The Steam Machine is destined for the console colosseum.

Before you start making plans to evict your current rig, I want to explain how the Steam Hardware Survey actually works. Rather than representing every player that uses the storefront, the monthly census instead uses a small percentage of participants who’ve opted in as a sample. It’s anything but representative of all 154 million active users, meaning the statistic above isn’t an absolute.

Does the author know how *survey sample sizes work?

Do I?

Is it because it’s opt in or opt out?

  • DigDoug@lemmy.worldEnglish
    38·
    1 day ago

    One could argue that the Steam Hardware Survey being opt-in means that it’s likely to overestimate the power of Steam users’ computers, since people with fancy rigs are far more likely to want to brag about it.

    • Ava@lemmy.blahaj.zoneEnglish
      4·
      16 hours ago

      On the other hand, I suspect that some people with high-end PC specs might be more on the “privacy inclined” side of the spectrum, and might prefer not to share data on principle.

      Though, if I were to guess, yours is probably the larger influence.

    • dan1101@lemmy.worldEnglish
      23·
      1 day ago

      Perhaps, but a couple years ago when I still had a GTX 1050 I was happy to participate in the survey to show that a lot of gamers didn’t have fancy RTX cards. I suspect there is a lot more low-end hardware on the survey than high-end.

      • Fushuan [he/him]@lemmy.blahaj.zone
        2·
        14 hours ago

        I’m just assuming but my assumption is that from the people that are active steam users, the ones that opt in on the survey are in general more involved in gaming than those who don’t, and if they are more involved, the chances that they invest more in gaming than those less involved are high, and part of that investment is the rig.

        So, 3 assumptions there, but they make perfect sense to me. If those are actually true, then the survey has a overrepresentation of extra involved gamers, and those have an overrepresentation of better rigs. Thus the survey has a survivorship bias into better rigs, not worse.

        If their conclusion is that, even with that bias, 70% don’t have that good rigs, I’d say that their conclusion is sound.

        Mind you, I’m not saying that people who do the survey have good rigs, I’m saying that the amount of people that have good rigs compared to those that don’t is probably higher in percentage in the survey due to above.