Attack Increases Bugged? Soft Capped?

Dave held a live stream earlier this evening, linked below, in which he tested attack damage w and without suicides, attack synergies, and boosts using Stark vs RoL WS.
In these tests, none of the attack rating percentages described were represented accurately in the damage against WS. One viewer stated that since Stark’s poise charges raise his attack rating, he may run into a soft cap, whereas a champion with a lower base attack which is not boosted by his kit would benefit more. Another viewer also pointed out that 1 point of attack rating does not equal 1 point of damage, but even then, the percentage increase did not accurately correlate at all.
I would appreciate any clarification the community or mods could offer. Is there a soft cap to attack rating increases outside of a champions kit? Are masteries, boosts, and synergies working as intended? Would people invest in these if they knew the described percentages were not equal to the actual increase in damage?
Start watching at ~1:48:00.
https://youtu.be/NAmXo6eqzec
In these tests, none of the attack rating percentages described were represented accurately in the damage against WS. One viewer stated that since Stark’s poise charges raise his attack rating, he may run into a soft cap, whereas a champion with a lower base attack which is not boosted by his kit would benefit more. Another viewer also pointed out that 1 point of attack rating does not equal 1 point of damage, but even then, the percentage increase did not accurately correlate at all.
I would appreciate any clarification the community or mods could offer. Is there a soft cap to attack rating increases outside of a champions kit? Are masteries, boosts, and synergies working as intended? Would people invest in these if they knew the described percentages were not equal to the actual increase in damage?
Start watching at ~1:48:00.

0
Comments
The test is muddled by using both suicides and spider man stark enhanced which both increase base attack. Spidey gives up to ~200% attack increase via his poise charges.
I’m going to give sparky a simple base attack of 100 to illustrate what’s happening. He has 100 base attack and increases it by 200%(200) via poise, with suicides by 60%(60) and 10%(10) by attack boost.
Now lay this out in a formula.
Poise
100+200=300 *poise boosted attack by 200%
Poise plus suicides
100+200+60=360 *suicides boosted overall attack by 20%
Poise plus suicides plus attack boost
100+200+60+10=370 * the attack boost added 3% to over all attack
Yes recoil boosts base attack as well.
Add it to the for formula in place of the attack boost. I’m assuming 1/3 recoil (30)
Poise
100+200=300 *poise boosted attack by 200%
Poise plus suicides plus recoil
100+200+30+60=390 *suicides boosted overall special attacks by 30%
However I do feel that as YouTube journalists it is their duty to issue retractions and corrections when the are proven incorrect. If they don’t it’s just tabloid journalism.
I watched the relevant part of the video and based on my back of the envelope calculations that is both what is happening in the video and also explains many of the stream's audience comments. Dave and his audience are not accounting for relative dilution.
As far as I'm aware, all percentage buffs apply to the base value of the thing being buffed, because all percentage buffs are additive to each other. In other words, the game adds up all the buffs (and debuffs) that apply to a stat like attack, and only *then* applies the total. That means if you have something boosting attack by 20% and another thing boosting attack by 15%, that means your total attack boost is 35%. Neither applies first with the other applying second. So in net effect all buffs are relative to the base value automatically.
It would be probably more accurate to say that all percentage boosts are really percentage point boosts. Meaning: 20% boost is twenty percentage points of boost, and 15% more boost is just fifteen more points of boost. It isn't really "fifteen percent" of anything.
If you don't know this your intuition is to convert "one hundred percent" into "about double." But while you can sort of mentally deal with "one hundred percent" being just one hundred more percentage points stacked onto other percentage boosts, when you mentally convert that into "double" your intuition now wants to see something being doubled, and that intuition is wrong.