I just released SpaceCat HD for Android TV. What it should have been a very simple and straight forward process turned out to give me quite a headache.

If you have a game that is gamepad-ready. Chances are that you are supporting MOGA controllers, simply because they are one of the best. If that’s the case you may also have this problem.

The not-so-obvious problem

If you want to publish for Android TV, you essentially have to do a few things in your manifest (see the full article on developer.android.com):

• Declare a TV Activity using an intent filter of the category CATEGORY_LEANBACK_LAUNCHER.
• Add a banner to the application (android:banner=”@drawable/banner”).
• Declare it as a game (android:isGame=”true”).
• Specify that it knows how to handle controllers.
• Specify that touchscreen is not required.

All these are very straight forward, but some of them require you to compile with targetSDK=21 (a.k.a. Lollipop) because is when they were introduced.

Another interesting change on Lollipop is the requirement of all service intents to be explicit. Any non-explicit one will throw an exception on runtime. This requirement is triggered when you compile with targetSDK=21.

And, obviously, the MOGA library does use a service with a non-explicit intent.

The invalid approaches

The first approach is to compile with a lower targetSDK, but then you can’t be on Android TV since the required manifest options aren’t available.

The second approach is to remove MOGA support, but then, the people that has those controllers will stop to have it supported.

The third option is to have multiple apks, one for Android TV with targetSDK=21 and another one for everyone else with a lower targetSDK. This sounds reasonable, but there is no clear way to separate Android TV from the rest and -as I was told by a Google Dev advocate- this was done in purpose.

But hold on, there is one more option, and it is the most obvious one, once you think about it.

The obvious solution

You just need to modify the code of the MOGA library to make it be an explicit intent. Except that the library is not open source.

It was pointed out to me by Robert Broglia, developer of Snes9x-EX+, that you could just use a decompiled version of the library, and then modify the init method. As you can see on his github project.

I’m really not a fan of having full libraries inside the source code, so I tried to just extend the controller to replace the init method for one using the explicit intent, that would have been a few lines of code, but the class is final.

At the end I created a new Controller that uses the explicit intent (and basically everything else is the same) and placed it under the package com.bda.controller.

It could have been easier

All this could have been much easier if

• MOGA had updated the library
• The Controller class wasn’t final

If you have a game that is MOGA-enabled and you want to port it to Android TV, I hope to have saved you some headaches.

A few weeks ago I released SpaceCat HD for OUYA.

SpaceCat is the only game I have that makes sense to be played with a gamepad, and because of that it supports most of the gamepads already, including Gametel, MOGA, NVidia Shield and some others. It was also ported to GameStick, but that is another story.

A different SpaceCat

There is SpaceCat (3D), which is a Free2Play game with really low conversion. The first 24 levels are free and the remaining 24 levels can be purchased with In-App Currency.

Then there is SpaceCat HD, which is a premium game. All the levels are unlocked and it has no ads.

OUYA has a Free to Try mode, so I needed to find some middle ground.

SpaceCat HD for OUYA has the first 9 levels for free, then the game can be unlocked with a single payment.

Some other changes were required for OUYA, mainly updating some assets and strings to display the special names of the buttons. Also reimplement In-App purchases to use the OUYA library. All in all, It was actually quite straight forward.

A better App Store

One of the best things of OUYA is that it is a much better App Store than Google Play.

New releases get visibility, so does the top 30 games and so does the hand-picked features.

There are about 20 new games released each week, so you can expect to be in the home screen for about a week. Be aware that the screen has space for 4.5 tiles and to see the rest users have to scroll. Your first day is probably going to be the best.

Show me the numbers

This is the chart of the downloads of SpaceCat during the first weeks on OUYA.

We have a solid start on the first 2 days when there were ~75 downloads. This is when SpaceCat HD was on the top of “New Releases”. As you can see, the longer the users have to scroll to find you, the less downloads you get.

Then SpaceCat was featured. That put it on the front. It was the first thing you saw when you start your OUYA, literally the best spot, and it really made a difference: Over 125 downloads per day.

After a few days, the feature games rotate, and you can see how the downloads per day get down to just 50.

But here comes the last factor. Given the amount of people that downloaded (and played) the game, it also made it into the top 8 of the trending games, which helped it getting a decent amount of downloads. I’m proud to say that it even topped Final Fantasy III.

But after the featuring, and given that SpaceCat is a game you usually play for a week and you either finish it or get bored, it slowly disappeared from the top 30 and not it only gets a few installs per day.

How does it compare to Google Play?

The amount of downloads is in a complete different league. When SpaceCat was featured on Google Play (on 2012) it was getting over 50,000 downloads per day, 400 times more than on OUYA, Even today, SpaceCat (3D) get about 100 downloads per day.

But, partly because the monetization is different and partly because of the public is different, the percentage of users that purchased the game is way higher on OUYA than on Google Play.

The conversion of SpaceCat HD on OUYA is over 5%. Which is very impressive.

Also, there is the human aspect. The people from OUYA were a pleasure to work with: Diligent, responsive and professional. Google Play has improved that, but they still have a long road to go.

TL;DR;

There is a chicken-and-egg problem with OUYA: There are not many players, so it is not worth developing for it alone. Since there are not many games for OUYA (750 is still an impressive amount), it is not very attractive for players.

OUYA has the Free the games initiative to break the vicious circle. I like it, and I really hope they succeed.

If you have a game that already supports gamepads, you should totally port it to OUYA, it is not much work and it is a very rewarding platform. If you have a good game, the chances of being featured are very high, and that is always nice.

I wish there were more OUYAS out there so I could develop a game exclusively for them, but right now it is just not viable. Time will tell.