Reformatting some comments I made earlier since it’s probably worth a new thread specifically.
WyomingMist linked a stats sheet posted by Blizzard on the BlizzCon website, and I noticed it contained the number of Endorsements given in Overwatch.
This is interesting, because Endorsements are a measure of the number of games played. For the curious, here’s a way to calculate the minimum # Overwatch games completed per time and a guesstimate for the active playerbase based on this statistic.
Summary: there’s a minimum of about 40 Overwatch games per second being completed this summer. One very rough estimate for the current (Summer '18) playerbase is 3 million (note this requires order of magnitude style guesstimates for the calculation–but you can easily plug in your own numbers below and get your own estimate, too!).
Stat
Value
Calc as
(1) endorsements
7.4*10^9
const
(2) endorsements/mo
3.7*10^9
= (1) / 2mo
(3) min # games/mo
102.8*10^6
= (2)/(12 players/game * 3 endorse/player)
(4) max # min/game
60
upperbound guess
(5) min play time (h)/mo
102.8*10^6
= (3)*(4) / 60min/h
(6) min h / active player
2
const
(7) max # active players
51,388,888.9
= (5) / (6)1 incorrect, see note!
1. did this bit a bit too quickly yesterday--oops! as pointed out by @SeraphLance (thanks) this is incorrect bc for an upper bound on the active playerbase you'd need to assume 0 endorsements/game. i.e. this equation doesn't work! you can still use it as explained in italics below to get a guesstimate for the active playerbase, though (no absolute bounds). note, through (3) there is no guesswork at all (there is no more than 12*3 endorsements per game in any mode of Overwatch currently). the min # games/mo still applies, and given (4) then (5) holds as well.
*Note if you change (3) to a more realistic (but still very rough!) estimate instead of a bound (say 10 endorsements/game), and change (4) to average # min/game (say 10), and change (6) to average h / active player (say 20), you’ll get a guesstimate for the # active players. In this case: 3.1 million. Again, feel free to plug in your own numbers and arrive at your own guess! I am just providing a framework here.
**Also note, 40 games/sec is from (3) converted to games/sec instead of games/mo.
If you have anything concrete to mention, feel free to do so. Right now, sounds like you’re just trolling and/or don’t understand at all. If you have nothing constructive to add to this thread please just leave. Unfortunately, I have no way to mute you as this is my thread.
I’d like you to please refrain from posting further in this thread. You are just handwaving at this point and wasting my time; to me this feels like trolling. Please take it elsewhere.
I wouldn’t even call this a rough estimate, with all due respect.
The problem is that you start off going for a lower bound on the number of players, and the logic checks out until you get to #6. When you’re trying to get active player counts, that’s an upper bound. Mixing upper and lower bounds is bad juju, because now you’ve got a number that’s effectively meaningless.
If you wanted the lower bound, rather than 2 the number ought to be 24*31 = 744 (the most hours a person could be logged in a month), which puts the estimate at… 138k. A more reasonable lower bound, but with such a margin of error as to be useless.
Incorrect. Trolling is baiting someone. Leading them on and handwaving without offering anything constructive. In addition, use of thinly veiled insults is also something I’d consider trolling and I’d encourage everyone to stay polite here so this thread isn’t locked.
I did not attempt to calculate the fewest players possible. For the active player base estimate, the 3 million estimate is a very rough guesstimate, agreed. As should be obvious from the guesses taken. I make no attempt at even confidence intervals; I’m guessing at orders of magnitude for that.
The point for the paragraph in italics is to note how you can follow roughly the same steps to get a different answer. It’s there so people can plug in their own numbers and easily get their own guess for the active player base.
As for a lower bound on the playerbase count. To get the fewest players possible (the lower bound of active players), we need:
a. fewest games possible given # endorsements -> given by (3)
b. each game lasts as long as possible -> given by (4)
c. every player has as much of the global playtime as possible (plays as much as possible) -> given by your post (24*30h/mo/player)
In other words, I agree with you. But that’s not what I calculated above (the upper bound on the playerbase–which admittedly is not a very useful upperbound as I note: it’s greater than the confirmed number of sold copies!)
I will edit the language in my OP however to make the roughness of the guess clear.
I make no claims the active playerbase count is accurate to even within an order of magnitude. As for the upper bound, yeah it’s not a useful upper bound–as I mentioned it’s more than the # of confirmed sales! (lol)
You actually didn’t count the upper bound either, because your endorsement assumptions are a lower bound on the player base, as well as your game time.
An upper bound is literally infinite, since a game can have zero endorsements.
And yes, of course you can guess the in-between values to derive an estimated number of players, but those guesses are completely arbitrary. 10 endorsements per game? Why? 20 hours per month per player? Maybe? These are very low-confidence value, which is why you have a low-confidence result.
The problem is that “total endorsements” is simply too far removed from the statistic you care about to get decent numbers. And every derivation step you make reduces the accuracy further and further.
Yeah, you’re right. The max # players a month figure is wrong. I made the post very quickly yesterday before leaving somewhere, and it’s quite early right now–not to mention I was tilted earlier at the only response earlier being non-constructive trolling. I’ll fix the OP, thanks!
Originally I was interested in calculating the min # games / month from endorsements. I quickly tacked on the bit about playerbase counts because I thought it might be interesting too, without thinking about it too much. I’ll revise…
Actually, let me link you to another thread where I posted something very similar. I calculated that during my usual playtime and SR and region, the feeling I got that there was about 1-10 concurrent lobbies at once using guesswork from endorsements was reaffirmed using guesswork from the endorsements stat.