Hacker News Reader

Comment by theshrike79 | Hacker News Reader

Backend. Do the game backend stuff for online/mobile games.

Everyone and their sister wants to be the one that does the pretty graphics and UI with Unity or whatever. Nobody wants to be the one who does the backend code that actually makes the game work.

It's actually challenging when you work with games that might have a million daily users or even more. If it's a competitive game, people will try to cheat and you can't trust anything the client sends you. The code also needs to be cost-effective to run, if your game makes $10k a day, but the backend costs $5k a day to run - that's kinda bad.

Also especially mobile games need a custom frontend for customer service so they can do refunds, check player purchase history etc. Those are done with JS and Python etc.

Source: Moved from traditional tech to making CS tools and backend for games. My salary went up hilarious amounts, as did bonuses and options.

View on HN

>Nobody wants to be the one who does the backend code that actually makes the game work.

This is only true for a certain class of games that rely on multiplayer. Many games are singleplayer where the meat exists on the game engine. There's two sub sections of this as well. Real time and non-real time.

Realtime means streaming updates from all users. Games like fortnight or pubg are like this. Non-realtime means like checkers or multiplayer chess. If you are doing the later rather then the former I would say you're still in traditional web development where shit is still relatively easy because an HTTP server is good enough.

The former is what is really hard, and that kind of thing is what I would refer to as backend game development. If you are doing that, then likely you're not a traditional web developer anymore.

Typically in both cases Backend as a role will refer more to people creating and running services like server orchestration, matchmaking and so on. So the real-time/non-real-time distinction doesn’t really matter. The engine level role is usually Multiplayer or Networking Engineer and the actually making the game role is Gameplay Engineer.

It does matter. For real-time the person will likely need a custom solution. That means coding the entire server in C++ or some other zero cost language with custom protocols.

For the non-real-time, it means using a pre-existing backend server.

Many backend engineers don't know how to write a custom real time server from scratch. It's two completely different levels of skill. If you are hiring for this role and put out of job ad for typical backend engineers you won't get the right people for the role because, again, backend engineers typically don't have this skill.

Right that’s what I’m saying, the game server and associated protocols will be designed and written by the engine team in close collaboration with the gameplay team that will be making the game.

Yes, you're right. My mistake.

Are there any billion dollar game franchises that lack an online component? Feels like a safe bet to me.

God of war, Cyberpunk 2077, witcher 3, skyrim, fallout, Zelda, Tomb Raider, Metal Gear, Final Fantasy (not the mmorpgs), Assasins Creed, there's tons.

I wouldn't say they are billion dollar game franchises but I would say billions and billions of dollars in total are part of games where multiplayer is not the central aspect. This sector is huge.

I think games with a huge online component tend to be bigger franchises but fewer in number. Overall online gaming as an industry is bigger then single player gaming but they are both huge enough in size that they are comparable.

Witcher 3 shot CDPR from under a billion to over 8 billions market value. So yeah, that game single-handedly made that, up to the point when CDPR dropped the ball with the buggy release of Cyberpunk 2077 - but nowadays they released a nice update for CP2077 which is what should've been the actual game launch 2 years ago.

But in the context of career advice, it still seems like a decent bet that the online gaming industry is and will remain pretty robust. Especially for people transitioning from traditional 'enterprise' programming jobs.

Why specifically the focus on online gaming vs. gaming in general? I would imagine the "safety" of both travel in lock step. And I wouldn't think gaming is "safer" then other fields of programming.

In fact even in good times, I feel gaming is less safe.

That was the point of contention upthread. Someone said doing backends for online games was a good transition for a general programmer, which intuitively seems plausible. The concern was whether that was too niche but the size of that market implies otherwise to me, however much bigger the whole pie might be.

initially, minecraft.

Minecraft with multiplayer was launched only 3 weeks after the initial release.

Actually game graphics people are hard to find but no one is going to hire into that role with no experience.

That said, you can leverage your normal tech chops to get a foot in the door. Games do need backend and database people, etc.

We've never had issues finding artists, much of the bulk work is outsourced anyway.

Cloud infrastructure/devops, data analytics, backend code, data storage all need people especially in mobile games.

> We've never had issues finding artists

I think they meant graphics programmers, not artists.

"Graphics programmers" are historically some of the highest paid and hardest to hire programmers in gamedev. But they're probably less of a thing in the age of Unity and Unreal.

Yea, nobody needs a "graphics programmer" unless they're making their own game engine.

In the mobile world very few companies bother. If they do, they're either so old that Unity mobile wasn't very good and already have their own engine or they're doing something niche where Unity doesn't make sense.

Everyone else just uses Unity and Unreal, because trying to launch a new game is hard enough without having to build everything from scratch.

This isn't true at all. You can possibly get away without a graphics programmer on those platforms due to the immense market of graphics techniques available due to... graphics programmers publishing them.

There's a ton of shader development happening. Optimizing the graphics stack takes some specialization.

> due to the immense market of graphics techniques available due to... graphics programmers publishing them

Depends on how one defines “graphics programmer”. Those people are arguably more “tech artists” than “graphics programmers”.

Lots of people can write clever shaders in node graphs. Those folks are usually more “tech artist”. Although the line between “graphics programmer” and “tech artist” is very very very fuzzy.

Not to mention graphics is way harder then web. Let's be honest... Web, which encompasses frontend and backend is significantly easier then graphics programming.

Web programming is incredibly easy in general, and gets even easier every year. Game programming is much harder in my experience. Game Engine programming is not even comparable.

You mean creating a HTML/CSS website is easy in general. Because you can also develop games on the web and create whatever complex kind of application you can think of. Best to be precise when saying these things.

It depends. If you're implementing something like K8s or Spark it's just as hard as a rendering engine, and you only need one K8s and one rendering engine. These days unity, unreal, or Blender can't really be beat.

If you're just working on top of a platform, it's trivial for sure.

No spark and k8s are configuration problems. Hard but you are not creating the tool, just configuring it. It's more of a dev ops /sys admin thing and, though I agree it's very hard, it's a very left field comparison IMO.

With engine creation you are creating the tool. It's programming through and through.

Though I can see how a lot of backend people are taking on the role of "configuring" stuff so it's hard to delineate backend programming away from it completely.

Still even when taking this into account, it is of my opinion that a rendering engine is significantly harder then even K8s or spark.

You are right though that with existing game engines graphics programmers are not as needed anymore.

> Hard but you are not creating the tool, just configuring it. It's more of a dev ops /sys admin thing and, though I agree it's very hard, it's a very left field comparison IMO.

They specifically are talking about creating the tool.

They're talking about the developers working on spark and K8? That's different. But I don't think they are talking about that. They are more likely talking about using these tools.

No they aren’t they said “implementing k8s”

I see, my mistake.

Easier in terms of less math for sure, but there is still plenty of complexity with web. Async, constantly changing environment, and specific device quirks to name a few.

I've done both. Graphics programming is significantly harder. It's both harder to program for and harder to learn. Make no mistake, they are different, but in terms of challenge, Web is trivial compared to it.

You'll get a lot more people in Highschool who know Web simply because of how accessible and easy it is and a lot less people in Highschool who know how to create a physically correct rendering engine or a realtime graphics engine. It is brutally harder.

You get all of that in gamedev as well.

lol

Render programmers are not all that well paid, but they are absurdly hard to hire for.

The highest paid IC's at my last few companies (gamedev) were people that had been there 20 years (He was AI programmer) or the infrastructure people running the gameservers and multiplayer backend.

That might seem weird. But it's hard to hire infra people as they can work basically anywhere and will have as much enjoyment.

Render programmers can work in CAD or Games, there's much less demand, and they tend to want to work in games.

Question is, is it moral to work on that kind of game? You're describing the predatory IAP fests.

Edit: although I suppose it's equivalent to working on "engagement" algorithms in social networking.

Nobody is forced to buy anything in mobile games, it's bad business to do so.

The reason mobile game companies have hordes of data analysts and data scientists is to figure out the optimal point where enough people pay to speed up progress so the game is profitable, but not so much that non-payers stop playing because they feel they can't progress without $$.

Nobody bats an eye if someone spends $80 on the latest console, but everyone is sharpening their pitchforks if they spend a similar amount in mobile game IAPs. Even if the hours played and enjoyment received are perfectly equal.

Most people who buy IAPs exchange money for progress in games, they work 8-10 hours a day so they don't have the time that a Monster Energy Drink fueled teenager has to grind equipment. But they have money they can use to catch up somewhat so they can enjoy their evenings playing the fun bits instead of grinding for that 1% drop like a second job.

>Nobody bats an eye if someone spends $80 on the latest console, but everyone is sharpening their pitchforks if they spend a similar amount in mobile game IAPs. Even if the hours played and enjoyment received are perfectly equal.

I'm all for people spending money in the way they want to... but I think people that buy little bits of virtual powerups and collectables for a mobile game lean closer to "addiction" than "gameplay" or "enjoyment."

A console is a physical device that you can play tons of games on. An IAP, even a collection of $100s is just not the same.

I used to have friends that would spend $100s of dollars on those mobile games. They don't play them at all anymore. So what was the point? They spent $100 to get to the next $100 barrier to spend another $100? And then they quit?

Again, I'm all for people doing whatever they want and think any regulation would cause casualties but I will always say a game designed in such a way where you either have to "grind 8 hours on monster energy" or "buy your way up" is just a predatory and ridiculous "game" right out of the gate.

(meant to type "latest console _game_" there)

People usually have two commodities, time and money.

When you're young, you have virtually endless amounts of time. You can just sit down and grind a game to get better at it and get better equipment.

When you're older, you don't have as much time, but you have money. Then (in my opinion) you should be able to use money to get on par with the people who have spent all the time gaming you spent at work and with your family.

Of course there are games where you just need to "git gud", because it's all your own reflexes and knowledge, in-game items and boosts don't matter that much if at all. If people using money to offset time used bothers you, you can play any of these games.

This is why people selling MMO accounts is a thing. People with money want to enjoy endgame content with friends who have unlimited time.

>When you're older, you don't have as much time, but you have money. Then (in my opinion) you should be able to use money to get on par with the people who have spent all the time gaming you spent at work and with your family.

I suppose. I feel at that point, its more a chore than a game tho, lol.

I don't play those games, exactly for this reason. I feel the reasoning "offsets time" is honestly just a coping/irrational justification. Imagine paying to skip to the good part of the book or movie. Makes 0 sense.

So you have to buy the game and then pay to skip parts of it? That sounds like a system designed to extract extra money from people, not a game...

>This is why people selling MMO accounts is a thing. People with money want to enjoy endgame content with friends who have unlimited time.

This is a bit more justifiable, as long as the game isn't designed around the idea of creating the very situation... I say this as someone that almost failed out of college because of spending too much time on an MMO (talk about no time _or_ money, lol) I'd suggest the friend group just play a different game entirely. MMOs are big commitments... same as any other hobby you could get deep into though. I don't think its good for a group of friends to constantly be throwing time or money at something just so they can stay "at the same level."

*edit* Just want to add to the idea of regulation and casualties. I wouldn't fault a father for spending a little extra to "catch up" with his son so they can play together or something. I can't think of a situation off the top of my head but in that instance, I wouldn't say that's "bad" or "predatory." Everyone's situation is unique. So these comments are more directed to those mobile "games" that involve sinking time or money into them to "advance."

> So you have to buy the game and then pay to skip parts of it?

Games need to be designed for both groups to be successful.

They need to satisfy the people who pre-order the game and are sitting there with controller in hand hopped up on Monster the second the game launches. Then they'll play a 100 hours straight and complain there wasn't enough content because they finished it in 5 days (...of 20 hour/day playing, but they're not sharing that bit on the forums).

Games also need to be fun for the people who don't have a 100 hours to invest in a game, but still want to experience it.

This is why games have Achievements that are nigh-impossible. They're for the first group. Nobody in their right mind would attempt some of them unless they're sitting on a literal mountain of free time and nothing else to do.

Some games (like a bunch of Assassin's Creed games nowadays) allow the second group (money, no time) to buy better equipment with actual money for the single player game. They can still experience the plot of the game and have some challenge, but don't need to spend hours fine tuning their dodge reflexes for that one boss they got stuck on.

> Some games (like a bunch of Assassin's Creed games nowadays) allow the second group (money, no time) to buy better equipment with actual money for the single player game. They can still experience the plot of the game and have some challenge, but don't need to spend hours fine tuning their dodge reflexes for that one boss they got stuck on.

Let's call spade a spade.

This is a convenient rationalization to justify adding elements of unnecessary grind to with intent to get people to spend more money on a game they already paid for.

Games used to (and to a degree, still do) have difficulty selectors and cheat codes to accommodate different types of players. Publishers of AAA games like the ones you mention (Assassin's Creed) monetize this because of greed.

> Games used to (and to a degree, still do) have difficulty selectors

I finished Resident Evil: Village recently. On the "Casual" difficulty. First RE that has that. Every boss went down in a couple shots and I was swimming in the supposedly scarce ammo at the end.

Somehow they forgot to charge me extra for this.

Haha right?

I can sorta understand in some cases but I have no idea why there is a market for "skipping around" in single player.

>Some games (like a bunch of Assassin's Creed games nowadays) allow the second group (money, no time) to buy better equipment with actual money for the single player game.

You have blown my mind, as I hadn't realized this has gotten into single player games like Assassin's Creed.

I generally agree games should be fun for everyone... but at that point is it even a "game" if you're paying money to skip around? What happened to difficulty selection? Or just plain old grinding? That's part of the experience.

There's a game called Cuphead that I really enjoy(ed) but never beat. I couldn't in a million years think "let me just pay 10, 20, or 30 dollars to skip these hard levels." The very idea just seems so antithetical to gaming in general.

What about putting them in Elden Ring, just go right to the last boss if you pay $200-300 (assuming $ scales with progress).

If you're finding yourself with not much time, I recommend rogue likes. I never liked them until I understood that the "playing" was the experience, and grinding to the end was the reward. Now I like them b/c with a Steam Deck, I can play short sessions. Not once have I ever considered paying actual money to "skip" a game. That is just wild to me.

Again, at the end of the day, ppl can spend their money how they like... I just find it a bit ridiculous and that's not even touching on the "gacha" mobile games.

Cuphead, Super Meat Boy and Soulslike games like Elden Ring are firmly in the "git gud" -category.

They don't hand out skips and it's just purely about the player's skill.

That's why I don't play any of them, I don't have the free time to spend that's required to get good enough.

I currently play mostly story-driven single player games like the new Tomb Raider games, Quantum Break, Control etc. on my Steam Deck.

> I don't have the free time to spend that's required to get good enough.

Want to buy my save just in front of the last Elden Ring boss fight? :)

Considering your earlier posts, it's the logical thing to do.

Actually the game doesn't stop after you beat the last boss, I could do that and sell you a save right after.

> When you're older, you don't have as much time, but you have money. Then (in my opinion) you should be able to use money to get on par with the people who have spent all the time gaming you spent at work and with your family.

the problem with this is that you assume "all the time gaming" is a bad and nonfun thing, at that point you're essentially paying the developer to not play the game. Sure those situations do exist, but they are design flaws* that should be fixed. Accepting a quick way to get thru them gives the developers incentive to not fix them but to actually make them more prevalent, and this has been happening in games more and more, just look at overwatch 2 and the hero unlocking speed as an example on a huge AAA game.

>This is why people selling MMO accounts is a thing. People with money want to enjoy endgame content with friends who have unlimited time.

This could be argued as a reason for selling currency, but for selling accounts I harshly disagree. The vast majority of people that buy accounts in MMOs do it to get "stolen valor" such as buying gladiator/rank 1 accounts in WoW or exclusive items/mounts/etc. The endgame activities most take part in takes orders of magnitude more time to progress in than the lead up to it nowadays, makes no sense to buy an account to skip it if you're genuinely time constrained. I could see that being a relevant case back in Ultima Online days for setting up a new PK character but that's not really a thing today in any big capacity.

*: some exceptions exist to make balancing passes I suppose, just to not speak in absolutes

You're somehow assuming all computer games involve grinding for some 1% drop.

You're basically describing old single player games that did rare drops to pad gameplay time, then the first MMOs where grinding was there to keep you paying for the subscription, and now the IAP fests where the grinding is there to make you pay something.

However, there still are games that do not involve grinding out there. Even some AAAs. You do not have to be predatory to enjoy some success making video games.

Edit: I missed that: "Even if the hours played and enjoyment received are perfectly equal."

They aren't equal. Do you call grinding for rare drops equal to enjoying a story or developing and applying some skill?

And how is hours played a metric of enjoyment? If you count hours played, the game i spent the most time in ever is Minecraft. Paid all of $20 for it.

I'm describing games like World of Warcraft, where I spent WAY too much time grinding for some elusive drop to get to the next quest in the chain. It's amazing how many wolves you can kill in Azeroth without finding a single paw :D

I would've gladly paid money to skip that bit and get to the fun parts even then and I was a mostly broke student at the time.

Now I play mostly story-driven single player games that don't rely on fetch quests to pad the runtime. Or games that really don't have an ending like Factorio, Dwarf Fortress and the like.

> get to the fun parts

What fun parts? You just grind more, perhaps in a group. Or a larger group. I know, I could probably still lead a Kharazan (of the Burning Crusade flavor) raid from memory. The only fun part of WoW is the group interaction.

The sad thing is when people have only played WoW and its successors and start thinking this is the only way to do a game...

Factorio has an ending... kind of.

No, it is not. If a reader has a gamedev position outside these predatory markets, hit me up.

I thought game dev salaries were bad, but people still applied because of the thrill.

In a team of 20 making a mobile game, you have 17-18 people who don't do backend and 2 people who do.

Those two are the ones that are invaluable, there's a literal queue out the door to fill the sexier positions of game designer/producer/2d/3d artist/unity developer and whatever.

You can just pretty much plug in a new Unity developer to the team, since the skill set isn't that hard to learn.

...but finding someone who can do 1M+ DAU game backend that's reliable and cost-efficient isn't something you just hire off the street.

> but finding someone who can do 1M+ DAU game backend that's reliable and cost-efficient isn't something you just hire off the street.

So, how do you get that particular skill? I've done backend development, but nothing with that kind of scale.

Basically you're paying for someone who has seen shit fail 42 different ways and has a pretty good feeling/knowledge what works in which case based on their experience. And they also know which solutions are overkill.

Simpler is always better, don't do stuff for your CV, build it so that you don't need to wake up at 3am on Christmas to fix some obscure deadlock issue.

Log stuff so that you can tell most of the time what went wrong just by looking at the logs and graphs.

But also be mindful of costs, logging everything to Cloudwatch can get surprisingly expensive.

Benchmarking is important, you can use the logs to see where you're actually spending time and optimise it if possible.

Sometimes the illusion of real-time is better than being real time. Everything that can be cached, should be cached. Utilise the TTL in Redis to make stuff simpler.

Ok, so how do I convince a prospective employer that I'm capable of these things? Well, except for seeing shit fail 42 different ways. We have a bit of a chicken and egg problem there. How can I have seen shit fail 42 different ways if I need to see shit fail 42 different ways to get the job?

Not all games are 1M DAU, you can leverage knowledge of smaller games first.

Only the biggest games with huge marketing budgets go from zero to 1M DAU in a day, so there's always the possibility of slowly improving the system while the player count grows.

You can still design the systems and try to test them. Build programs and throw load tests at them.

Scaling software is more about careful choices than some ultra deep understanding (though really understanding your data source is very important).

The difficulty really depends on the game. Some of these things are glorified webapps and 1M+ DAU is pretty trivial on the cheap. But if you have more real-time multiplayer interaction, it would take a lot more work.

Those would be the roles that “everyone wants” as alluded to.

Any good resources to get started

Try to build your own prototype.

I found a fairly low-friction way to add a custom online services subsystem into UE4/5: https://github.com/RyroNZ/UE4MasterServer

You can replace the python with anything that lives across JSON. I had no issues getting this to compile & load in UE5.

Once you have this in place, you could grab one of their arena shooter MP prototypes and focus almost entirely on the sessions, servers, matchmaking, accounts, logging, fleet management, etc.

Reply on news.ycombinator.com