Mathematical effect of Sentinel's (or anyone's) health nerf on win %


#1

I did a statistical study on the effects of a reduced life bar on win percentages. If we call Sentinel’s starting health to be 13 and his handicapped health to be 9, I would expect one-on-one 1-round win percentages to be adjusted by a factor of (13/9)^2, or about 2. In other words, if one of his opponents had about a 25% chance of winning before, I would guess that that opponent now has a 50% chance. These numbers would need to be adjusted to account for the three life bars in play and other system differences between CVS2 - which my study was on - and MVC3, but it’s a starting point.

My data and the reasoning behind my math can be found here:

If you wish to look at my other articles, my main page is here:


#2

Your test seems to be completely AI driven, and it assumes that a player would use a 910K Sentinel in the exact same manner as he would have used a 1.3M Sentinel. It also seems to ignore any and all character match-up specific information.

It’s clear that you put a lot of work into this, but I don’t think it actually addresses real-life, high-level decision-making very well.


#3

Sixfortyfive:
I did say ‘would expect’ and ‘guess’ in my post above, and I did request human trials in the description beneath my article. True, I would expect someone with a few ticks of life to play differently. Whether that throws off the numbers significantly, I don’t know. But I would guess that if two opponents did not change play style, irregardless of matchup, that the win % would be effected by roughly the same factor, as I drew data from multiple match-ups with unchanging play styles. See my article for more detail on using the factor.

I understand that Sentinel’s health total is 905K, though I rounded to 900K for my example in the post above.


#4

I drew a picture of Sentinel from memory.


#5

The proper way to put this relation is through a gaussian curve, not a direct relation. If an oponent had a 50% chance before, it’s not going to have a 100% chance now.


#6

How about he’s a lot easier to beat and people will find other characters that are better.


#7

You’re not taking into account that many characters could already kill Sentinel in one combo, so this changes nothing.

Your usage of math is fine, but it has almost no correlation to an actual match.


#8

*910K Health

There’s been a few videos now showing he still lived at 905 and died 910

Nice attempt though.


#9

Haha! Okay now do Mango Sent.


#10

I lol’d. :rofl:


#11

rofl


#12

It has been suggested that some ?rounding errors? result in the damage display becoming inaccurate. My original test to determine the 905k value used primarily single hits to reach the 905k determination, and a single two hit combo at the end to test for the values below 10k.
Died at exactly 905k damage received. (Exact two hits varied to test for different ranges of damage)

Some of the more recent videos used to result to determine a 910k result used several combos for the determination. So if what’s suggested is true, using several combos is probably not the most accurate way to test his health value.


#13

Mikros:
I was giving a simple example. I thought it might be a particularly relevant one, because if people have a feel for how a one-on-one match with Sentinel goes - if they feel his opponents win about 25% of the time - and if my theory is correct, then it would seem the health adjustment Capcom made is more likely to balance him, at least before accounting for assists, etc. If you read the article, I explain what to do if you start at the 50% mark or pass it. I tried to make it so that it could roughly be done in one’s head. If, for example, Sentinel’s opponent had a 50% chance of winning before the handicap, you would divide his chance of losing by 2, leaving Sentinel’s opponent with a 75% chance of winning.

Symph0ny:
I don’t do much to try to explain why the mathematical relationship exists (or why it doesn’t), but it seems to, based on the data.

Mr. X:
I used the 1st post in the ‘Marvel vs Capcom 3 Adjustment Patch’ thread as my reference. The difference between 900K, 905K, and 910K seem relatively minor, regardless.


#14

Reply of the Year.


#15

Mathematically, this is the most sound argument yet


#16

that’s quite the math problem.


#17

Speaking of math problems, this may not put the 905k vs 910k health issue to rest (only Capcom’s official word will), but it seems clear to me Sentinel’s health value is now 910k instead of 905k. If you think of those numbers in relation to Sent’s original health, you’ll see that 910k is cleanly 30% less than 1.3 million whereas 905k is 30.384615384615…% less than 1.3 million. Sure, Capcom may have taken away 30.384615% of his health, but I find it easier to believe they simply took away by 30% (390k) leaving us with 910k. That and the fact that people have shown not everything under 910k damage kills Sent, but anything over that value does. Not a big difference either way.


#18

That 910K really gives him a bit of leverage to continue on like his former 1.3 self.


#19

I don’t think that’s true at all.

Just mathematically speaking, yeah, Sent dies quicker when you touch him… but he’s still really hard to get in on… he still does more damage than anyone… he still has the best XF… he still has silly hyper armor… he still has super safe defense in Jump back S, that leads to 100% combos… he’s still better than 95% of the cast…

He still has more life than any character on my team…

He’s still really fucking good.

This idea that Sent somehow sucks now, because of some random nerf that didn’t even begin to address what makes him good, is insane. Sent had like 6 reasons why he was a ridiculous character, and his vitality was the least of them all.


#20

He’ll still be alright. I never said he sucks…it’s just there’s no way he’s going to be the dominant force he once was. Assist is easier to kill, he’s easier to kill, he already had bad matchups…just dont see it. I just can’t put him in high tier anymore with that low health. He has good shit in a sense like if he catches you not blocking he can hurt you really bad or pretty much kill you but a play style like that I dont see lasting. People are going to figure out new shit and he’s only going to go down unless people find out some real good new shit where like just blocking whatever he’s doing or outzoning him wont shut him out of most of his options.

He’s still better than 95 percent of the cast in probably like an ease of use sense but…even with his ability to 100 percent people I just can’t see why it’ll last minus using him as an assist/anchor. If he really was that good Combo and Justin would use him instead of using characters that have like real ways to put you into situations and dont just sit around waiting for one or 2 things to hit. On top of having a bunch of matchups issues. High mid at best.

All I’m calling out is the end of Sent being in the top tier. Nothing else. People are going to have to figure out a lot more about Sentinel than the 2 things they ever do with him now to keep him top tier. He’ll still probably get played because his assist helps out a good bit of characters pretty nicely but the fact that he would survive shit that killed most other characters IMO was the only reason he hanged in a game where people haven’t completely tapped any of the other character’s potentials.