As indicated in the title page, this has been a good year for us. Much of that advance has been down to optimising our ad usage. We have had no new apps but instead have been piling in improvements to our existing ones.
The most significant of these has been the Chinese/Japanese game of Go, which saw a major upgrade to provide 18 Kyu to 3 Dan play, an increase of 10 grades stronger than previous version. This was being prepared at full speed in the run-up to the coming Sedol vs AlphaGo match. We managed a release 3 days after the match started and just 4 days before GDC in San Francisco. This confusion of getting the product released and then updated again just before I left for GDC was enough for me to miss my flight(!) and then have to re-book at the last minute. That is a first for us. Somehow I managed to see all games live, one of which was during the flight from New York to San Francisco.
GDC was the same huge event, with 27,000 attendees. This is the place to meet up with all colleagues from across the world. As ever it is very stimulating and inspiring. Clearly this year's "big" thing was VR.
The next most significant app update has been Spades, where we put in some
2 man months to advance the AI. This was very successful in that the strength
of the program made its biggest jump ever between versions. Part of this advance
can be attributed to migrating the more advanced generic tools developed for
the later app Euchre, back into Spades. These allowed more intelligent testing
and removed the need for the developer to analyse plays in huge detail as
the advanced framework allowed this to be done automatically using whole game
play outs. This allowed a much more rapid turnaround. In addition some new
ideas were successfully tried for Euchre and these were migrated back to Spades.
AI Factory tried home made and semi-professional promotion videos a few years back, but this year invested in professional videos from France and India for Chess and Solitaire. These added some polish to our store listings but the impact in downloads was less than another venture: optimising the Store Listing using the Google console A/B testing, which showed that seemingly minor changes could boost downloads. Overall AI Factory downloads have since increased by some 12% using minor tweaks to the listings.
Ads still offers us much grief and a substantial body of our work has been to combat defective ads. For example banner ads that consume a high volume of CPU and/or data usage, even when the app is in the background. These are unacceptable problems that we have had to create internal fixes to combat. This flakiness of the Ad infrastructure is bad for us and bad for the Ad ecosystem as a whole. We welcome the time when this is no longer a problem.
Now we have to think about Winter 2016 edition, which we will probably publish just before Christmas.
