Soft cap and PVP

truedim

 Actually fully covered Strange lvl70 have 9 saved covers, but I can't respec him using 5 of them.

bbigler

Taking a character past level 325 will not change your 4* mmr. I have a single 5* leveled to 360 (Kitty) and everyone else is 277 and below. I never see a single champed 5* team as an option to attack. And I only get attacked by 5* teams when I reach the upper ranks, like < 20. 

justsing

OJSP said:

jamesh said:

I thought the MMR generally ignored the weekly boosts. 

That’s right. Otherwise, no 5* teams can find my 4* teams in SHIELD Sim. In fact, because my team usually has one or 2 with lvl 350+, those with 2 5*s can find me easily from 0.

However, in a way, Bowgentle was also correct. The 4* would be around lvl 450 when they are boosted. I couldn’t remember when exactly they reach that level, so I put a range of levels as an estimate.

I’m guessing justsing is actually right there at the crossline. Your lvl 327 is counted as one championed 5*, while the lvl 325 is still counted as a 4*. That’s why you could queue single championed 5*s below 900.

Let us know what happens if that lvl 325 becomes 326. I’m guessing you will start seeing double championed 5*s a lot sooner.

It’ll be interesting to see. That’s for sure. Good thing I have a bunch of max champed 3*s I can sell if I’m forced to champ some 5*s

bbigler

OJSP said:

bbigler said:

Taking a character past level 325 will not change your 4* mmr. I have a single 5* leveled to 360 (Kitty) and everyone else is 277 and below. I never see a single champed 5* team as an option to attack. And I only get attacked by 5* teams when I reach the upper ranks, like < 20. 

A lvl 360 5* is different to a lvl 325 4*. I’ve been experimenting with MMR since 5*s were introduced.. I think I know what I’m talking about. And ranks doesn’t have anything to do with MMR, it‘s the number of points we have and the number of unshielded players in the shard. add: For the 5*s, it is commonly accepted that it is safe to level them up to 375 before we are matched against other 5*s.

From attacks I've gotten from double 5* champ teams, it's all about your rank and character levels, not points. For example, if I climb early in an event and reach top 10 with a score of 500 points, I start getting attacked by 5* players. I don't see them, but they can see me.  This has happened many times. If I'm the only 4* player in the top 10, then I can only queue other 4* or 3* players that are near (but below) my rank. 5* players in the top 10 can also queue other 5* or 4* players near their rank, which includes me. That's why I can't see them, but they can see me. 

This exact same thing happens late in the event too when points are much higher. If I rise above most other 4* players in rank, then I start getting attacked by 5* players while I can only queue 4* players with lower ranks and fewer points. This happens in the beginning, middle or end of the event. The same thing also happens in Lightning rounds and shield Simulator. 

Rank and points are generally correlated, so I understand the confusion, but from this evidence, the MMR will match you against teams with equal or slightly lower levels and near your rank in the time slice. This also explains why I can queue 3* teams when climbing. 

I deal with this by simply keeping my rank around 25 during most of the event and then climb higher near the end and shield (or shield hop). But shield hopping when high in rank is very dangerous as 5* teams can beat me in half the time I can beat a 4* team. 

bbigler

OJSP said:

I guess we’ll have to disagree on that then.

Try climbing when an event first opens, get up to top 5 and see what happens. I assume you're a 4* player?

Bowgentle

bbigler said:

OJSP said:

I guess we’ll have to disagree on that then.

Try climbing when an event first opens, get up to top 5 and see what happens. I assume you're a 4* player?

If you climb when the event opens, you break MMR at 500, that's why you get attacked by everyone and their dog.
I think at this point even 550s can see you.

bbigler

Bowgentle said:

bbigler said:

OJSP said:

I guess we’ll have to disagree on that then.

Try climbing when an event first opens, get up to top 5 and see what happens. I assume you're a 4* player?

If you climb when the event opens, you break MMR at 500, that's why you get attacked by everyone and their dog.
I think at this point even 550s can see you.

Exactly. That's my point.  MMR breaks when you pass a certain rank, not a certain number of points. I get attacked by 5*s everyday despite my roster. And yes, sometimes they are level 500+.

Phumade

Why are you trying to talking about multivariate calculus when most of the forum posters are just taking Pre calculus?  

Tier, Level, points, and Timing reaming are all components of MMR.  Who you que and who ques you are two different kinds of MMR, which  operate independently.

AardvarkPepper

OJSP said:

But anyway, this is the exact reason why I was always reluctant to talk about MMR. I said I was going to post a guide about it on day 1000 based on my experience and other players in S1 whom I’ve played with for a year. It’s day 1485 now and I still can’t post it, because I’m tired of explaining it.

but if you posted it then you wouldn't have to explain it at length any more, you could just reference your post

no?

bbigler

Understanding the MMR is a worthwhile endeavor, but it will ultimately be a theory as none of us can prove it. I'm afraid of people taking their own experience with it and formulating their own theory based off of that alone. The best theory is one that explains everyone's experiences in PVP.  So, if someone states a "fact" about MMR that contradicts my own experience, then I know it's wrong, or at least not quite right.

jamesh

I think he's just saying that theories are a lot easier to disprove than prove.

If we treat this like a scientific theory where we make observations and then create a theory that explains those observations and can predict future behaviour, then all that's needed to disprove the theory is a single observation that conflicts with it.  We could make hundreds of observations that conform with the theory and all that means is that we haven't found any conflicting evidence yet.

bbigler

@jamesh is right, that's what I meant. 1 conflicting case can disprove a theory.

Almost everything on this forum is an opinion, including my opinions on Kitty, clearing methods, counter teams, whatever. Nevertheless, I think it's helpful to share our thoughts, insights and experiences for the benefit of others. You don't have to believe what I say, but I hope you'll try it out or at least consider it. I do that for other people's ideas because they might know something that I don't. 

AardvarkPepper

@OJSP:  If you're not going to write a post about MMR, but you're saying you're already posting about MMR but you don't want to explain it - well what's going to happen if things go on as they are?

1)  You continue doing things the way they are, but apparently you're not happy about it

2)  You stop talking about MMR but apparently you don't want to stop looking at MMR either

OR

3)  Even if you alienate some readers or only five posters are interested in what you're writing, you take the time to write up your thoughts and theories and observations anyways

Real talk here.  Even if you look like a potato, you can't let that stop you from leaving your house in the morning.  Even when people point and laugh and shout "hey there goes the potato" or "did you escape from the Pringles factory" hey, you gotta do your stuff so it is what it is.

If you don't do what you do, nothing's going to change.  Is it about whoever else, or is it about you?  If it's about you, then why do you expect other people to fight your battles for you?  If it's about whoever else, then why bother yourself in the first place?

just keepin it real

also realized I'm out of potatoes

jamesh

I'd also add that disproven theories are still useful: they obviously managed to predict some behaviour, so there must be some kernel of truth.  It could very well form the basis for the next theory.

So please don't stay silent for fear that someone might come out with a counterexample.

Rod5

I’m on day 1875, I have been on both sides of the MMR “break” and I can say that I 100% am in alignment with OJSP’s perspective, both in terms of my own personal experience and how I feel logically it could/should work.

I don’t want to hijack the thread any more either, but l’d be happy to take it offline if you want me to explain a bit more @bbigler

Phumade

Most mmr tactics are based on player psychology, habits and tendencies.  Simply put, many scoring/placement strategies would fail if put infront of a shield rank 190 player for just being too obvious.  Yet those same strategies are consistent winners when applied against newer rosters or inexperienced players because they exploit a preconceived notion of play.

In the long run, its better for you to continue the mmr discussion.  It exposes your thought process and your roster building process.  Those are the two easiest holes to exploit and use against another player.

AardvarkPepper

bbigler said:

1 conflicting case can disprove a theory.

Not so much.

1)  There is the possibility the developers change the mechanics slightly at different times without reporting those changes.  What is true for one player at one time may not be true for another player at another.

I really suspect this is the case.  Sometimes I can get 10 wins against computer AI seed teams in PvP before human teams start to show up, sometimes 9, sometimes 8.  I haven't kept careful records to investigate why, but I suspect it's a developer thing.

2)  Reports may be inaccurate.  Player bias, player mistakes, or hardware/software interactions may cause unforeseen and unusual circumstances that are not normally the case.

For example if the game refreshes or they exit then enter they may lose the opportunity to face computer AI seed teams in PvP.  Players that didn't think the refreshing or exit/entering was a factor might not report their actions, then the reports would end up giving inaccurate data in context.

3)  Even if a conflict is shown and corroborated, a theory may not be "disproved".  The theory may be mostly or partially correct and merely in need of refinement.

bbigler

Given my experience in software development, I really doubt that Demiurge periodically adjusts the MMR algorithm without telling people. I'd bet it hasn't been on their to do list for years. 

As for disproving theories, I agree, conflicts usually require small adjustments, but if those conflicts question the fundamental mechanics of the theory, then it may need to be thrown out. The theory that is, not the conflicting case.

AardvarkPepper

bbigler said:

Given my experience in software development, I really doubt that Demiurge periodically adjusts the MMR algorithm without telling people. I'd bet it hasn't been on their to do list for years. 

Here's a theory.

If there's a single "right way" to do things then players feel pressured to pursue single optimal-path play, which becomes drudgery.  To avoid just this, Demiurge has made a point of NOT releasing information regarding mechanics.

Is it such a leap to think Demiurge may have implemented multiple MMR algorithms that the game randomly switches up on, precisely to confuse players trying to figure out the mechanics?

It's what I would do.

Nyeh heh heh heh.

Rod5

bbigler said:

Given my experience in software development, I really doubt that Demiurge periodically adjusts the MMR algorithm without telling people. I'd bet it hasn't been on their to do list for years. 

As for disproving theories, I agree, conflicts usually require small adjustments, but if those conflicts question the fundamental mechanics of the theory, then it may need to be thrown out. The theory that is, not the conflicting case.

Totally agree on 1st point, why would they?

I also agree to point 2 in principle, but with something as nuanced as this I would suggest it’s worth trying to understand and make sense of the perspectives of players with a ton of experience before drawing conclusions. 

Your fundamental theories may seem like they are truth to you but there may other perspectives which on balance you may find explain your experience better than your initial conclusions.