1
\$\begingroup\$

I made a game for my bachelor final project. How many testers should I get to play-test my game? Is there any theory about that?

Would be really helpful if you can give any citation or source, since most likely my examiner will ask about that.

Thanks!

\$\endgroup\$
6
\$\begingroup\$

Get as many as you can reasonably manage. The more prepared and organized you are for testers, the more you'll be able to manage. If you have an automated system set up for collecting data and deploying your game for testing, you'll be able to manage more testers than if you were attempting to manage them manually.

It's also important to get the correct variety of play testers. You want play testers who are part of your target audience. You'll likely want your testers to be people who are not connected to you. They're more likely to give you an honest opinion.

Read this question and answers to ensure you're getting the most from your testers. There's also some good information about getting play testers and tips to teach your play testers how to test in this question.

\$\endgroup\$
  • \$\begingroup\$ Thanks for the reply! However, if the examiners ask me why I chose that particular number of test subjects(now at 22 testers), how will I answer? \$\endgroup\$ – Vhutama Jun 6 '13 at 5:49
  • 1
    \$\begingroup\$ @Vhutama This answer and the questions he linked seemed like adequate enough sources. If you need to find officially academically sanctioned sources, then you might be on your own. \$\endgroup\$ – Jesse Dorsey Jun 6 '13 at 12:07
  • 1
    \$\begingroup\$ "It's as many as I could reasonably manage" sounds like a reasonable reply. \$\endgroup\$ – MichaelHouse Jun 6 '13 at 13:12
  • 3
    \$\begingroup\$ Also, this is a non-blog site. It's full of professionals with professional answers. If you need to link, link here. \$\endgroup\$ – MichaelHouse Jun 6 '13 at 13:22
  • \$\begingroup\$ @Byte56 I wish could vote up your comment ;) (But I can't. Maybe I don't have enough rep?) Anyways, thanks for the answers guys! \$\endgroup\$ – Vhutama Jun 6 '13 at 13:30
1
\$\begingroup\$

Calculate how much testing will have to be done (as time, say you come up with a figure of 200 hours), then figure out how much time a single tester would spend on the task, and you have a figure for the number of testers you need.
Of course that should take into account retesting bugs they report, duplication (especially with volunteers testing games), people not actually doing the work they sign up for (especially with volunteers testing games, many just want to play something new and shiny before anyone else, and get a free copy at release date as a reward).

There's no magic number, depending on the size and complexity of your system, the number could be anywhere from two or three people (you do want some duplication of effort, to cut chances that someone misses something) working for a few hours over a weekend to dozens of of people working full time for months.
And with consumer products there's the added factor of disparate hardware/software combinations it will run on. It wouldn't do to test a graphics heavy game for example on only 2 computers, both with the exact same videocard, operating system, and drivers.
It will need to be tested on a broad range of hardware/software combinations to determine if it works properly on all of them (or at the very least to be able to mention in your documentation what the hardware/software requirements are to run it).

All of this is true for both games and any other system (and not just for software either).

\$\endgroup\$
  • \$\begingroup\$ Same as the other answer, sadly I need a source to back this up, or maybe, a source or citation that state I don't actually need a magic number. \$\endgroup\$ – Vhutama Jun 6 '13 at 6:10
  • 2
    \$\begingroup\$ @Vhutama You will find that just simple cold logic and real life experience has a much bigger part to play in IT than any number of websites or "studies". \$\endgroup\$ – jwenting Jun 6 '13 at 6:11
  • \$\begingroup\$ I'm totally agree with you, but this is an academical project. I will have examiners, and they will ask me for academical or statistical proof, though probably any link to a book or any non-blog site will suffice, and I can't find any. Still, thanks for the answer, I'll keep it in mind when I start my own game. \$\endgroup\$ – Vhutama Jun 6 '13 at 6:38
  • 1
    \$\begingroup\$ @Vhutama For established "best practice" you'll likely find little or anything that your professors would find sufficiently "scientific". This is a constant area of tension between science/education and industry, in that people get educated into practices that should work in theory but in the field turn out not to work at all. You could take a look at books about software testing, those will tell you how to determine testing criteria and methods, but can't give you numbers for how many people you want to involve. \$\endgroup\$ – jwenting Jun 6 '13 at 6:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.