Jump to content

The Math Behind The Tournament Rank


39 replies to this topic

#1 Red squirrel

    Member

  • PipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 1,626 posts

Posted 22 February 2013 - 04:00 PM

I'm just a curious squirrel, so I came up with this.

The Score in the Light/Medium/Heavy/Assault rank depends on a lot of unknown variables.
What I could determine from the leaderboard is that you can get between 1 and 2 points per match. At the time I checked the numbers there are a few players that got exactly 1 or 2 points per match. All others were in between.

So what does that tell us?
First I believe this means that if you are a super crappy player but play much more games than everyone else you end up on top of the board. But then who cares about that.

So what interesting facts can we get from this?
It'ts very interesting to compare how many points the different weight classes get per match.
So I calculated average number of points per match - 1 ( -1 because 1 seems to be the minimum score per match).

Et voilà:
Assaults: 0.384
Heavies: 0.309
Mediums: 0.324
Lights: 0.681

So we see that the 3 heavier weight classes perform quite similar with Assaults getting some more points. Interesting is also that Mediums seem to generate more points than Heavies but not really much.
But then the lights! wow they seem to get a twice as good rating per game.
That menas more assists more kills less deaths. For me this is an indicator for ECM Raven madness.

okay going to sleep now .... getting slowly crazy posting stuff like this.

Edited by Red squirrel, 23 February 2013 - 01:13 AM.


#2 Vulix

    Member

  • PipPipPipPipPipPipPip
  • 712 posts
  • LocationSouthwest USA

Posted 22 February 2013 - 04:03 PM

I wish PGI would admit there is a balance problem with Raven-3L's, and I say this as someone who uses light mechs almost exclusively, including the 3L

#3 Red squirrel

    Member

  • PipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 1,626 posts

Posted 22 February 2013 - 04:04 PM

We should see how those numbers change with knockdown enabled.

(...still not asleep :P - squirrels never sleep)

Edited by Red squirrel, 22 February 2013 - 04:05 PM.


#4 Misfit73

    Member

  • Pip
  • 18 posts

Posted 22 February 2013 - 04:11 PM

I hope there is more into the ranking than powergaming, otherwise it will be ruled by shared accounts.

#5 ERescue

    Member

  • PipPipPipPipPipPip
  • Philanthropist
  • Philanthropist
  • 203 posts

Posted 22 February 2013 - 04:16 PM

Sorry, you will have to redo your math... It is entirely possible to get zero points from a match. I know this, because I currently have less points than played matches in light and heavy categories.

#6 Red squirrel

    Member

  • PipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 1,626 posts

Posted 22 February 2013 - 04:20 PM

View PostMisfit73, on 22 February 2013 - 04:11 PM, said:

I hope there is more into the ranking than powergaming, otherwise it will be ruled by shared accounts.


I think this event is more about getting everyone to drop as a lone wolf in order to balance the elo system.


For me it is much more interesting to get an idea how balanced the weight classes are.
And as I (and many others) expected lights seem pretty overpowered.
Of course in the end we dont exactly know how the score is calculated but the tendency is obvious.

1) Heavies and Mediums seem very balanced
2) ECM SSRM lights rule the field.
3) Assaults are a bit stronger than the 40-75ton mechs (probably a mixture of D-DC ECMs and difficult to hit Stalker CT hitboxes)


Would be nice to see scores for different mechs - or to get a post from one of the pilots on top of the list what they've been playing.

View PostERescue, on 22 February 2013 - 04:16 PM, said:

Sorry, you will have to redo your math... It is entirely possible to get zero points from a match. I know this, because I currently have less points than played matches in light and heavy categories.



It's just an approximation based on the public leaderboard - thanks for the input.
But at the end of the day the picture is the same if you assume 0-3 points per match.
The light chassis generates more points.

#7 J4ckInthebox

    Member

  • PipPipPipPipPipPipPip
  • 832 posts
  • LocationBritanny, France

Posted 22 February 2013 - 04:21 PM

tldr lol, but it is spelled "et voilà", not "E voilá"

#8 Pat Kell

    Member

  • PipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 1,187 posts
  • LocationSol, NA, Iowa

Posted 22 February 2013 - 04:31 PM

I was in the lead for a bit on the all around leaderboard and I was rotating between a raven 3l (2 streaks, 3 med laser and ecm), a centurian AL(2 lrm 15, 1 med laser and tag) a catapult k-2 (2 ppc,1 erppc) and a d-dc (2 ppc,guass and 3 streaks and ecm). but I am or have already lost the lead cuz I need to take a break. I think scores do depend a lot on how many games you play rather then an average of how good you do which would be preferable but its ok, I got my 15 mins of fame hehe. I may be wrong about how the scores are figured...not real sure now that I read more but any way its fun so I will play when i can.

Edited by Pat Kell, 22 February 2013 - 04:39 PM.


#9 Bryan Ekman

    Creative Director

  • Developer
  • Developer
  • 1,106 posts
  • Twitter: Link
  • LocationVancouver, BC

Posted 22 February 2013 - 04:33 PM

View PostVulix, on 22 February 2013 - 04:03 PM, said:

I wish PGI would admit there is a balance problem with Raven-3L's, and I say this as someone who uses light mechs almost exclusively, including the 3L


Before we do anything to the 3Ls we are waiting for state rewind to come in. 3L's will be a lot less desirable when you can't hide behind lag shields.

#10 Vulix

    Member

  • PipPipPipPipPipPipPip
  • 712 posts
  • LocationSouthwest USA

Posted 22 February 2013 - 04:35 PM

View PostBryan Ekman, on 22 February 2013 - 04:33 PM, said:


Before we do anything to the 3Ls we are waiting for state rewind to come in. 3L's will be a lot less desirable when you can hide behind lag shields.


I think the problem is more so that 1v1 against any other light a 3L will almost always win

#11 Carnivoris

    Member

  • PipPipPipPipPipPip
  • 463 posts

Posted 22 February 2013 - 04:39 PM

Can someone explain to me how the numbers even matter? It seems to just add all your points together, rather than taking an average. That doesn't mean you're good. That just means you've got the free time to play tons and tons of matches in a given day. It's a thinly-veiled attempt at getting people to play more to get data in for ELO.

#12 Vulix

    Member

  • PipPipPipPipPipPipPip
  • 712 posts
  • LocationSouthwest USA

Posted 22 February 2013 - 04:41 PM

View PostCarnivoris, on 22 February 2013 - 04:39 PM, said:

Can someone explain to me how the numbers even matter? It seems to just add all your points together, rather than taking an average. That doesn't mean you're good. That just means you've got the free time to play tons and tons of matches in a given day. It's a thinly-veiled attempt at getting people to play more to get data in for ELO.


Pretty much, I dropped out of trying to win the tournament when I realized I have work to do this weekend

#13 HighlandCoo

    Member

  • PipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 772 posts
  • Locationbehind you

Posted 22 February 2013 - 04:43 PM

View PostBryan Ekman, on 22 February 2013 - 04:33 PM, said:


Before we do anything to the 3Ls we are waiting for state rewind to come in. 3L's will be a lot less desirable when you can't hide behind lag shields.



Interesting. Ive been hitting commandos with SRMs and AC lately (something that was impossible a couple of months ago).. so the Raven still has hitboxing issues that other mechs do not?

it never even occured to me that different mechs would have varying levels of "lagshield" regardless of their speed.

#14 Noth

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Infernal
  • The Infernal
  • 4,762 posts

Posted 22 February 2013 - 04:43 PM

View PostBryan Ekman, on 22 February 2013 - 04:33 PM, said:


Before we do anything to the 3Ls we are waiting for state rewind to come in. 3L's will be a lot less desirable when you can't hide behind lag shields.


I think it is more the fact that it mounts both ECM and Streaks. The other ravens are not nearly as good as that despite lag shield. Jenners are less desirable despite packing more of a punch simply because they do not mount ECM. Lagshield has very little to do with why the 3L is so good. If it was lagshield, all lights would be being called OP often. Instead it is just the 3L and occasionally the Com2D.

#15 TB Freelancer

    Member

  • PipPipPipPipPipPipPip
  • Shredder
  • Shredder
  • 783 posts
  • LocationOttawa

Posted 22 February 2013 - 04:47 PM

View PostVulix, on 22 February 2013 - 04:03 PM, said:


I wish PGI would admit there is a balance problem with Raven-3L's, and I say this as someone who uses light mechs almost exclusively, including the 3L


They're annoying but I have little trouble taking the legs right out from under them as it is. Until a few things like state rewind, more server optimizations, knock down and general performance tweaks happen, I don't see any point at all in buffing or nerfing them. Too many of those things directly contribute to their present success....

....much like playing premade while PGI was collecting seed data for Elo, a lot of them are crying now. I expect to see a lot of lights crying when state rewind, knockdown and other general performance tweaks make them almost effortless to kill.



EDIT: Didn't notice this post. But it pretty much reinforces what I just wrote.


View PostBryan Ekman, on 22 February 2013 - 04:33 PM, said:


Before we do anything to the 3Ls we are waiting for state rewind to come in. 3L's will be a lot less desirable when you can't hide behind lag shields.

Edited by TB Freelancer, 22 February 2013 - 04:48 PM.


#16 Noth

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Infernal
  • The Infernal
  • 4,762 posts

Posted 22 February 2013 - 04:50 PM

View PostTB Freelancer, on 22 February 2013 - 04:47 PM, said:

[size=4]

They're annoying but I have little trouble taking the legs right out from under them as it is. Until a few things like state rewind, more server optimizations, knock down and general performance tweaks happen, I don't see any point at all in buffing or nerfing them. Too many of those things directly contribute to their present success....

....much like playing premade while PGI was collecting seed data for Elo, a lot of them are crying now. I expect to see a lot of lights crying when state rewind, knockdown and other general performance tweaks make them almost effortless to kill.



EDIT: Didn't notice this post. But it pretty much reinforces what I just wrote.


Why is the 3L so good when the other lights aren't? It isn't because the 3L gets anymore lag shield than the other lights. If all of a class are getting effected by lag shield, and one of those stands out as completely better than the rest, it is not lag shield that is the problem with them being too strong.

Edited by Noth, 22 February 2013 - 04:51 PM.


#17 Vulix

    Member

  • PipPipPipPipPipPipPip
  • 712 posts
  • LocationSouthwest USA

Posted 22 February 2013 - 04:54 PM

If I pilot a Raven 3L and encounter another light mech, one of two scenarios happens:

If the enemy light mech has ECM, I such to counter and use my Streaks and energy weapons to circle strafe my opponent to death. Spiders have no missiles and are thus screwed by my streaks, while commandos have less armor and will die quicker. My only equal fight is with another ECM Raven.

If my enemy has no ECM, like a Jenner, I just keep my ECM to shut down their streaks and dance around to avoid their other weapons while I launch auto-aimed missiles at them.

It's not a lag issue, it's the Raven-3L itself

#18 Lukoi Banacek

    Member

  • PipPipPipPipPipPipPipPipPip
  • WC 2018 Top 12 Qualifier
  • WC 2018 Top 12 Qualifier
  • 4,353 posts

Posted 22 February 2013 - 05:38 PM

View PostERescue, on 22 February 2013 - 04:16 PM, said:

Sorry, you will have to redo your math... It is entirely possible to get zero points from a match. I know this, because I currently have less points than played matches in light and heavy categories.


How do you see your points?

#19 Lukoi Banacek

    Member

  • PipPipPipPipPipPipPipPipPip
  • WC 2018 Top 12 Qualifier
  • WC 2018 Top 12 Qualifier
  • 4,353 posts

Posted 22 February 2013 - 06:00 PM

Never mind...it just added me to the queue :)

#20 DEN_Ninja

    Member

  • PipPipPipPipPipPipPipPip
  • The Blade
  • The Blade
  • 1,097 posts
  • LocationCrossing, Draconis March

Posted 22 February 2013 - 06:05 PM

I lost a game and got 2 points. I don't understand.





3 user(s) are reading this topic

0 members, 3 guests, 0 anonymous users