31 replies to this topic

[PhoenixFire55]

Medicine Man, on 14 June 2017 - 10:24 AM, said:

Nonsense. 1 person out of 12 can't determine victory or defeat. "Random factors average out etc etc," is just noise you have no basis for being able to prove anyway.

You can spew "nonsense" all you want, its math. Since you are ignorant and choose to remain ignorant I won't bother to explain in detail. You can read it up yourself in any of the MM-related threads on the forum.

[Medicine Man]

PhoenixFire55, on 14 June 2017 - 10:41 AM, said:

You can spew "nonsense" all you want, its math. Since you are ignorant and choose to remain ignorant I won't bother to explain in detail. You can read it up yourself in any of the MM-related threads on the forum.

Lol because you can't. Now you are just saying, "it's math!" As your argument. Weaksauce bro.

[Nemesis Duck]

PhoenixFire55, on 14 June 2017 - 10:41 AM, said:

You can spew "nonsense" all you want, its math. Since you are ignorant and choose to remain ignorant I won't bother to explain in detail. You can read it up yourself in any of the MM-related threads on the forum.

What math? Where? Empirical, not Imperial.

Edited by Nemesis Duck, 14 June 2017 - 11:00 AM.

[Kaethir]

MischiefSC, on 14 June 2017 - 10:37 AM, said:

You can seed with a bit over 40 matches. That's enough to shake out any statistical anomalies based on who you drop WITH. You can create some inaccuracy there by playing bad mechs but I would argue that's still a "you" thing.

Within 40 matches of playing decent robbits in pug queue you've washed out the bulk of "good luck/bad luck" randomness from the results. What's left is a passing accurate view of the players impact on a teams odds of winning or losing. That's a combo of how they play and what they play.

You can inflate damage and KDR by doing things that don't win matches. Only thing that inflates win/loss is consistently doing things to drive wins.

I'm of the opinion that there are too many variables for 40 to be enough, although I will admit that if you're not looking for 'skill' but just for what is anything within your control, yeah it's not as much as I was thinking.

We can quibble over over the level at which it becomes relevant, but the point is still the same. After a certain point, your W/L is reflective of how much you contribute to your team. I'm about average (1.03 W/L on ~1650 matches), which jives with my in-game experience (occasionally do very well, occasionally do terribad, usually somewhere in the middle).

[G4LV4TR0N]

MischiefSC, on 14 June 2017 - 10:28 AM, said:

W/L is accurate beyond about 50 matches. Overall is playing with you more likely to result in a win or a loss? You're 8% of your teams performance, how much you carry with that is all you.

I'm in favor of splitting pug and group queue stats though.

And then you will have to split it again because 2-man groups are so different than 12-man.

[Mystere]

MischiefSC, on 14 June 2017 - 10:28 AM, said:

I'm in favor of splitting pug and group queue stats though.

Players hanging on the coattails of their mates might have a not-so-pleasant word with you.

[Mystere]

Medicine Man, on 14 June 2017 - 10:44 AM, said:

Lol because you can't. Now you are just saying, "it's math!" As your argument. Weaksauce bro.

Did you say "math"?

[Sjorpha]

Winning consistently over time is the best indication of skill, maintaining high win rate at increasing tiers in a matchmaking system is the best indication of a progression in skill. Maintaining high win rate in competitive play is the best indication of real expertise.

Now other stats like kills and damage aren't very good data because they mean totally different things in different tiers and contexts. It is much easier to do high damage and kills for an ok player at lower tiers than it is for a good player in a high tier match. It's much easier to score good if your teammates are bad, and harder if your teammates are good, and so forth.

Competitive matches don't have higher damage/kill/whatever averages than then worst of the tier 5 scrubfest matches because the damage it takes to kill the enemy remains about the same, so looking at those two matches there is no way of telling that one is full of really good players and the other is full of really bad players. Therefore damage and kills and any other in game performance stat can't be used to measure skill.

Now it's true that group queue win percentage can be "inflated" by consistently running with a team of players that are a lot better than yourself, I'm not convinced this is a big factor though. Planning for a match by coordination of mechs and such things is a part of your skillset so there is nothing really wrong with having the ranking system reward you for that.

The problem with MWO's current ranking system isn't that it measures the wrong things, but rather that there is no zero sum qualifier regulating how many players can be in each tier. This means everyone can be tier 1 in theory, the "XP bar" problem. It would be much better if tier 1 was capped at for example the best performing 10% of the players, tier 2 the next 15% and so on, if people improve to surpass you it should be possible that you are pushed down in tiers. That is a huge problem, but measuring skill mainly by winning is the correct way and not something that should change.

Edited by Sjorpha, 14 June 2017 - 11:21 AM.

[MischiefSC]

G4LV4TR0N, on 14 June 2017 - 11:06 AM, said:

And then you will have to split it again because 2-man groups are so different than 12-man.

Group queue stats are always going to be iffy. Pugging works for an Eli concept because everyone is swimming in the same sea, so to speak. Any time you're consistently paired with others and on top of that have varying tonnage restrictions you're looking at a far less trustworthy sample.

[Nameless King]

I would like them to add in a KMDD stat.

[Spr1ggan]

MischiefSC, on 14 June 2017 - 10:28 AM, said:

W/L is accurate beyond about 50 matches. Overall is playing with you more likely to result in a win or a loss? You're 8% of your teams performance, how much you carry with that is all you.

I'm in favor of splitting pug and group queue stats though.

That's fine by me, haven't played a single group queue match since coming back and then you have dudes with K/D's and W/L's going from 10 all the way upto 40 in some cases xD

Edited by Spr1ggan, 14 June 2017 - 01:14 PM.

[MischiefSC]

Spr1ggan, on 14 June 2017 - 01:14 PM, said:

That's fine by me, haven't played a single group queue match since coming back and then you have dudes with K/D's and W/L's going from 10 all the way upto 40 in some cases xD

Usually group queue stuff.

I'm all for splitting that out.