DOC Update percentages for eagerloading benchmark

[GuySartorelli]

Description

While working on silverstripe/silverstripe-framework#10929 I had to do some benchmarking, and I wanted to validate my findings against the ones that were done for 5.1.

The percentages don't seem to be correct - this PR updates them based on https://www.calculatorsoup.com/calculators/algebra/percent-change-calculator.php

Issues

• Provide a way to pre-filter and pre-sort eager loaded data silverstripe-framework#10929

[emteknetnz]

I'm pretty sure this is incorrect

Putting in 0.1819 and 0.1453 into the linked calculator shows a 20.1209% decrease, so I see where you got the 20% from.

However the changelog says that it's performs 25% FASTER, not that it's resulted in a 20% decrease in time

e.g. You make a 500 km trip at 100kmh in 5 hours. If you make the same trip in 4 hours that's a 20% decrease in time. However you'd be going at 125kmh which is a 25% increase in speed, e.g. your vehicle is now 25% faster

We had some argument when the work was orginally done about which metric to use i.e. "time taken" or "faster". We ended up going with faster

I'll close the PR