Really, really needs a has_trackpoint feature flag :)
Added to the roadmap :)
The roadmap is very long though. Everybody has different features that are impportant to them. And we research all the data manually.
It would be interesting if you could change specs. Say you wanted a laptop like the one you currently have, but brighter and with more RAM.
I find it disappointing that it doesn’t compare performance metrics. It’s easy enough to pick a benchmark that’s more or less accepted by the industry to measure CPU and GPU performance, and map those out as well.
Which benchmark would you pick and how would you get that data for specific laptop models?
Notebookcheck has a decent selections of benchmarks for components https://www.notebookcheck.net/Mobile-NVIDIA-GeForce-GTX-1060-Laptop-Benchmarks-and-Specs.169547.0.html and laptops https://www.notebookcheck.net/Alienware-x14-Review-The-world-s-thinnest-gaming-notebook-requires-compromises.608607.0.html
Good idea. I guess I would have to ask them if using their performance values on Product Chart is ok with them.
I second notebookcheck as a source for GPU benchmarks. Their methodology is pretty solid, they cover a lot of devices, and they’ve been collecting benchmarks for more than 15 years now.
They also cover other interesting metrics, such as display brightness and color accuracy, which is not always available on a spec sheet.
It would be hard to get the scores for a specific laptop, given that the TDP and heat management are often very different across models with the same specs.
For CPU I’d personally go with the Passmark CPU benchmarks, since they are a pretty good indicator for the type of CPU loads I run as a developer (they tend to map linearly to compile times for example).
For GPU, I think the problem is a little bit trickier, but one starting point could be selecting a number of common benchmarks from Notebookcheck.com, which does a lot of in depth laptop benchmarking and reviews, and average them out.
It wouldn’t be perfectly reliable, but it would be a lot better than nothing.
Does this work for anyone else? I couldn’t find some way of spelling “Oryx Pro” that it would understand.
My laptop, the T480, isn’t listed; only the inferior T480s is.
It doesn’t seem to put sufficient weight on the weight; I entered a laptop with a weight of 2.6lbs, and the top result returned was 4.6lbs. I would classify these two laptops as not similar at all.
Which laptop is it?
Acer Aspire S3-391-6046
When it was new there were quite a few similar laptops on the market (4GB RAM, 13.3” screen, core i3 processor, under 3lbs), which is why I was surprised to see a heavier laptop with larger screen presented as similar.
In the ultrabook market, weight is such an important factor that I would consider e.g. a 2.6lb 4GB device to be more similar to a 2.8lb 16GB device than to a 3.8lb 4GB device even though just looking at the ratios you might come to the opposite conclusion.
Interesting. For reference, it is this one:
Which laptop would be more suitable as the “most similar” than the Dell it finds?
I wonder how they normalize the factors such that one doesn’t contribute much more than the other. For example, I imagine changes in price (~±$1000s) would weigh in much more than changes in screen width (~±5 in)
I convert the values into percentiles. Over the set of all laptops.
So a $1000 higher price and a 5 inch bigger screen might both mean 80 percentiles of difference for example.
Can’t compare on the difficult bits…. quality of screen (glossy or not, viewing angle, screen type), type of keyboard, layout of keyboard, type of trackpad / pointer.
Exact specs of the hardware are usually not very relevant, any current/recent CPU will be fast enough, memory and storage can/could usually be upgraded after buying (I usually buy the lowest spec machine, and upgrade after-market)
So yeah, I’ll stick with my T480 (which isn’t in the list even, only the T480s)
Update: typo (and -> any)