The Gold is in the stream(ing)! Yet another tech goldrush.

"Company X is considering creating a game streaming business."

This headline, or similar, has popped up so often over the past six months that you'd be forgiven for losing track of which companies are considering it now. Or even caring which ones for that matter. But the question to ask here is why. Why are so many companies suddenly "considering a game streaming service?"

An answer is of course that they're chasing yet another tech goldrush, and just like the goldrushes of the past many will fail miserably having wasted a ton of time and money in the process. But to understand why we'll have to start with what game streaming is, and the tech behind it.

We'll start with the "game" portion. Video games are an app, in modern parlance. While in their own, incredibly popular category of app (they're far more downloaded as a category than any other) they're just another executable with code that runs on a computer. Most everyone reading this knows what they are and almost certainly has used these apps.

So the interesting part of the equation becomes "Streaming", that second word. "Streaming" here means inputs go into the computer device the end user is on, over a network, to another computer, which calculates the output, who's representation is then "streamed" via audio and video, back to the end users computer. In a normal app that whole "network" thing is removed. Meaning a normal app has a lot less steps, one less computer, and overall a lot less cost if you just remove the whole "streaming" part.

So why then is this streaming part, the one that adds complexity and cost, seen as the next tech goldrush? The answer lies in the idea that the computer device the end user has is either not good enough to do the "calculates the output" part of the equation, or at least isn't fast enough to do so that temporarily renting a faster computer to end user overcomes the cost of all those extra steps.

But is this true? Is the device in the hands of the end user really "not fast enough"? The best selling category of computer devices is the "smart"phone. These devices have been steadily increasing in their ability to "calculate the output" for quite a while now.

Let's take a look at the bestselling video games of all time and see how many of the 50 a smartphone can run. For the sake of argument we'll split this into: Games pretty much all smartphones could run for the sake of computation power: 38/50. Games modern, high end smartphones could run: 49 out of 50. Yes, a Galaxy S10 should, going by GFXBench and Geekbench results, be able to run GTAV(link to minimum GTAV GPU benchmarks). of this list only 1 game, Red Dead Redemption 2, is necessarily blocked by hardware.

Alright, so we've established that a modern, high end smartphone can run most every game on the bestselling games of all time list. Even if you extended this to free to play, they're mostly already on smartphones or runnable on smartphones, even Apex Legends theoretically could run on a Galaxy S10+. Thus streaming does not make sense for the vast majority of gamers, if you've enough money to pay for a game only streaming service you've certainly got enough spare money to buy a Galaxy S10e before that, and get the compute power along with a host of other features such as a very solid camera that game streaming services don't provide.

So to find out what audience game streaming could target we'll have to narrow down. What if we chose a more "hardcore" gaming market? The Steam Bestsellers List for 2018 should serve well. And the result is much better for game streaming! Only 30 of the bestselling 2018 games could be playable, given reasonable dev time, on a modern high end smartphone. It looks like we've found our audience for game streaming.

So now that we've found an audience, what competition does game streaming have? Well, consoles for one. A PS4 and Xbox One could play all the games listed, even if they're not supported, in a computational "output" sense. But looking forward, in the future they may not be able to, hardcore games tend to ramp up system requirements over time. Which is when a Playstation 5 might come to hand, as just one example of competition. In order to appeal to consumers over a hypothetical Playstation 5 that looks much like a Playstation 4, what advantages (and disadvantages) would a game streaming service have?

We'll start with how many game both could play on hypothetical bestsellers list of 2021: The PS5 would play all of them, assumedly. Game Streaming service X would play: 99/100 of them, assuming VR becomes no more popular than it is now. Game streaming services can't play VR titles, the input to output lag would be far in excess of the requirements to not have players hurl. But 99% of games isn't bad, so what other points of comparison would there be between the two?

We'll start with mobility: Game streaming can be played from anywhere* well anywhere with a very good internet connection, that has servers in the area. So offline is out, some countries are out, airplanes are out, very rural areas are a no go. But otherwise that's fairly mobile. Versus the PS5? Well the PS4 already has a Remote Play app, one that works decently as long as you have a good internet connection and are close enough to home. Considering most people spend most of their time close enough to home for this to work, and you need a good internet connection for both services, it does not appear that Game Streaming has a huge advantage here for most people.

Alright, what about price then? Well that's the big question: Google's Stadia has been announced, sans a price as of this writing. There's the "Shadow Ghost" streaming service, and at $35 a month... would take only a year to equal the average successful launch price of a new console, which you get to keep until it breaks, a point which most console manufacturers have managed to make quite a few years away.

Now theoretically "cloud computing" is supposed to cost less than a dedicated computer. If you have elastic demand for compute power, and have a dedicated server then anytime your not using that server you are losing theoretical value! Thus cloud services capture that value and sell it back to the customer, you only use what you need! That's much of the point of cloud computing, well it's half the point. The users of today's cloud compute tech also pay cloud computing for all associated IT services, and being close to any point around the globe, and heavy bandwidth pipes, all costs that can be amortized by sharing.

These second points are all things that, by necessity, a game streaming customer is paying for as well. Yes, the system will do updates for you, have games installed for you, and etc. These are all niceties, niceties that don't just come for free. While the cost can be amortized by sharing, there's still a cost that's going to go against any lowering of cost "elastic demand" materializes. Game streaming customers also pay for being close to any point around the globe, a service 90%+ of them won't need 99% of the time. They'll also be paying air conditioning costs, server rooms get hot, so do consoles but that's almost never a problem when there's just one. GS customers also pay for heavy bandwidth usage, which isn't a cost associated with a dedicated computer if you have it in front of you.

So what about this bandwidth thing? Well by necessity the GS customer pays in their monthly fee twice, for the data to come from the servers and then to them. We'll ignore the former and concentrate on the latter. Google's Stadia's connection stats put it at a lot of bandwidth, bandwidth not all home users, not to mention mobile users, can't afford in the US (or Canada, or similarly dispersed population countries).

Then there's the idea of reselling, lots of people resell their consoles. A Playstation 4 goes for $150, or more, on Ebay today. If you bought one new 5 years ago you'd be paying the equivalent of less than $7 a month already, if you didn't sell it. If you did that drops the price to just over $4 a month, which if we assume an efficient market is the market price per month of owning a PS4. Let's be clear, neither I nor any reasonable person should have to justify why Google Stadia or other game streaming service will be more than $4 a month.

So is Game Streaming cheaper than a console? For consistent users, No. And that's a big No to swallow.

There's a lot more that could be gone over. "But smartphones can't actually play all those games!" Is both an input and dev support problem. The former must be solved by any game streaming service as equally as it is solved by inputs to an Android device, and so is moot. And looking forward, with Grand Theft Auto 3-San Andreas, and now games like Ark, Fortnite, and PUBG on Android it seems dev support is slowly ramping up as well.

Then there's the Nintendo Switch, a mobile gaming console. If a mobile gaming console were released today at the cost of say, $425 it could then most likely play both all 2018 Steam Bestsellers, and hypothetical 2021 Steam Bestsellers as well. And most likely look pretty good, and in 4k if you must, as well. Not only could this theoretical console cost less than a streaming service, it could be played offline and in rural areas and on a plane and etc.

Then there's the assumption that a Game Streaming service offers an exactly equivalent experience to close by computer. It does not. Input to output lag is increased by the travel time over a network, degrading the experience in one way. As John Carmack, founder of Id Software put it, this lag is now similar to the processing time many TV's add for lag anyway. Except, of course game streaming does not make TV lag disappear, so this lag will just be added to many peoples experiences. And in motion video compression is often apparent when streaming games, not so if a console is hooked directly to the screen. So the experience is not directly equivalent.

Then there's the idea of a "Netflix like subscription!" to games that could be played instantly. Except in monetary terms those already exist in numeration. None of them have been a rousing success, because games are not what is known as a substitute good. When streaming music, music is often decently substitutable. Yes it's annoying Tool isn't on there, but plenty else is, few complain, for one example.

But as an example of a less substitutable good having Friends on Netflix appears to be worth $100 million a year. Hard to substitute Friends, it would appear; and games are worse. According to many players there's no substitute for that game they play. Harvest Moon isn't Stardew Valley no matter the similarities, just like The Surge is no Sekiro. People will pay $60 happily, and would probably pay a hell of a lot more (and do judging by games like GTAV) for their handful to one game they play. Not having to install a game to play your Netflix of Games service! Is a nicety, but will not change the vast majority of peoples habits.

So after all this: What is game streaming for? Well this is all logical supposition, instead of benefit of hindsight. But here's two possibilities: Flexibility and demos. The first is: If you want to play that ONE game that requires high compute power, paying for a month a year, then game streaming does indeed cost less than a console, even at Ghost Box level costs. The second thing is that demoing a game, while trying to find a new one that you like, would indeed be quicker and easier without download and install times, on average.

Neither of these, obviously, pay a lot. Cloud Computing relies on steady, high end demand to make money, not on sporadic sharp spikes of customers. Demoing a game costs someone money, but are customers really that into trying games quickly and easily to pay for it themselves? Probably not to the tune of a hell of a lot of money.

That's the sort of picture that this goldrush is hoping to get their theoretical gold out of. Somewhere in there might be gold, perhaps? But there's a basic business tenet those in areas that have had historical gold (or silver) rushes have learned as a kind of common knowledge history/business lesson. The people that most reliably win in a gold rush are not the gold diggers, but those that sell the picks and shovels to the gold diggers.

And ultimately, perhaps most importantly, those pick and shovel sellers are Game Developers and Cloud Compute Sellers (and those they buy from). More competition for selling their games can only be good for game developers, already competition has lowered the cut Steam takes from highly successful games, by offering even lower cuts from the likes of the Epic Games Store.

For Cloud Compute there's the usual 3 sellers that stand to win, potentially. Microsoft, Amazon, and Google. Should they wish all 3 have the supply chains, server rooms, IT staff, tax cuts, and etc. to undersell any attempt by any other company to start selling cloud computing themselves; no matter that the service may be somewhat different from those they already provide.

Further it appears at least two of the three will happily pick up their own shovels and try for gold themselves. Stadia is announced, and Microsoft's rumored next gen consoles include "Streaming" heavily in one of two SKUs. What's more all 3 cloud providers already have other tools required too, such as familiar user bases to sell to, working relationships with game developers, and etc. Even if these companies withdraw from Game Streaming, an eventuality Google has already prepared for by test running whether it can transfer save games from their cloud to a developer cloud, they could still happily sell enough shovels to other companies to dig their own financial graves with.

So who is Game Streaming for? Well consumers. They'll win on average no matter what, more competition to provide them with hardware and software outlets should be an easy win for them. Secondarily for game developers, without any substitutable good just emerging from Game Streaming, well they stand to be able to sell more games, if not at least then at most, either way that's a win. And then perhaps the companies that sell picks and shovels to cloud providers. Those servers and "blades" and whatever need to be paid for, and no doubt AMD, Intel, and Nvidia will be happy to provide.

It's the Cloud Providers that are taking a gamble to see if there's gold, well two of them so far. It's also anyone else putting money into game streaming. But if you're not a cloud provider already? Well there's always plenty of losers in a goldrush. Maybe, as the old adage goes, selling the picks and shovels is a better idea.