What is Stadia?
Last week Google announced Stadia, a cloud gaming service. The idea to play games in the cloud and stream the video output over the internet isn't particular new, OnLive, was announced ten years ago, also at the Game Developers Conference (GDC), Gaikai was also shown at GDC, but one year later. Since both companies aren't giant names nowadays you might think that cloud gaming was tried but never really caught on. Well, this is kind of true, Sony bought both companies (Gaikai for the technology and OnLive for their patents) and launched PlayStation Now, which allowes to stream games. I mean it is there, but to me it seems like it's more of an afterthought for Sony, not something they will centre their gaming business around. Just like 10 years ago, there seems to be a hype around cloud gaming right now: Microsoft develops Xcloud, Verizon and Amazon are apparently also working on services like this and then there's NVidia Shield, which also goes into this direction.
Testers of services like these seem to agree that controlling the games works surprisingly well, but the image (aka graphics) quality was occasionally suffering. Since everybody is worrying about the games lagging, it makes sense that the engineers focused most of their work into fixing this, at the cost of image quality. Needless to say that streaming something interactive like games is much harder than streaming something linear like films or music. But still, the streaming technology behind those game streaming services is pretty much the same as for streaming linear media: when the connection is getting bad, the software reduces the quality to save bandwidth. In the case of music only very few people will be able to tell the difference, for video it is more obvious when the quality drops. Technically speaking streaming a game is nothing more than streaming a video, except that the player decides what frame he is going to see next via the controller input.
Google claims to have solved the problems with latency. The requirements for Stadia aren't particularly low: 25 megabits per second for 1080p with 60 fps, 30 megabits for 4K (Sony recommends a steady connection between 5 and 12 megabits per second, but it only supports up to 720p). Reading what testers wrote about similar services, this actually sounds reasonable. So, to whistle with the choir: "if anyone can solve the issues regarding latency, then it is Google". Google has the infrastructure and know-how to take on challenges like this. One thing that also came to my mind is, that Google could also use machine learning to predict where the player is going next and send that frame (or at least a compressed version), before the player is actually going there. This idea is not particular new, it is called prefetching and processors are using it since decades, the only new part would be to predict the next input of the player, but using machine learning this is not impossible. Another way to save data would be to use something similar to Foveated rendering, again combined with at machine learning to predict the regions which are most likely ignored by the player and encode them in lower quality. Those two are just speculation from my side about how Google could tackle the latency problem.
Even though the technical challenges are tough, there are a lot of advantages. Unfortunately most of them are for the developers and the owner of the platform (obviously, otherwise Google wouldn't be interested) and not so much for the players. By taking away the game from the machines of the players, the game developers don't have to face with problems like piracy, cheaters or different hardware configurations. The advantages for the gamers are that they are now don't have to worry about their hardware being powerful enough to run the newest games, buying new hardware for games is a thing of the past. So is paying €50 - €60 for a new game, just a fixed price per month will do, also having to download a game is a thing of the past, the idea of Google is that you see something you're interested in on YouTube and a couple of seconds later you are able to play it. Gaming in the cloud also allows for new possibilities we haven't even thought of yet, like style transfer, as they showed off during the introduction.
Pros and cons
All those things sound nice and dandy, but don't forget that you don't own any of the games you play. From the developers' perspective, this is of course a good thing, as I already mentioned. Games are no longer just games-as-a-service, gaming itself is becoming a service, one where the developers (and Google) have full control over. Conserving games for later generations is no longer possible, once the servers are down this game is gone forever. Another thing that is troublesome is, that Google has full control about which games are on the platform, in case Stadia is becoming the de-facto standard for gaming it could be that it is almost like a game doesn't exist if it is not on Stadia. Although I have to say that if there is going to be an oligopoly it is also not very good for the consumers: there are different cloud gaming services and each one of them has different games, so if you want to play multiple titles, you have to have multiple accounts (that's something we are seeing now with Netflix and its many competitors, e.g. Disney's streaming service). In the gaming market there used to be a "friendly monopoly", with Steam being by far the biggest player, and smaller services like GoG or Itch.io sharing the remains or serving some niche. Big names like EA, Blizzard or Ubisoft tried to launch their own stores, but they usually remained exclusive to their own titles, fragmenting the market a bit, but without ever being a serious competition to Steam (I guess most players thought of them more as a nuisance than anything else). Last year this changed, when Epic started their own store and spending quite some of their Fortnite money to buy exclusive titles and keep them away from Steam. So yeah, we now have an oligopoly, but honestly it is much worse the previous monopoly was (which is something odd to say). I could go on about the Epic Store, but I guess, I leave that for another article. My point was that the whole games-as-a-service trend spawned oligopolies and from a consumer's side this is not good, game streaming services could have the same problem. Of course I'm not alone with these worries, David Brevik (the "inventor" of Diablo), tweeted about the impact these kind of services will have on game design. He is especially concerned about the fact that game developers get paid per played minute of their games and how this will affect game design, imagine all the worst things from free-to-play games on steroids.
The development of cloud gaming services actually follows a rather logical linage: Steam was launched in 2003, at first it only attracted a small audience, but 15 years ago, when Half Life 2 was released and Steam exclusive (meaning that you needed to have a Steam account and an internet connection to play it), Steam got more and more widespread (a bit like a virus if you ask me). I guess it is difficult to imagine today, but the internet wasn't available everywhere back in 2004 (I know, because I didn't have internet at home, I wanted to play Half Life 2 so badly, but couldn't) and what followed was a rather big community outrage. People were complaining using similar arguments to mine above. The outrage didn't help (I basically boycotted Steam until 2010 btw.), Steam won in the end, but I have to say that Steam also improved a great deal in the last 15 years and Valve did a great job toning down some of the criticism, by just being there for the community. Difficult to say, if Google is doing the same since it is not a small company that grows with its community, but one of the biggest corporations in the world.
I mentioned that Google seems to have handled the technical issues surrounding latency. Even if they did, it still doesn't help you when you live in a rural area where there is no internet connection with 25 megabits per second. A similar point was also used against Steam in its early days, but for Stadia it is even stronger: even with a slow internet connection it is possible to play games, once they are downloaded (again, I know what I'm talking about: downloading Warframe or Dawn of War III, which are both around 30GB, almost takes a whole night when on the country side), but Stadia needs a good internet connection, if there is none available in your area then well, tough luck. You might respond with: "if your internet connection is slow, you anyway can't play a multiplayer game like Warframe". Well, no: when I wrote my review of Warframe I was actually on the country side and the internet connection didn't keep me from playing quite a bit before I wrote the review. Another thing to keep in mind about latency is, that even if you have a good internet connection, you also need to have a good WiFi connection, otherwise your gaming experience will suffer. Google announced that a special controller will be directly connected to the WiFi to minimise latency. That sounds good, but if your WiFi is far away from where you want to play, you will run into bandwidth issues, it is the same when you are sharing your WiFi with someone else, and he decides to download a torrent or watch something on Netflix. Maybe 5G will be the remedy for this, but it still takes a while until it will take off (even though the first tests started this week in Austria, unfortunately my village is not part of it, but a town nearby is).
Google advertises Stadia with the promise that game developers can really push their games to their technological limits and allow gamers to play them exactly in the way the developers envisioned it. But now I want to raise the question: do we really want to play the game in the way the developers envisioned it? In the early days of gaming, computer games were created by nerds for other nerds and therefore it is not surprising that those nerds started to play around with certain aspects of games they were playing. They changed some things and added new things, modding was born. Modding is and was an integral part of PC gaming, mods like Counter Strike or Defense of the Ancient (better knows as DotA) are just two examples of mods that defined or even invented genres that spawned some the most popular games today. Without the ability of fans to modify the games they love, none of these would have been possible. In some cases a mod even put a game on the map (like DayZ did for ARMA 2). I'm not sure how much the modding scene helped Minecraft, but I guess that for sure it didn't harm its success. Being creative with games is highly limited if you are not in control of your game (Stadia allows you to directly upload stuff onto YouTube, but seriously, that's as creative as painting something colour by numbers).
After pointing out the negative aspects of Stadia (or cloud gaming in general) let's talk about what will decide about whether it is going to be a success or not: the price. Google didn't announce a price yet, PlayStation Now for example right now costs €14.99 per month or €99.99 per year in Austria. That looks like a pretty good deal, considering that a new AAA game usually costs €60, but there are two drawbacks: if you take a look at the list of games available on PlayStation Now, you'll see that there are only a few AAA games and that those games are usually older. The second drawback is, that prices are not static and the operator can increase them, especially if the service has already gained a lot of traction and doesn't necessarily rely on a low price, that's something that Netflix likes to do. Considering that interested gamers will save a lot on hardware and maybe also on games (if Google gets big names to put their games up there or starts developing their own games, which they do) then Stadia is definitively the cheaper option. More convenience and a smaller price are usually a recipe for success, even if the cost is giving up the freedom to do whatever you want with your games. So from that point of view, if the technology delivers what it provides, it looks like Stadia has the potential to be a success for Google. Although I'm a bit wondering who the intended audience is: from what I saw it is more targeted towards core gamer, but they are usually less willing to give up ownership of their games, as well as more sensitive towards technical issues like a higher latency, which even though Google tried the best they could to do to get a hang of, will most likely still be noticeable.
My thoughts on Stadia
After stating positive and negative aspects of Stadia, now I'm going to write about my opinion, some personal wishes and some other aspects. As you probably noticed in the previous part, I'm obviously not the biggest fan of the idea of streaming games. This is partly because of my experiences with the launch of Half Life 2 and being a free software advocate for years, which means that taking away almost all rights from the user is a huge no-go for me. But even with all the negative aspects, Google is also doing quite some things right and let's start with the biggest one (for me): Google uses Linux as operating system where the games are run, Debian to be more precise. That's huge news for Linux gamers like me! Seriously, for the longest time Linux gaming was a niche for nerds who were (and still) are very grateful for every game released on their system. AAA titles rarely came to Linux, with a few notable exceptions like Borderlands 2, Dawn of War III or Shadow of Mordor. Since Unreal Engine and Unity added support for Linux, the number of Linux games drastically rose, especially more and more indie developers are happy to port their games to Linux, in order to increase their potential player base (I had to chance to listen to a talk of the creator of Spinnortality and thanked him for porting the game to Linux, and he stated that his motivation was to increase the potential player base, even though the percentage of Linux users on Steam is very low).
In general, I have to say a big thank you to Valve for making Steam available on Linux. Whatever Engadget says, Steam isn't Linux gaming's life support, it is rather responsible for the huge grow that Linux gaming experienced in the last couple of years. Just last year Valve released Proton and gave yet another big boost to Linux gaming. Valve was doing great things for Linux gaming, but also Google's decision to use Debian as platform of choice will have a potentially huge impact on Linux gaming as whole. The biggest argument against games being ported to Linux was that it is a lot of effort and not worth for the little player base, but now that any company that wishes to release a game on Stadia has to have a Linux build, this argument isn't valid any more. Linux is now more than just an afterthought or something by nerds for nerds, it is crucial to have Linux support if you want to have your game on Google's cloud service. And that is great! Since the burden of porting a game to Linux is already handled during the development of the game, the biggest drawback from a game developer's perspective is gone. Not also releasing an already ported game for Linux would be a rather silly decision from a developer, because even though the market is small, there is a market and the developer would also have a product for that market, so he would pass on an opportunity to make money there. Increased supply of games would also further drive adoption of Linux as a gaming system, meaning that this market is likely to grow (I read many times that the only reason why people still stick to Windows is because of games, over 10 years ago when I switched to Linux also the only reason why I kept using Windows was for gaming).
You can now argue, that games won't be released outside of Stadia, because Stadia is the only platform where games are going to be released. But I would disagree on this one, because Spotify doesn't keep CDs from being released, neither does Netflix keep films from being released on Blu-ray or DVD. I think that game developers will still release their games for platforms other than Stadia (except for Google's in-house studio of course).
After stating the obvious, let's talk about some pipe-dream scenarios that came across my mind: earlier I mentioned that cloud gaming will make modding a thing of the past, but honestly this does not necessarily have to be the case. Since all the computations are done in the cloud, what does keep developers from making their games open source? Sure, it is very unlikely that the big old developers like Bethesda will jump onto this, but Google's in-house team might will. TensorFlow is a good example, where Google open sourced a very powerful machine learning library. You might think why they released a tool which would give them an competitive advantage to the public? Well, in order to make good machine learning models you don't just need powerful algorithms, but also a lot of computational power and data, two things that Google has, but are very difficult to get for the general public. Open sourcing the algorithms has the advantage that you will create a community of developers around it, which at one point also will need a lot of computational power. Where will they go to when this is the case? Google cloud of course.
Let's imagine that Google's in-house team is developing a multiplayer AAA title and release its source code. The vast majority of the player base will be on Stadia and it is very safe to assume, that most of the multiplayer part will happen there. It is also safe to assume, that Google will use some special features of Stadia, that only they can provide with the demanded quality and on the demanded scale. Therefore, opening up the source code will not be too big of a thread to the business model, because a multiplayer game without players is just boring. There are many multi player games out there, which are more or less dead, because nobody plays them. The bright side is, that interested developers can take a look at Google's code and learn from it, which will further drive adoption of Stadia among developers and be beneficial for Google in the long run. The big but here is, that if the game is available for free, why would someone still pay for Stadia, when he can just download the game and play it on a community server? Well: 1.) playing on Stadia will be the most continent way to play, so most of the people will do that 2.) a multiplayer game needs a big player base and 3.) Google will utilise other special features to keep players on their platform. Special features don't necessarily have to be only of technical nature, just think of simple progress mechanisms. Since progression system are getting more and more important (especially in multiplayer games), I think that players are much more committed to stay the platform where they started playing 4.) Google needs developers to use their platform, therefore it would make sense to release the code, so that other developers can see how the whole works.
I also talked about the death of modding earlier, but if Google releases the code of one of their games to the public, they could also facilitate a whole new dimension of modding. Google of course needs to support this, I admit, that modding just got a lot more difficult, but it is not impossible, but depending on the goodwill of Google, not only because of them releasing the source, but also of providing infrastructure to run the mods in the cloud, similar to what Valve is doing with Steam Workshop.
It looks like gaming is in a very interesting bifurcation right now: on the one hand side there are those cloud gaming services like Stadia and on the other hand side there are blockchain based games, which are trying to give users back the control over their games in many different ways. It's interesting to see, which one will win, Google Stadia got a lot of buzz from the "mainstream" gaming media, while blockchain based games are still a very small niche, even within gaming. In the short term I would say that things look good for cloud based games, in the long term blockchain probably has the bigger potential, meaning that when people realise that they can earn money from selling virtual goods they earn in-game. But actually both can co-exist and nurture each other: right now cheating is a nuisance, but not a problem that has some profound effects. When real money is in play (e.g. for items), cheating is a much bigger issue, because cheaters have a much higher incentive and could ruin the whole in-game economy an therefore the whole game. Cloud gaming would be a possible solution for this problem, because if everything is executed in the cloud, then the attack vector for cheaters and hackers is rather small.
Google Stadia is an interesting project and I will keep an eye on it, even though it has some worrying implications and contradicts with my free software ideals.