League of Legends: Penetration Changes

I played a ranked 3v3 game on the weekend with Robb and Adam that really got me thinking. The other team played 3 spell damage based champions. This feels wrong on the surface for a couple of reasons... The first being that typically melee bruisers dominate the map and they didn't have any at all. The second is they all did the same type of damage. As such our team could get away with ignoring armour and focus entirely on magic resistance as a defensive stat. They ended up blowing us apart. We stacked magic resist (maybe not as early as we should have, but we definitely had a lot by the end) but they stacked spell penetration and it really seemed to do quite a job on us. I did some rough napkin math after the game and thought my MR of around 150 was cut down to around 50 with the items they bought. That's a lot like not having any MR at all!

The season 3 changes made resists more expensive to buy and changed the way penetration stacking worked. Did those two changes combine in such a way to make it impossible to actually build defensively? Would we have been better off stacking health? Should we have ignored defenses entirely and just hoped we could burst them out before they could burst us out? We did have Kayle on the team so it's plausible that we could have glass cannoned them out. I want to look more into how penetration is set up now to get some answers...

Enemy Rumble:
8.55 magic pen from runes
8% magic pen from masteries
15 magic pen from blackfire torch
35% magic pen from void staff

Enemy Diana:
8.55 magic pen from runes
8% magic pen from masteries
15 magic pen from blackfire torch
15 magic pen from sorcerer's shoes
35% magic pen from void staff

I had 173 magic resist at the end of the game if my wit's end was stacked up. With the new way of calculating penetration, and using Diana as she had the most in this game... My actual magic resist against her was 173*(1-.43)-38.55=60. Under the old way it would have been (173-38.55)*(1-.43)=77. A difference of 17 which doesn't sound that huge, but it works out to about 11% more damage taken.

What about if I'd gone for health instead? 130 of my magic resist came from items bought more or less solely for the magic resist. Without those I would have had 43 magic resist. Under the new way of calculating penetration that would leave me with 43*(1-.43)-38.55= -14, except the formula has a cap of 0 for magic penetration. So I would have had 0. For reference, under the old way I would have had (43-38.55)*(1-.43)=2.5 magic resistance.

I have to spitball an amount of health here, but I figure I spent 7350G on the magic resist items and could have bought something like 1550 more health with an overlord's bloodmail, phage, and rylai's. That's only 6820 but it's close enough. In the actual game I had 2400 health (with 300 from my magic resist items), so I could have gone up to 3650 health. 60 magic resist means I take 62.5% damage from magic sources, so my actual EH would have been 3840. With the health items I would have been at 3650 EH. Unfortunately the blackfire torch also does a magic damage dot and scales based on my max health so even if they'd been about even it would have been worse to stack health. That said, focusing a good chunk of my money on either health or magic resist has about the same impact on surviving, and neither one would have been good enough in the actual game as it turned out. Magic resist was still better than health even with some of the magic penetration wasted in the health build by dropping me below 0.

Ignoring any defensive stats at all and going full glass cannon would have meant I'd die very quickly. I'd have had 0 magic resist and only 2100 health. Maybe this would have been a reasonable plan against a single target team but having melee bruisers against Rumble/Diana/Alistar meant bad things with all the AE damage they did. I don't know that we could have bursted them out even if we'd skipped defensive stats, especially since they had 2.5s immunities from items themselves.

So maybe it was a team comp failure? I donno. We'd just won a game pretty handily with the same setup, but they were playing a typical setup with Olaf/Rengar/Chogath. Maybe we could have won if we'd played tighter? One thing that is pretty clear to me though is stacking a damage type is not the end of the world anymore. At least with magic damage there's more than enough good penetration items around (they didn't even buy abyssal scepter) to completely negate any threat of stacking magic resist.