Feedback & Analysis of PvE Personal Scaling Over 471 Battles

nwman

I've been playing a lot more without Cage than I used to, and I want to say it feels like scaling has been slower so this seems to confirm that.

grunth13

I have been playing like I usually do since the community scaling has gone off. I find that the nodes do not get much higher than when I start the pve. The max I have noticed is about 10 level increase. What I have noticed is that nodes never go down. I have wiped 3 times in a row on the same node and the node not come down at all. Maybe it was already at its floor level, but 6 healthpacks being used should tell the AI that maybe its too hard (I won it the 4th time with a different team with minimal issues and a lucky cascades galore so maybe I should have just changed my team instead of using healthpacks). Since community scaling has been dropped, I feel my nodes haven't varied much even if I wipe a goon only node with no damage over and over. Granted I have almost all my 3* at 135-140 levels so they are all useful for me. The nodes have increased in level from me increasing my team levels to 135 from 120. I think that this is the problem with even neutral scaling. Even though I am not having a hard time with the scaling, the overall levels are increasing with my increasing. This in itself doesn't pose a problem, but it does cause problems for end of sub grinding because it takes longer to fight the same fight against vs the levels I was facing when I was at level 120. Covers are definitely worth more than levels as many people have noted. I have a level 108 Fury who is 1/5/3 who's EP does base 2541 damage. With him boosted to level 198 it does 3511 damage. Thats a 1000 HP which is nice, but the enemy increases its health but more than that and there are 3 of them. This is where leveling characters hurts you in PVE. You have to grind much earlier and much longer for the same result. I think someone had done an analysis of this about 2 weeks ago that explained it much easier than I did now, but that is my 2 cents. So in conclusion, level your characters for the fun of it because it won't do much in the sense of making PVE unplayable, but don't expect to get the same placement unless you want to devote more time to PVE.

Unknown

Phantron wrote:

There's obviously a limit to how much taking damage can affect your scaling...

For the duration of this experiment, the only observed limits to downscaling were initial node levels.

Phantron wrote:

I suspect there's a point where your roster/past performance is so strong that the game figures anytime you played poorly/lose you must be faking it...

This will likely depend on whether Demiurge_Will and MPQ customer support are correct about initial node levels, or if IceIX is.

Jack0fAllGames wrote:

Did you notice any evidence of a delay in scaling when playing quickly? I've had numerous occurances where I had a performance that I fully expected to result in an uptick and got no change, but then quickly moved to another battle that should have resulted in either no change or downtick and instead "resulted" in uptick(s).

As you can see from the data plots, the majority of battles do not produce any change to node levels, even when the health lost is substantially above or below the "100 damage per level" line. This is consistent with the model I proposed in the scaling ticks section, and the delays you are experiencing may just be an artifact of the breakpoints built into the system. The unpredictability of when scaling ticks will trigger is the main reason I had to fight so many battles to accumulate statistics. In an ideal performance vs. node level graph, the points for scaling downticks and upticks should lie in distinct clusters, and their boundary would then demarcate the performance level required to keep scaling perfectly neutral.

As for the unexpected upticks, the only way to see if they should really be unexpected is to start recording them. I certainly had to adjust my assumptions about what a "hard battle" actually is in the scaling system. The amount of health lost the game expects to see in a high level node can be quite demanding.

Sundance2112 wrote:

1) Did you record simply health(before) - health(after)? Or record how much damage you actually took throughout the fight?

2) Do you have any reason to believe that the scaling algorithm looks at anything other than health(before) - health(after)?

1. I recorded the maximum health, initial health at battle start, and final true health at battle end. The two plots I posted were based on true health damage, and percentage of true health lost relative to initial health. I also made an effort to avoid using characters with true/temporary healing abilities, so this factor should have a limited impact on the data.
2. That one scaling downtick in the first plot appearing below an uptick indicates that absolute true health lost cannot be the sole factor in triggering scaling ticks, so there are almost certainly other factors at work here.

Zen808

I don't suppose you kept data on the EotS Survival nodes? I cannot for the life of me remember if I've ever seen scaling go up after finishing a Survival node, even though any that start with all goons tend to be pretty easy to escape with little damage.

Also, as to other factors that might influence scaling, you have to assume that they would build anti-sandbagging safeguards into the algorithm. I know if I was designing it, there are 2 that I would certainly put in:

-Factor of final AP. This is to catch the guys who only use match damage, and don't fire off powers, in an attempt to prolong a match and take more damage.
Also, because I have a little troll in me, I would make it so that someone who finishes with over 100AP unused automatically gets an uptick to their scaling. I don't _think_ the devs are that cruel, but who knows.

-Minimum damage inflicted. This is to catch the guys who sacrifice massively underlevelled teams to tank their scaling. If the player fails to inflict "x" amount of damage before wiping (say 10 HP of damage per level of enemy), then the wipe doesn't count towards downticking their scaling.

Which is not to say that's exactly what they do, but it would be naive to assume they don't have something in place to trip up sandbaggers.

Xenoberyll

Phantron wrote:

I suspect there's a point where your roster/past performance is so strong that the game figures anytime you played poorly/lose you must be faking it because there's no reason why someone so awesome like you should ever lose. If you have crossed that point, which anyone seeing high level 200s most certainly have, you're not getting into the game's good graces, ever. Maybe prior to that you can convince the game that you suck and need some help, but there's no going back once the game decides you're awesome.

i have a feeling i may have crossed that point, but probably everybody feels that way sometimes

Anyways...awesome post Vynyv

Unknown

I've done well in PvE for a while and scaling has increased because of that, thing is I oftentimes don't have a good second team and scaling only increases...

How do I lower scaling? Do I join an event and ignore it? Do I join it and sacrifice the weak characters of my roster? Will the next next event be less difficult?

rdevil32

amazing post and this should be a sticky!

if i could, i would give this man some HP or tokens!

Bowgentle

Breezy wrote:

I've done well in PvE for a while and scaling has increased because of that, thing is I oftentimes don't have a good second team and scaling only increases...

How do I lower scaling?

You can't.

The only thing that has been working for people is to take a break for a few months.

Sitting out one or two PVEs won't help.
Taking damage won't help.
Finishing badly in a few PVEs won't help.

Basically, once your scaling is bad, it can only get worse.