New Steam survey results

Nice Win7 64 bit up 1.94% while 32 bit up only 0.43%.

As well Vista 64 bit up 0.04% while 32 bit down 0.86%.

So the trend towards 64 bit appears to be accelerating. Prior to this 64 bit take up versus 32 bit take up was roughly 2:1. Now it's over 4:1 (Vista and Win7 combined).

AMD actually made back a little ground on Intel, good for them. :)

I wish they showed Windows System Ram stats also like they do for Mac.

Regards,
SB
 
I wish they showed Windows System Ram stats also like they do for Mac.

I wished they showed nothing for the Mac, why Front Page something that only represents a small percentage of the users? unless.. of course.. someone.. payed.. nah.. I'm talking crazy.
 
I wished they showed nothing for the Mac, why Front Page something that only represents a small percentage of the users? unless.. of course.. someone.. payed.. nah.. I'm talking crazy.

Don't you know any Mac owners? Someone at Valve bought a Mac and now they turned into that abnoxious person trying to justify their expensive purchase down everyone's throats.

/only half kidding.
 
Yay Win7 64-bit on top! Bout time on two counts!

Man really highlights that Fermi-based GPUs have a ways to go to catch up to ATI though... hopefully the 460 will help.
 
Damn, AMD already lost the CPU share they gained last month.

At this rate, it's possible that single core CPU's will be extinct (for all intents and purposes) for the most part in ~6 months or so as far as Steam enabled machines are concerned. As well dual core continue to lose share to quad core.

And awesome to see Win7 64 bit with more share than XP. Although 32 bit OS's are still the majority of the market. :(

Regards,
SB
 
Damn, AMD already lost the CPU share they gained last month.

At this rate, it's possible that single core CPU's will be extinct (for all intents and purposes) for the most part in ~6 months or so as far as Steam enabled machines are concerned. As well dual core continue to lose share to quad core.

And awesome to see Win7 64 bit with more share than XP. Although 32 bit OS's are still the majority of the market. :(

Regards,
SB

The important thing here is that the uptake of 64bit W7 is much higher than the 32bit W7.
 
Is there any info yet how long does a system stay in survey if the account is inactive, before it gets removed?
 
Along those lines, how often do they poll? I haven't seen a message asking me to participate in the survey in a long time (perhaps they don't ask?).
 
They might resurvey systems based on re-installation or the yearly account anniversary date. My system was resurveyed last month.
 
Is there any info yet how long does a system stay in survey if the account is inactive, before it gets removed?

I guess the survey randomly samples. Which means every month there are people asked for participation untill (some number fe: ) 10000 have been reached and that's it. Not accumulative, without memory. At least that would be the statistically right way to do it.
 
I guess the survey randomly samples. Which means every month there are people asked for participation untill (some number fe: ) 10000 have been reached and that's it. Not accumulative, without memory. At least that would be the statistically right way to do it.

So in theory it could still have some ancient machine which was used to complete the survey once when the survey first launched, if it just hadn't been selected for random survey after by chance?
 
So in theory it could still have some ancient machine which was used to complete the survey once when the survey first launched, if it just hadn't been selected for random survey after by chance?

Probably, The GTX285 has been End of Sale for months and it's still gaining marketshare. so we might as well be looking at half current and half 2009 numbers for all we know.
 
Probably, The GTX285 has been End of Sale for months and it's still gaining marketshare. so we might as well be looking at half current and half 2009 numbers for all we know.

Though is that overall share or DX10 share?
Since every DX10 card switched to DX11 card grows the marketshare (in steam anyway) of other DX11 cards, and GTX285 being fastest single chip DX10 card, there's probably in general more users not upgrading from it compared to some slower cards
 
So in theory it could still have some ancient machine which was used to complete the survey once when the survey first launched, if it just hadn't been selected for random survey after by chance?

I'm not sure I understand what you're asking. :cry: (non-native here)

But I can try to be clearer. Statistically correct sampling would require never to reuse timed data. It is okay to have a random hit on a same person in the last month + in this month. But it's not okay to use the data of the last month of that person without asking again. Every months you should elaborate data only from inquiries done in that month.
This will be an accurate distribution for the steam-users if you just sample randomly even if the number of people asked is not really big (1000 * choices would be sufficient). And it will also be an accurate depiction of trends, if you do the random sampling over and over again every month.

If you would accumulate inquiries across larger time-frames you have to take care not to accumulate answers of a single person (you have to track 20mil. person's answers over several years [in theory]). Additionally the interpretation of the survey would change to the following: the market-share graph shows the change of the all-time gravitation of the user-base towards a vendor. And that's ... very pointless data, except you want to identify if you (as a vendor) and how fast you are reaching your goal of market domination. It's not really usefull data.

Valve is not uncapable and dilettant and you can asume that the statistics have been collected the right(tm) way, and every month represents only results from that month's.
 
Back
Top