When choosing a winning desktop PC, the raw numbers only provide part of the picture
“Deciding which PCs won awards in this month’s Labs was a fuzzy logic decision that I’d struggle to explain to DeepMind”
There were times when I lost track of the day, the month, even the year. When you’re buried in testing, things blend into one. Your life is a constant battle against benchmarks that misbehave, computers that run out of storage space and reams of identical sheets of paper with something scrawled on them that’s probably, definitely, maybe a three. Or an eight. Not sure? Time to rerun the benchmark.
In a fog of results such as this, it’s easy to be distracted by the siren call of numbers rather than the factors that will genuinely make a difference if someone buys the computer on test. And trust me, I take this responsibility seriously: 99 times out of a 100 the people who write into PC Pro after buying something we recommend do so to say thank you, but there’s always one out of 100 who is left unhappy with the purchase. Was it something I said? Something I didn’t say? A test I didn’t run?
This isn’t the same thing as the cynics I occasionally meet who, after finding out what I do for a living, declare that it’s all paid-for and that they don’t trust magazines. I’ve grown used to that over my 22 years of testing kit, and now know that nothing I tell them will change their minds – even when I point out that, if our reviews were paid for, PC Pro would have winked out of existence many years ago. Although I do confess, my voice dropping to a whisper, that Dell did treat me to a rather nice Christmas lunch once.
It helps that I have a mantra when writing reviews, one that I’ve shared with everyone who reviews a product for this magazine – and that was shared with me when I started in 1999. That is, “no opinion without fact”. Whilst there isn’t always room to explain exactly why I’ve come to a conclusion about a particular product, I’ll do my best to provide a set of reasons so that someone reading the magazine will understand that I’m not giving a product an award because of that nice lunch.
So, facts are good. When it comes to a Labs like this one, however, you can drown in them. With reluctance, I’ve had to grossly simplify my hours of games benchmarking to the three graphs you see opposite. There’s a spreadsheet that contains all the results across all the games, with yet more results on sheets when I “just wondered” how fast that particular system would be with that setting rather than the default. If you’re interested, feel free to email me
( editor@pcpro.co.uk) and I’ll send you my master spreadsheet.
Ultimately, though, deciding which PCs won awards in this Labs and how many stars out of five to give them was a fuzzy logic decision that I’d struggle to explain to DeepMind. Not just how fast a PC was, but the quality of the case, how noisy it got during testing, the level of support I believe each company will provide (not based on gut instinct but by what
PC Pro readers tell us in surveys) and many factors beside. By the end of the testing, the answer is clear. But, hey, that’s just my opinion.