That is a theory but basically impossible to prove (unless someone invents a time machine).

There are however quite a few games that have sold well and were well received and the sequel couldn't match the sales (LBP2 for example).
What's the reason for that ? More than enough people could try it, so many decided that the game wasn't for them.

The same would happen to games that already start at a budget price. But the publisher gets only half the revenue from the guy who doesn't like the series.

The huge flaw with that logic is that there are already tons of games out there and people have barely enough time to play them. Price all games at $30 instead of $60, but then people would need to buy twice as much games as they already do and that's something that simply won't work.

Which the same reason why I don't think that services such as PS+ are very profitable, especially for the publisher. They get a tiny cut from the monthly fee, but their games go out for "free" to millions of PS+ users.
Let's say 1m PS+ users downloaded a game (for example Sleeping Dogs) and the publisher gets 10% of the monthly fee from these 1m people. That's ~$400k.
The publisher would need to sell ~30k copies to get the same amount of money.
Since these people downloaded the game, they were at least somewhat interested in it, so the publisher basically lost 1m potential buyers.
So on the long term the publisher could've surely earned more with the game.

Now the other side is that it could've attracted more people to buy Sleeping Dogs 2, but at the same time that could be offset by people who deliberately wait for SD2 to come to PS+ instead of buying it right away.


About iOS and Android games:
http://www.biu-online.de/de/fakten/marktzahlen/datentraeger-und-downloads/mobile-games.html

19m sold games in Germany, 2% of the gaming revenue.
Tiny.

What are forgetting is that loads of AAA titles make most of their revenue during the first month after release, and that the sales afterwards are just ... inconsistent. Price drop occurs after 1 year, most of the time. It's not about making revenue, it's about expanding fanbase

Actually outside of Steam deals and huge sellers (Diablo 3, Starcraft 2 etc.) don't expect much to change on a weekly scale.

http://www.biu-online.de/de/fakten/marktzahlen/datentraeger-und-downloads/alle-plattformen-im-vergleich.html

Most PC sales come from budget priced games and indie games.
A top seller list would become meaningless quickly when there's one HumbleBundle on sale for example and at the same time a PC version of a popular multiplat game is released but only found on place #6.
So the future top seller list would need to be sorted by revenue and then the HB games wouldn't appear either (just as they don't appear right now).

Furthermore:
http://www.biu-online.de/de/fakten/marktzahlen/datentraeger-und-downloads/distribution.html

38% of all games in Germany are sold digitally, but they only represent 11% of the revenue.

Another thing:
http://www.biu-online.de/de/fakten/marktzahlen/die-deutsche-gamesbranche-2012.html

1.5bn Euro revenue from games, 11% of them are digital, so 165m Euro revenue.
PC gaming brought it 464m Euro revenue, so AT BEST 36% of all PC gaming revenue is from digital sales.

Great data. Thanks for the heads up!

There's no law needed.

They just need to ask enough customers to get accurate data, instead of asking the retailers how much copies they have sold of a certain game.
Probably won't be as accurate as retail data, but even Famitsu and Media Create don't agree on everything now do they ?

That is true. But I think that might become to unaccurate. Let me give you an example: On the 3ds e-shop there is an review system in place. In Norway, a Pokemon game has sold at least one copy more than pushmo/pullblox. However, the pokemon game has 200 reviews, while pullblox/pushmo has 1000. 

Both games are sold only in online stores too. Asking consumers just covers up a little bit of the market, as many won't bother to do the simple task it is to say vote in a pool (you press one button). 

I suppose they could try to scale it up or something, but I still don't think that would be a smart thing to do. VGChartz does the same thing (I believe), but this company trying to do this probably would need to be rather accurate with their numbers.

No really you just have to ask enough people (not look at reviews/ratings/views/demo downloads/leaderboards or whatever).
But yeah the problem would be that Norwegians may buy less of a certain games than British people do (percentage wise) and if you only ask British gamers and not Norwegian ones, there are bound to be errors.
Still that can be solved by surveying Norwegians.

Now the real problem would be that you'd need to survey about 1000 people for every platform, in every country, every week (or month) to get the data and that's going to cost, so most would probably only do that in the most important markets and then extrapolate the numbers.