Yeah but the important thing here is that nobody will know what games Triggercut owns anymore.
Oh wait, they didn’t know that before either since he always had the option to set his profile private.
What a bummer.
I agree it will need a concerted effort, but Steamspy was not something easy to build either.
My guess would be: take the history of concurrent players stored in Steamspy or Steamcharts servers and try to correlate that (sampling those specific milestones) with sales behaviour. Use game tags to distinguish between different correlation behaviours (that is, see which tags influence the correlation and in which way). Run an analysis until you get numbers within a ±20% of the scurrent Steamspy estimates (for the data stored prior to the cutoff) for at least 85% of the games. Once you get that apply it to new games to estimate sales.
Of course no correlation when controlling for tags might be found. I seriously doubt that, but I could be wrong. It could also be a better estimation than the current method (which was dubious for games in the sale range of the median, so most games had quite inacurate numbers anyway)
I’m just saying it’s doable and that the data is actionable with enough effort. Steamspy also used statistical methods to estimate games it had not sampled a lot.
You run into the serious issue of forecasting outside of range since behavior shifts over time. And you have no way to recalibrate.
e.g. see Google’s flu trends from 5 years ago suddenly being way off as search behavior (and news reports) changed before they got back on top of things by modifying the models involved. Won’t be able to do that.
Bad predictions are probably worse than no forecasts.
That’s what I said upthread something like this would deviate over time. But for a time the numbers could work.
I still think the correlation of peak active users over the first two months with sales will be a very strong one nonetheless. It’s on the long term sales where stuff like genre and game length would need to be thoroughly modeled and where you would see deviations. But in terms of market analysis, the first two months are the most important for the viability of a game. And, in terms of judging the health of a game online population, current active users is an excellent data point and is available.
I just don’t see too much added value in long term sales predictions (which is what we have lost access to), except in the indie low side of the scale where they are indeed meningful, but then at that scale even Steamspy’s method were so unreliable as to be useless.
So, most of the value Steamspy provided for developers, and again, we used it and I’m sad to let it go, can still be had.
(as an aside, number of user reviews also really strongly correlated with number of sales, specially before discounts kick in, and again, with some adjustments for genre and size of the game, and it’s a measure we used for other platforms. You do need to have access to several private data points to be able to build the rough model, though)
Interesting… Steamspy will go on!
Now you can see how much money, for reals, have you spent on Steam
https://help.steampowered.com/en/accountdata/AccountSpend
I shouldn’t have clicked that link.
What’s “TotalSpend” vs “OldSpend”? I don’t know if I should be adding those numbers…
STOP TRYING TO MAKE THINGS WORSE
Reddit says
TotalSpend is all money put into your account
OldSpend is money put into your account before 17 Apr 2015.
PWSpend is something related to Perfect World accounts which is something to do with CS:GO I think.
Really? I’ve spent a lot less on Steam in the last 3 years then.
I spent half my total in the last 3 years. Probably because I buy more new games even though I buy fewer games.
I’m around $1600, not too bad.
The numbers confirmed what I suspected:
420$ total spend
400$ old spend
I pretty much stopped shopping at Steam. Their prices simply aren’t competitive where I live. There is always a cheaper source, not even counting the grey market shops.
I shouldn’t have looked…
TotalSpend $2,268.03
OldSpend $1,693.65
Hmm. I suspect if I added all the money for all the games, it would be ten times that.
halp
What do you mean, “add the money”? You mean games you bought elsewhere like GMG? For me, I’d guess external purchases are probably at least half of my total, so another $800 or so.
Thanks for this. I think I just threw up a little bit…My wife cannot see this number… ever…
$10,400 total spend.
This is my cocaine habit.
I think we have a “winner”!