable, 2 Gameboys, and the holy trinity of Red / Blue / Yellow versions. I’ve got no on-chain proof for that, though… only PTSD of the Safari Zone….

Dark Forest is a browser-based 2.5D (2D plane, 3D assets) game that takes place in “rounds”, with each round lasting a few days (the current round is underway and will last 10 days). If the following in-game image makes your eyes bleed, the gameplay might still make you stay!

Generally, a few months pass between rounds, which players use to plan for the next round. The current round in February preceded by one last year in October, and game mechanics are added between rounds. While a round is underway, players try to earn the highest score on the leaderboard, viewable in real-time on the homepage. (Yeah, I’m not doing too hot. And yeah, the CEO of Figma won the first round). The top ranked 63 players at the end of a round get a reward from the devs, but otherwise you just play for fun.

The core game loop is:

• All players spawn with a home planet and vision over a small area in their corner of the universe, with fog of war (i.e. no vision) covering the map

• Planets have energy that regenerates over time

• You send energy from your planets to other planets to conquer them

• Some planets have “silver” and “artifacts” which give you points that determine your rank

• You break the fog of war with an “explorer” whose speed is dictated by the compute power of the real life, physical computer you’re playing on

If it all sounded great until that last point, hear me out. No, this isn’t some shadow-realm version of Boktai: The Sun is in Your Hand where you destroy the environment to play. Nor is the game pay-to-win. Nor is it using your browser to mine Monero.

First, AFAICT the game’s miner seems to have fairly limited scaling capability, with a Raspberry Pi giving reasonable performance (1-2k hashes/sec is roughly equivalent to my browser using a Ryzen 3600 XT & 2 cores). On the backend the game runs on the Gnosis Chain (an Ethereum sidechain — don’t worry if this means nothing to you for now), which is Proof of Stake, not Proof of Work. Shadow-realm Hideo Kojima still has a game to make.

Second, yes – faster computers have an advantage. It’s not as big as you’d think (running on 16 cores vs 2 cores is ~2.5x better, and 16 vs 4 is ~1.7x, so diminishing returns kick in). The bigger advantage comes from using a “remote hasher”, which is a separate server that does hashing. But in the early game, you’re mostly limited by your strategy, not the fog of war. In the mid-game, players need to direct their explorer’s search regardless of how much compute they have. If I were to make an analogy to other competitive online games, having a large amount of compute would be like playing a First-Person Shooter a la CS:GO or Valorant with a 240 Hz monitor and a low-latency connection. It makes a difference, but anyone can upgrade their gear to be at their competitive best. You can still have fun and be competitive without top gear.

Third, if you’re afraid the game might be using your computer for nefarious purposes, it’s all open source (both the code and the live smart contracts and frontend). If you don’t have the skill to audit it, then yes, you’ll have to trust a friend or someone else to do it for you. When I quickly read through it, it looked good, promise. You won’t be sending any excellent privacy coins to anyone.

(That last part was a joke, please don’t buy or mine Monero if you’re trying to make a quick buck.)

There are other mechanics like upgrading planets, a planet leveling system (you progressively conquer larger planets as you expand), and “artifacts” that boost your planets’ creation of energy and other resources. These feel natural and are staples of any good strategy game, but the tech tree is less developed than the richness of a Civilization entry, and there isn’t customization on the level of Stellaris.

What Dark Forest lacks in more traditional strategy game features it makes up for in competitive Player vs Player (PvP) design, though.

Dark Forest’s ladder and competitive rewards are sort of similar to other online games like League of Legends. There’s a publicly viewable leaderboard and end-of-season ranked rewards. What’s unique about Dark Forest’s leaderboard is that the game doesn’t use an “Elo system” or feature any skill-based matchmaking at all, which is a departure from the most popular competitive games today. Instead, all players compete with all other players within a round. This may not seem like it sets up the average player for a good time, but we’ll touch on how this doesn’t lead to new players getting Goomba stomped by more skilled ones.

Players earn points by interacting with the game universe – either discovering “artifacts” on conquered planets, or exchanging “silver” for points. Players interact with other players’ planets using the same moves as they would an unconquered planet. The only difference is that players can beef up their planets’ defense so it requires more energy to capture. So rather than a “Player vs. Environment” (PvE) MMO experience, the resources themselves are a battleground for “Player vs. Player” (PvP) gameplay.

Note that the exact competitive mechanics change between rounds, for e.g. there has also been a “race to the center” round, which keeps the game fresh.

This is where the titular mechanic of “a dark forest” comes in. Since there are finite resources and you expand to the edges of your fog of war, you might think that you must “tread without sound” while exploring the universe. I.e. you should avoid other players. Hey, you’d be in good company, sci-fi author Liu Cixin thought that, and he inspired the game!

Surely you wouldn’t want to discover a player more advanced than you, with more planets and energy, and have them conquer you. And on the flipside, if you encountered a less progressed player you’d crush them under your thumb!

It turns out, it’s not that simple.

The multiplayer metagame is where the strength of Dark Forest’s game design emerges. The current round of the game offers a few interesting levers to reign in more progressed players and generate more sophisticated strategies than “more planets = more energy = me strong!!”. (Note again that these mechanics and the win conditions change from round to round. There’s a good write-up of a previous round’s strategy).

First, engaging with similarly progressed players is encouraged through a “junk” cap on how many planets you have conquered. This is similar to “population” in StarCraft, except, sorry, you can’t construct additional pylons here! Your junk limit is fixed at the start of the round, and each planet comes with junk if it’s unoccupied. I.e., players are incentivized to take already-conquered planets since they’re free of junk. Otherwise players must abandon old, smaller planets to free their junk limit and expand. Since smaller conquered planets typically have their silver and artifacts removed, this mainly pushes similarly progressed players to war over conquered-but-unmined planets.

Second, actions take time to perform. No, not because there’s a lot of clicking (there is, but more on automation with the plugin system in a bit). Once you decide to send a resource – energy, silver, artifacts, or ships – to another planet, actually sending that resource takes real world time proportional to the distance it must travel. In the early game, this is on the order of a few dozen seconds. In the mid-game, this turns into tens of minutes! There’s no worse feeling than knowing you’ve executed a poor game plan while there’s still 20 minutes before your stuff has arrived. The high latency between actions and outcomes rewards sophisticated planners and not the ones with the highest actions per minute APM (sorry MarineKing, this one ain’t for you).

Third, there is a “Broadcast” mechanic that allows a player to reveal the location of a planet through the fog of war, whether owned or unowned by them, every few hours. This leads to emergent meta-strategies around large resources. Need a bit of help to make sure a nearby expanding neighbor has to worry about multiple encroaching players? If you’re ranked 64 and 65, why not join forces if you stumble upon a top-63 player?

Fourth, taking an action costs real-world money. Don’t worry, I won’t be shilling a shitcoin and I’ll explain what the payment stuff is like next. The important thing to know is that choosing the right action matters, because you’re wasting money otherwise.

Yes, you have to pay to play Dark Forest. But you pay in a USD stablecoin (xDai) so the price doesn’t fluctuate and you aren’t pumping some trending coin on Twitter. For the current round one action (e.g. sending energy to a planet) costs $0.002 (2/10ths of a cent). I paid ~$0.50 to play for ten hours but it’ll depend on your gameplay. All players received $0.15 at the start of the round from the devs. I put $10 into my wallet to play when that ran out.

The money you pay doesn’t go back to the devs – it’s a “gas fee” you pay to save your action to the network (in this case, an Ethereum sidechain called the Gnosis Chain, formerly the xDai Chain).

If you enjoy the game, you can buy in-game hats to support the devs. Yes, hats, like the kind on your head or in Team Fortress 2. I bought a Fez hat for one of my planets for $1

Was it worth it? Hey – I report, you decide.

“Ugh… it’s crypto, just tell me how to make money playing this…” is probably what you’re thinking. Again, I’m not hiding anything – you really don’t make money playing Dark Forest. You really do have to enjoy it and find it worth paying to play.

Dark Forest is not play-to-earn (P2E) like other crypto games (e.g. Axie Infinity) are. The top ranked 63 players at the end of a round do get an NFT prize. Since it’s an NFT, it can be sold like any other NFT on the Ethereum network. It’s my understanding that in some previous early releases, the top players got a cash prize (paid in USD-equivalent crypto tokens), but this was a small amount ($1,000 pool).

Before you claim “NFT trading? that’s P2E!” I would ask you to think hard about what P2E means. The NFTs you earn after a round of Dark Forest are akin to an end-of-season reward on a competitive game’s ranked ladder. They’re a personal achievement. Is League of Legends P2E because you can sell a Challenger-tier (the highest competitive ranking) account for a few hundred dollars? While most players with a challenger account would be proud of their rank and never think to resell, some people really do make a living flipping high ranked accounts. E.g. a top ranked Korean League of Legends player, “Dopa”, has been banned from competitive play in Riot’s professional league for selling his high-ranked accounts.

Personally, I think a variation of P2E is inevitable in online games, like the freemium model before it. Dedicated player communities generate a lot of value for a game, whether through streaming, professional play, community art & music, guides, or other content. It’s up to publishers to figure out how best to craft the experience. Even Riot bootstrapped their professional scene by paying $25k/year salaries to players early on! That’s P2E! Just for a small subset of very skilled players.

If you still think Dark Forest and all games with secondary economies are P2E then sure, you can snip this section, add it to web3isgoinggreat.com and dunk me on your social media platform of choice! I’m making an argument of degree (the play-to-earn moniker depends on how a game is monetized and the player experience; just as a game with real-money items is not necessarily pay-to-win). You’re making an argument of kind (players making money off of a game makes it P2E). We fundamentally disagree on the definition of that phrase and you can evaluate whether Dark Forest is still worth playing on other terms.

Plugins are probably the single biggest feature I’m excited to spend “playtime” on in Dark Forest. Plugins let you write short Javascript programs to manipulate the game. You can find a list of them here, and they range anywhere from “show an outline around my planets” to “automatically conquer all planets fitting these criteria around a center point”.

By my estimation, most of the top players in Dark Forest have automated the bulk of their gameplay via plugins or writing programs that interact with the game’s backend directly. After playing for 10 hours, I only had 300,005 points, whereas other players had 2500x as much, despite only playing for 4 more days. Rewards aren’t linear with game time but they don’t scale that aggressively AFAICT.

You might be thinking “wait, if it can automate game actions… and this is browser based… isn’t that like cheating? And how does the game stop people from cheating? Isn’t the blockchain public? Can’t shadowy super coders data mine where I am?”.

This is the coolest part about how Dark Forest is built. How can the game 1) run in a browser, 2) on a public blockchain, 3) allow modding, but 4) have no cheating? (Ok, there are bugs - that’s not the same).

It’s because the player actions are encoded with something called “Zero Knowledge Proofs”. Yes, I can see your eyes glaze over – don’t worry, I’ve saved the nerdy stuff for the next section on blockchain & cryptography. The point is, even though your actions are recorded on a public blockchain, they’re hidden with a kind of cryptographically-secured game engine. So while you can automate your own actions, you can’t do more than “click faster” with automation -- the game’s backend can verify that the moves you submit must be legal and protects itself from denial of service attacks. And even with automation, you still need a guiding strategy to effectively compete with other players.

The other benefit of the game’s backend and state being public on-chain is that other games can plug into the Dark Forest ecosystem. Without having to construct a formal agreement with the Dark Forest team! This is equivalent to an open API, similar to Riot’s API for it’s titles, and is often referred to as “composability” in web3 parlance. If you’ve ever played an online game where your Twitch Prime subscription unlocks items or skins, a la Epic’s Fortnite or Riot’s Valorant, this effectively allows for that same experience but for any developer, without API rate limits, and unconstrained to the limited features the publisher exposes.

And this goes beyond extending the game’s capabilities -- since users are on-chain addresses, player guilds (in MMOs, guild are groups of allied players) can be formulated as Decentralized Autonomous Organizations (DAOs). Guilds can then re-appropriate tools to vote for on decisions! No more spamming in-game guild chat or impersonating players on private guild forums! Will Robinson has a full series on his experience as part of a Dark Forest guild.

This brings us to a more general point about mods. With all of the cool stuff you can build using mods, why is it so rare to find a rich mod framework for any new games, much less a multiplayer game these days? Well, for multiplayer games it makes sense, developers don’t want yet another vector of cheating or other more subtle competitive advantages. Riot has a thorough third party application policy for their games. Hell, they’ve been criticized for installing “a rootkit” on players’ computers to ensure Valorant has an impeccable anti-cheat system (it works well btw, no complaints from me).

But in giving up mods, I think the gaming community has lost out on an immeasurably large amount of rich new gaming experiences. Seriously, this isn’t hyperbole – it’s a hard counterfactual to prove, but the history of modding makes a strong case. Sure, Skyrim mods are amazing and add cool stuff. But the most revelatory story is the birth of League of Legends, DOTA, and the MOBA genre.

DOTA and League of Legends came from a series of mods of Warcraft 3, notably DOTA: All Stars by a developer named Guinsoo. The mod was eventually passed on to another, pseudonymous developer named IceFrog. Guinsoo eventually went on to work at Riot Games on League of Legends and IceFrog went to work at Valve on Dota 2. Today Riot’s League of Legends boasts 180 million monthly players and more than a billion in revenue. It’s at the vanguard of the esports scene with hundreds of professional players globally. And it all started from a mod and some pseudonymous devs circa 2005!

The ironic, unfortunate thing is that Dota 2 and League of Legends don’t have mods. Players are still creative – the popular “All Random All Mid (ARAM)” mode in League of Legends started as an unofficial custom game format in the community before getting productized. What new phenoms would we have if these games turned themselves into true platforms?

All of the prior stuff may have you excited, but still wondering “...so why is this built on that Big Bad Blockchain thing again? I’d like my earthly oceans unboiled, thank you.” I’ll go deeper here. Fair warning this is going to get nerdy.

First, the game runs on the Gnosis Chain which is Proof of Stake. No boiling oceans here! If you don’t know what that means, instead of crunching numbers (Proof of Work), the network is secured by people dumping a million bucks onto a computer in the form of tokens and saying “I won’t lie, fingers crossed winky face!~ Oh but if I do then you get to take all my money!~”. The network is Ethereum-compatible so when Ethereum swaps over to Proof of Stake, the game may move off of this sidechain. It doesn’t really matter to players either way.

Second, the game’s designed to run on basically no developer-owned servers. The game logic runs only on players’ machines and the Gnosis blockchain. Turns out the devs ain’t paying the infrastructure bills because they don’t have any. They do have some nominal bills for their website I’m sure, but the actual backend logic that you’d normally have to run in a data center isn’t costing them anything. $0.

Concurrent interacting players is something MMOs normally struggle with, because it’s a hard problem (you have to resolve O(N^2) player interactions for N players). Developers use tricks to get around this like having small zones, making some actions unimportant like player collision, and splitting playerbases into different instances. Dark Forest is no different -- blockchains don’t have high throughput. The game’s use of “Zero Knowledge Proofs” (ZKPs) affords them a clever workaround. We’ll touch on these more in the next sections, but the main problem here is that Dark Forest must support many concurrent players within a round, but they can’t validate every little interaction a player makes like a real-time MMO a la World of Warcraft does.

Server costs are real, ask any developer that supports a few thousand players! Even a simple Minecraft server with 20 players can you a few thousand in initial setup costs and a few dozen in recurring costs. And this will scale up, depending on how many players you want to support. Dark Forest elegantly skirts recurring infrastructure costs via it’s game design and using a blockchain for it’s backend.

In a communication with the Dark Forest dev team, they told me they spent more on Github than they did on any cloud costs (e.g. their blog/static site)! And this won’t suddenly start going up when they scale to more than a few hundred players!

If you’re a security-minded developer there’ve probably been some alarm bells going off in your head. Maybe akin to, “the game accepts actions client-side… in a web browser… this guy is fucking clown. Why have I read this far?”. Hold on. To explain how the game accepts rich actions executed in a player’s browser without cheating, we need to touch on Zero Knowledge Proofs (ZKPs).

Every action a player takes in Dark Forest produces a ZKP (a string of bits that follow some cryptographic properties). The ZKP is submitted to the game’s backend, where it’s cryptographically verified to have followed the game’s logic. This allows for richer actions to be taken client-side  without having a ton of chatter between front- and backends of the game. Here’s how it works in practice:

First, a ZKP lets another computer verify that a specific kind of computation was performed. In particular, when constructing a ZKP, a computation must match a pre-specified boolean logic circuit. For example, you can produce a ZKP of an addition of two binary numbers. This is the “proof” part of a ZKP. This lets each player run rich logic for their gameplay client-side, instead of sending individual inputs or commands to a central server that validates each command separately.

In the case of Dark Forest, game logic includes things like checking whether sending energy to another planet is valid by ensuring the target planet’s location falls within a certain valid radius. The developers wrote a blog post that goes into more detail about their ZKP circuits.

Second, the ZKP does not reveal any other information about the computation. Seriously, nothing else. Not the inputs. Not the outputs. For example, you can prove you’ve calculated w=(x+y)*z without telling me what w, x, y, z or the answer are. This is the “zero knowledge” part of a ZKP. The backend verifies the client’s ZKP to ensure it was valid.

Yes, this is mind-melting.

The combination of a decentralized database (here, a public blockchain that stores the game state but that can’t process a ton of real-time simple inputs) and decentralized compute (ZKPs created in players’ web browsers) lets the game run in a performant manner, without cheating, and with mods.

That’s the magic behind how Dark Forest works.

In case this isn’t quite clicking & you’re a curious developer, I recommend this repo of Battleship done via ZKPs by Kunal Modi, a colleague from South Park Commons. He also has a write-up.

I had another paragraph here explaining ZKPs with an analogy of a computer crunching numbers and how only you, as the chef of those crunchy numbers, could appreciate them. I axed it. No crunchy numbers today.

Sigh… math seems cool until you have to learn it well enough to teach it…

Does this sound too magical? A little too good to be true? Wouldn’t ZKPs solve a few other privacy problems? Maybe they could put an end to those annoying cookie banner ads? Why don’t all multiplayer games use them to prevent cheating? In fact, shouldn’t these things be used everywhere? At least for a few other useful things? Maybe proving that Berenstein Bears is really spelled Berenstain Bears?

The truth is, they’re legitimately new technology. Not new as in “Gee willikers, that internet thing is new!”. New as in the compilers and software to create ZKP circuits are 2 years old at the time of writing. Sorry, #TeamBerenstAin – the past is the past but the future can still be saved.

The price of casting a spell with this arcane, dark magick technology is… computations take a really really long time. Like, seconds to do basic stuff like arithmetic, even on today’s hardware. If you can write code, try to go through a SnarkJS or Circom tutorial or run a sample app. I instrumented & ran this repo of Battleship done via ZKPs so you can see for yourself.

ZKPs are slow. Devs needs to be clever about their usage. Imagine if adding a few numbers took a few seconds. We’d be back like 100 years in terms of compute power? Yikes.

Yet that’s the price you pay if you want a massively multiplayer game that can run on player machines without cheating, at no marginal infrastructure cost to the developers.

It’s good news! Culture isn’t stuck and the internet isn’t dead. Yaaaay! Ok so it’s slow, but we can make new hardware and software to make it fast. And things will get better, not just for games, but for all types of privacy-preserving applications. No one needs to shill Monero any more, woot woo (kidding, no one knows what Monero is any more)!

Remember my planet with the fez hat? Dark Forest using hats to monetize is a proof-point for how the development team is made of passionate gamers. Yes, it’s an in-joke that references another game that only some will understand. Even if you don’t understand the joke, it’s a signal that the team is not a fan of “pay-to-win” mechanics. This is despite the frothy market for such mechanics in existing games, crypto or not (enjoy paying for another DLC season pass for a $60 AAA game or buying your first Axie! Ok, Axie prices seem to have gone down, but I’m still not playing.).

You can also read their blog to find plenty of community-oriented educational content. Not just on game design decisions! They’re also bringing the cutting edge, dark magic whose mere utterance turns an anonymous Twitter anime profile-pic goon into an S-tier technologi– ok I’ll stop. They’re making Zero Knowledge Proofs more accessible to the broader software development community. Yes, you can cynically say “that's just content marketing for recruiting!” But it’s still great educational content! Plenty of tech companies have useful write-ups on their blogs. I benefitted greatly from them as a young engineer in Silicon Valley! And again, the team’s funded via donations and grants – check their funding page or their Gitcoin. No implicit growth mandate! Yaaaay regenerative economics.

Aside from blog content, there was an IRL meet-up at ETHDenver (which in dramatic irony I didn’t attend since 1) I just received my beta key, but 2) I had to fly to ETHDenver!). The team is active on their Discord. The pace of development and beta access shows they’d rather craft a great game for a few hundred players and make sure it can scale up before showing it to everyone.

There’s a famous Shigeru Miyamoto quote about how  “a delayed game is eventually good, but a rushed game is forever bad”. You may believe the quote is outdated since we’re in a world of online-connected games. Weekly or monthly patches that adjust gameplay are common. Hell, Riot practically re-released League of Legends 3 times with all of the visual and gameplay updates they shipped over the past 10 years! Surely a development team can launch early and fix a game based on player feedback – that’s the Lean Startup way!

(Here’s the above ice cream recipe, btw.)

There are subtleties and limits to the iterative approach of building products, games included, that make it hard for teams to get right. Development teams (even at mature, successful companies) can have egos, limited budgets, or a weak community muscle – it’s true! Anyone who has played both CS:GO (by Valve) and Valorant (by Riot) will tell you that a team’s specific approach to tweaking their game matters, and not all are equally skilled at interacting with their player communities. In a competitive game it’s particularly important. It’s the difference between a stale, solved game that only the most dedicated players can enjoy, and an interesting game that continuously appeals to all skill levels with a flourishing grassroots esports scene. (For anyone unfamiliar, Riot is better at this, don’t @ me).

The game is in closed beta so yes, there are issues.

If everything above sounded super nerdy, even for gamers, you might think Dark Forest is more like Eve Online than League of Legends. A cool part of internet gaming subculture? Sure. But not for the masses, just like blockchains.

During a previous round the Gnosis Chain was congested and couldn’t handle traffic, so players started running their own blockchain infrastructure nodes! Would the average Eve player do that? Maybe. The average League player? Eh, they’d sooner be caught running it down mid than running a piece of infra.

But there are also legitimate problems with the game as-is. The frame rate isn’t great. The community wiki isn’t fully up-to-date. The number of players and invite key list is small so you can’t talk to your friends about it. The new player experience is confusing (I personally re-onboarded 2 other new players at South Park Commons who were discouraged after onboarding friction). Half the plugins I tried didn’t work. Technically the way they store the wallet you use to pay gas fees isn’t super secure (it uses localstorage – if that is gibberish to you just know it has pitfalls but isn’t the end of the world).

But the kernel of the game and the direction of the team is enough for me to get invested (no no, not with a shitcoin, with my time) as a player. I trust the team to carry the game forward and listen to the player community.

More importantly, they’ve given me renewed optimism on the promise of web3 – that even if it’s not for you today, it will come to you in due time. Not because you can make a quick buck, but because the new experiences created will be worth it!

And if you want to build the next Dark Forest, hit me up at omar@exaverse.io :)

Acknowledgements

I’d like to thank Cameron Tuckerman-Lee & Stan Kirdey from South Park Commons and @gubsheep & Ivan Chub (@chubivan) from the Dark Forest team for helpful feedback on this post. I’d also like to thank Dan Romero (@dwr) for introducing me to the team.