Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I see how that can be misleading. It's globally not a percentage based score.

The benchmark is creating a global order based on which one is faster, including indirect comparisons. But that only takes care of the position, not the rating. That one is based on the percentage between direct neighbor iff there is a direct comparison in the benchmark data, otherwise it's a small .1 increment. So many small percentage increases that don't necessarily match the direct comparion between parts that are not neighbors. Sometimes that works great, sometimes not.

Here the example looks a bit pathological, that difference is further off than I expected when introducing the calculation recently. For the 12700K and 13700K, the direct comparison sees the 12700K at 85% of the 13700k:

https://www.pc-kombo.com/us/benchmark/games/cpu/compare?ids%...

For apps it's 79%:

https://www.pc-kombo.com/us/benchmark/apps/cpu/compare?ids[]...

So yeah, sorry, that part is misleading right now. I'll check the calculation.

But the ingested individual benchmark numbers and the ranking, so the position, is very solid I'd say. With the caveat that ranking position can change with more data.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: