Sunday, October 20, 2013

The hypothetical Apple non-console

I don’t know if Apple is going to introduce a gaming-capable Apple TV this week–in fact, I suspect probably not–but it makes sense to me that they could, if they wanted to.

Here’s what I imagine it might be like in terms of specs (look at me, I’m an ‘analyst’!):

Current Apple TV Hypothetical 2013 Apple TV 'Good’ Hypothetical 2013 Apple TV 'Better’
  • Single-core A5 SOC
  • 512 MB RAM
  • 4 GB flash storage
  • $99
  • Dual-core A7 SOC
  • 1 GB RAM
  • 16 GB flash storage
  • $99
  • Dual-core A7 SOC
  • 1 GB RAM
  • 32 GB flash storage
  • Game controller
  • $149

Here’s why I think so, by way of a point-by-point rebuttal of Andy Reitz’s arguments (via Gus Mueller).

In order to have graphics that compete with the upcoming PS4 and Xbox One, these games will need a strong GPU (for rendering graphics at 1080p), as well as tons of RAM (as Gus has speculated above). But these AAA games are also massive, requiring 8 - 30GB of storage per title.

Apple has never, in their whole existence, shown any interest in being part of what has traditionally been the gaming market, and I don’t expect that to change, notwithstanding their (apparently accidental) success with games on iOS. I don’t think Apple would ever try to compete directly with the PS4 or Xbone.

Instead, I think Apple would target their existing customers, who already own at least one Apple device, but are probably not very ‘serious’ about gaming, and may not even have any console at all currently hooked up to their TV (except perhaps an ageing first-generation Nintendo Wii). It’s arguably a much bigger market, and one that Apple already knows how to sell to. And, make no mistake, the emphasis for Apple TV will always be on movies and TV, with games as a distant and optional secondary feature. A ‘game console’ in the traditional sense, this is not.

If Apple were to produce such a box, paired with a game controller, I think that people would want to play AAA games, such as “Call of Duty”, “NBA2K14”, “Need For Speed”, etc.

Based on the above–no, they wouldn’t. The hypothetical market I’m referring to would want to play the same games they already enjoy on their iPhones and iPads, except now on their TV. Games like Infinity Blade 3, which by most accounts looks pretty good on existing iPhone/iPad hardware, despite the inherent constraints: apps can’t take up more than 2 GB of storage space, and in the best case (iPhone 5S), are limited to 1 GB of RAM and an A7 chip at resolutions similar to an HDTV1.

32 GB for the ‘gamer’ model would be enough to fit, at minimum, 15 or so games (probably more). And for people who aren’t interested in games (at first), they can save a little money on the cheaper model, and add a controller later if they change their minds.

The first is that Apple TV doesn’t run iOS 7, it appear to be running a custom fork of iOS 6.1. Of course Apple could upgrade this fork to include the relevant gaming bits of iOS 7, but this would be more work for an already taxed engineering team.

I don’t buy this. We have no idea what Apple does or doesn’t have working in their labs. It could be anything, or nothing. It doesn’t seem helpful to me to try to guess from the outside about how Apple allocates engineering resources.

In addition, according to the “Game Controller Programming Guide”, the controllers are optional … Now, I’m sure Apple could change their just-release specification for a hypothetical gaming-capable Apple TV. But how would game designers work with these constraints? Requiring games to be designed for both touch and gamepad controls, would be a huge limitation for games that run on the Apple TV.

This is a non-issue. In the same way that developers can currently target an iOS app to iPhone, iPad, or both (‘universal’)–Apple TV would simply be added as an additional target. If you want to target touch devices, then you must include touch controls. If you want to target Apple TV only, then you don’t. Simple.

Obviously, the current developer docs are based on currently announced products, which all have touch screens. And obviously, Apple would revise them if they decided to release a new product. If they were planning to release a new product, they certainly wouldn’t be so shortsighted as to tip their hand about it in developer docs six months beforehand.


My hypothesising is analogous to Apple’s choice of CPUs in iOS devices: Instead of trying to take powerful Intel x86 CPUs and scale them down to be more battery-efficient, they chose to take simpler but already highly battery-efficient ARM CPUs, and gradually grow them to become more powerful. In the same way, they could release an Apple TV that is not very powerful at first but ‘good enough’ for now, and comparable with the graphical output of current iOS devices; and then slowly grow it to become more powerful over the years, as they have always done with all of their other products.

This leads me to another point about traditional consoles: They have a very slow release cycle, so they often start out with incredibly overpowered specs in the beginning, because they are expected to last for about six years. Do you expect to be able to buy and run a brand new iOS game today on your original circa 2007 iPhone? No. But the PlayStation 3 debuted six months before that same iPhone, and new games for it are still coming out today (and planned for next year, at least!).

When has Apple ever entered a new market by trying to do things the same way that all of the existing incumbents do? If Apple made a games-capable Apple TV, they would do it in their own way, like they always do.


  1. An iPhone 5/5S screen is 1136x640, which is just under 720p (1280x720). An iPad ¾ screen is 2048x1536, which is considerably more pixels than 1080p (1920x1080). For that reason, I’m not sure it would be necessary to use a GPU from the iPad to drive an HDTV; perhaps one of the less-powerful ones from the iPhone would do fine. I don’t really know enough about GPUs to speculate with much confidence, though. And let’s not even get into 4K, for now–that’s a long way off. ↩︎