27 replies to this topic

[Nightcrept]

I like stats.

And since I don't think the current MWO stat system is very good I decided to see if I could not only think of the various relevant stats but to create a sort of pilot rating formula.

Thus far this is the relevant stats I started with before deciding on those needed for my preliminary power rating system.

KIlls-
Kills per game-
Deaths-
Deaths per game-
KDR-
Avg dmg dealt per game-
Disconnects-
Disconnect ratio-
Assists-
Assists per game-
Wins-
loses-
Win/loss ratio-
Avg tonnage lost per game-
Avg tonnage killed per game-
Games played-

If you can think of any others please let me know.

Now on to my initial formula. Mind you it is a work in progress and I will be working on it more in the next few days. If anyone can play with their own or modify mine etc feel free to post away.

I ended up using

Avg Dmg per game- ADpG
Avg tons lost per game- ATLpG
Avg Assists per game- AApG
Disconnect Avg- DA
Avg Kills per game- AKpG


ADpG/ATLpG+(AApG-DA*AKpG) = rating

It is simple but it gives a good indicator to me.

So for me today I played 6 games with the following stats.


Avg Dmg per game- ADpG =388.5
Avg tons lost per game- ATLpG =70
Avg Assists per game- AApG =2
Disconnect Avg- DA =0
Avg Kills per game- AKpG =2


So using my formula.
ADpG/ATLpG+(AApG-DA*AKpG) = rating

388.5/70 + (2-0*2)= 9.55

So my pug player rating would have been 9.55.

Now lets pretend that I dc'd twice in those six games for a avg of .3 dc's. My rating would then change to.

388.5/70 + (2-.3*2)= 8.9

So my DC rate would indeed change the rating but not so much as to prevent me from being a competitive player.

This formula is only viable for pug players. Any games played in a pre-made of any size should only be compared using team stats against equally sized teams.

Please discuss.

[superteds]

no stats. stats in approximately every game ever turn people into K:D obsessed, risk-adverse morons.

death2allstats

well no public stats. a hidden ELO rating is fine.

Edited by superteds, 07 January 2013 - 02:58 PM.

[Orzorn]

That's a good idea. I especially like assists per game, as that would help me, as a support styled player, know how many fights I actually got into.

Also, my power rating is over 9000.

[Nightcrept]

Orzorn, on 07 January 2013 - 02:58 PM, said:

That's a good idea. I especially like assists per game, as that would help me, as a support styled player, know how many fights I actually got into.

Also, my power rating is over 9000.

How in the world did u get a power rating off 9000?

[Nightcrept]

Yeah max rating should be around 80 depending on how they adjust the matchmaker for team tonnage.The avg should be around 5.00 to 6.50 for a pug.

Edited by Nightcrept, 07 January 2013 - 03:14 PM.

[Roland]

All I know is that mine is over 9000.

[Nightcrept]

superteds, on 07 January 2013 - 02:57 PM, said:

no stats. stats in approximately every game ever turn people into K:D obsessed, risk-adverse morons.

death2allstats

well no public stats. a hidden ELO rating is fine.

Stats can be fine if it is done using a power rating that self adjusts for cheating. If you noticed kills while have an affect on your rating is not disproportionate to the other stats. So kill stealing will not give you a large benefit. Especially since the most important stat in my formula is avg damage per game as that is going to indicate the level that you help out your team the most. Games where a person disconnects should not count in any positive respect. So you get the death, the game played, the loss of tonnage and the dc but not the pros.

[Lykaon]

superteds, on 07 January 2013 - 02:57 PM, said:

no stats. stats in approximately every game ever turn people into K:D obsessed, risk-adverse morons.

death2allstats

well no public stats. a hidden ELO rating is fine.

Agree!

[Nightcrept]

Roland, on 07 January 2013 - 03:20 PM, said:

All I know is that mine is over 9000.

Then my formula is broke so is you don't mind lets walk through this so I can figure out where. Please fill in this info for me if you can as it's not listed by pgi. Guessing works in this case. Avg Dmg you deal per game- ADpG =Avg tons you lose per game- ATLpG =Avg Assists you have per game- AApG =Disconnect Avg (Number of disconnects/games)- DA =Avg Kills you get per game- AKpG =

Lykaon, on 07 January 2013 - 03:31 PM, said:

Agree!

This is a warfare game. There will be stats. So instead of saying you don't want any how about helping me tweek a formula that will reward good game play and cut down on the game mechanics manipulation.

[Mack1]

I think the stats should be represented by a giant *****, the more kills you get the larger the ***** grows, this would suit the stat whores much more than mere numbers.

[Nightcrept]

Since pgi is most likely to rank players by simple and easily manipulated stats you may get your wish.

I was hoping however for at least one semi-intelligent forum goer to come along so i could bounce ideas around with.

[Voidsinger]

Well, I would say that mechs mastered should come into it, except I hit the 32 mech mastery bug.

Every stat has a story, and it is just a giant epeen contest. Since the decisions that matter will often go unrewarded in spite of resulting in somewhat suicidal behaviour.

[Nightcrept]

Voidsinger, on 07 January 2013 - 03:41 PM, said:

Well, I would say that mechs mastered should come into it, except I hit the 32 mech mastery bug.

Every stat has a story, and it is just a giant epeen contest. Since the decisions that matter will often go unrewarded in spite of resulting in somewhat suicidal behaviour.

I would say in your case mechs mastered would count more towards your understanding of the game then player rating. So you might know how every mech works but not be any good at running them. Not saying that's true in your case just the stat. Mechs mastered would be something pgi hopefully would look at if they where ever looking for in community testers.But I do agree with you. It is very hard to get a ranking system in a game like this that can properly gauge players ability.

Edited by Nightcrept, 07 January 2013 - 03:47 PM.

[Nightcrept]

The average damage you deal per game.
Divided by the amount of average tonnage you lose per game.

Then added to the sum of your

Average assists per game minus your average disconnects per game multiplied by your kill average per game.

So your average damage is say 800 per game.
And you average tonnage lost is 35tons per game.

800/35= 22.86


Then you take your average assists per game say 4.
And subtract your average disconnects say 0.
And multiply your average kills per game say 4. (pro tip...the two averages or assists and kills together cannot be more then 8)

So 4-0 times 4= 16

So the end result is 22.86+16=38.86

Mind you after the lag shield and ecm issues are fixed if your a light mech and getting 800 dmg-4 assists and for kills per game then you deserve a score of around 39 compared to my 9.6 in a illya.

[Caviel]

superteds, on 07 January 2013 - 02:57 PM, said:

no stats. stats in approximately every game ever turn people into K:D obsessed, risk-adverse morons.

death2allstats

well no public stats. a hidden ELO rating is fine.

I very much agree with this. "Gear score", "player ranking", "ELO" absolutely destroy the new/casual player experience. This sort of information quickly degrades into "You must be at least this tall to ride this ride". Yes, the information can be used positively, and yes, there are good things that can come from this data. The BAD things that occur are far more common and pervasive to make it worth making these stats public knowledge. I have not seen a community yet that hasn't degraded at the inclusion of this type of information.

Nightcrept, on 07 January 2013 - 03:35 PM, said:

Then my formula is broke so is you don't mind lets walk through this so I can figure out where.

It isn't nearly as funny when you have to explain the joke:
http://knowyourmeme....s/its-over-9000

Quote

This is a warfare game. There will be stats. So instead of saying you don't want any how about helping me tweek a formula that will reward good game play and cut down on the game mechanics manipulation.

That's the problem, there isn't one. ANY formula will be min/maxed and cause segregation of the "lower end" of the player base. For example: suggestions, such as this one, would be immediately shot down by the upper ranking community simply because you are a "n00b pilot" with a low score that doesn't know what they are doing and therefore couldn't possibly be knowledgeable or contribute in any meaningful way.

It already happens on THIS forum with just the join date!

[Voidsinger]

Nightcrept, on 07 January 2013 - 03:46 PM, said:

I would say in your case mechs mastered would count more towards your understanding of the game then player rating. So you might know how every mech works but not be any good at running them. Not saying that's true in your case just the stat. Mechs mastered would be something pgi hopefully would look at if they where ever looking for in community testers.But I do agree with you. It is very hard to get a ranking system in a game like this that can properly gauge players ability.

Amazing!

You know I'm a sucky player who just played a lot of games until I brute-forced the mechs (no, I don't disconnect for the XP). That said, a ping of 280ms makes playing at US peak times a brutal experience.

You were also spot on that understanding other mechs was precisely why I did it. Well, one of the reasons.

The others were I always prefer to be skilled up in mechs I own, and that I want to be able to never have to go back and grind XP on a mech again. I'm an old BT hand, so assault rides should be earned (I got stuck at Cataphract on the way up).

[Roland]

Nightcrept, on 07 January 2013 - 03:35 PM, said:

Then my formula is broke so is you don't mind lets walk through this so I can figure out where. Please fill in this info for me if you can as it's not listed by pgi. Guessing works in this case. Avg Dmg you deal per game- ADpG =Avg tons you lose per game- ATLpG =Avg Assists you have per game- AApG =Disconnect Avg (Number of disconnects/games)- DA =Avg Kills you get per game- AKpG =

I was just kidding, dude. It was a DBZ reference.

[Lukoi Banacek]

Nightcrept, on 07 January 2013 - 02:54 PM, said:

I like stats.

And since I don't think the current MWO stat system is very good I decided to see if I could not only think of the various relevant stats but to create a sort of pilot rating formula.

Thus far this is the relevant stats I started with before deciding on those needed for my preliminary power rating system.

KIlls-
Kills per game-
Deaths-
Deaths per game-
KDR-
Avg dmg dealt per game-
Disconnects-
Disconnect ratio-
Assists-
Assists per game-
Wins-
loses-
Win/loss ratio-
Avg tonnage lost per game-
Avg tonnage killed per game-
Games played-

If you can think of any others please let me know.

Now on to my initial formula. Mind you it is a work in progress and I will be working on it more in the next few days. If anyone can play with their own or modify mine etc feel free to post away.

I ended up using

Avg Dmg per game- ADpG
Avg tons lost per game- ATLpG
Avg Assists per game- AApG
Disconnect Avg- DA
Avg Kills per game- AKpG


ADpG/ATLpG+(AApG-DA*AKpG) = rating

It is simple but it gives a good indicator to me.

So for me today I played 6 games with the following stats.


Avg Dmg per game- ADpG =388.5
Avg tons lost per game- ATLpG =70
Avg Assists per game- AApG =2
Disconnect Avg- DA =0
Avg Kills per game- AKpG =2


So using my formula.
ADpG/ATLpG+(AApG-DA*AKpG) = rating

388.5/70 + (2-0*2)= 9.55

So my pug player rating would have been 9.55.

Now lets pretend that I dc'd twice in those six games for a avg of .3 dc's. My rating would then change to.

388.5/70 + (2-.3*2)= 8.9

So my DC rate would indeed change the rating but not so much as to prevent me from being a competitive player.

This formula is only viable for pug players. Any games played in a pre-made of any size should only be compared using team stats against equally sized teams.

Please discuss.

Interesting but arbitrary...people weigh different things differently.

[Nightcrept]

Thing is guys, rankings are coming. I'm just trying to figure out how to make it fair.

[HurlockHolmes]

I don't mind statistics, the more the merrier, so long as they are private.

Caviel, on 07 January 2013 - 04:12 PM, said:

It already happens on THIS forum with just the join date!

First of all, I have never seen this, ever.

Secondly, if somebody is going to do **** like that than they shouldn't be taken seriously.