With allegskill...

[MrChaos]

MW says: L33T!!!! L333T!!! I'm L33T I don't need no stinking math .

I guess I was a little rude to you MW (and Buyo), my bad. I'm just a wee bit weary of people using their guts and feelings to argue the point. Your welcome to your view since it's yours, and no hard feelings.



Pico that is a really BAD way to do ranks, however it is a wonderful way to perform behavior modification on the community. Remember bombers and what happened when the points changed for base kills?

[Malicious Wraith]

QUOTE (MrChaos @ Jun 17 2009, 01:58 PM) MW says: L33T!!!! L333T!!! I'm L33T I don't need no stinking math .

I guess I was a little rude to you MW (and Buyo), my bad. I'm just a wee bit weary of people using their guts and feelings to argue the point. Your welcome to your view since it's yours, and no hard feelings.



Pico that is a really BAD way to do ranks, however it is a wonderful way to perform behavior modification on the community. Remember bombers and what happened when the points changed for base kills?

How is my absolute understanding that the "Math" that you utilize to rank us based purely on our win/loss output "Guts and Feelings"?

No, I fully understand the level accuracy at which AS has succeeded in attaining, but I also understanding that its accuracy is limited purely to a rigid factor of winning or losing. The fact of the matter is that in Allegiance with teams numbering over twenty during most prime time games, your individual contribution to the team does not effect the only factor that is graded, loss or victory.

A much, vastly disproportionate, factor in your win/loss ratio has to do with the teams that you choose or fly with, rather than your own skill, or performance.

These are not "Guts or Feelings", these are inalienable facts that you are not able to factor into your skill system, which is why it is ultimately flawed. That is also the premise behind my request to enforce AB, as that is the only way to remove the human stack factor from the equation.

[sgt_baker]

QUOTE (Picobozo @ Jun 17 2009, 06:22 PM) Has anybody involved with the Ranking development ever considered using some or most of what MS originally did with AZ rankings? It was based purely on individual statistics. I know there are a lot of people who want to involve wins and losses and hair color or whatever else it is your basing these ranks on. But, in my opinion the best way to rank individuals is by their own stats.

Yes it leaves out people who like to scout (H_Mallow has been mentioned a lot here and deservedly so) so maybe concentrate on developing more recorded statistics based on support activities. It seems like all this hard work is going into mega-complicated math when it might ultimately be going in the wrong direction.

Just my 2 cents, don't want to put down Baker's hard work because I know he has good intentions and has innovated greatly with the stack rating. Though I did read what you posted Baker about your system not being 100% implemented. So I guess my other question is: When your stats system is 100% implemented, do you see a big change to the leaderboard?

For startes: Using something like MS's old system leads to a problem which is similar to that with curent core developement: How do we decide how many points to attribute to a certain in-game action? Then, when we've invariably given whores more points than nans (by a long shot), how do we correct for the error? Do we simply change the weights and factors from ground zero or do we recalulate all past games? If past cores weighted bombing over whoring, how do we account for that?

You can see how you suddenly become embroiled in a mess of numbers which have an extremely indestict eventual outcome. Ideally, there'd be a system which could account fairly for every single thing we did in-game. Currently we've got close to (if not more than) a million player-games in the ASGS database. Now imagine that every player performs 100 game-affecting actions in any given game. Suddenly you're faced with 100,000,000 events which you're charged to make sense of.

It'd require somebody to understand the intricate relationships between all of those actions before they could even begin to generate a system whereby decent rankings were generated.

Why is this a lost cause? The exact same problem is encountered by the core development folks. To date, they've been attempting to correct for unwanted effects given 100,000,000 inputs (and that's conservative). STILL to this day we don't have a core that people believe to be truly balanced.

The exact same problem would persist with a points-based ranking system.

Further to this, (I had a phonecall and have completely lost my thread.)

Another time maybe.

[Malicious Wraith]

Again, I reiterate that enforced AB designed by the AS team should be enforced, and I have the up most faith that we will see more balanced PUG's and more enjoyable games due to that.

[sgt_baker]

Actually, MW, your contribution to the game is inversely proportional to the number of players on your team.

[sgt_baker]

In that case: Hassle us (MrC and I) to simulate YOU LOT in order to get a decent post-launch AB system up and running. It's not supremely difficult, just far more difficult to model humans than it is to follow basic maths.

[takingarms1]

I find it amusing that geniuses have trouble with simple grammar

[Malicious Wraith]

Baker, do you, or do you not, deny that choosing your team is more important than your individual skill when it comes down to winning in Alleg.

With the assumption that Allegskill only ranks you according to your ability to win or lose, and you agree that the above is true, through syllogism, we can deduce that:

The ability to choose the best team is the more important than skill when determining your Allegskill rank.

[Picobozo]

Mr Chaos, I know that MS ranking wasn't perfect, but it did a pretty good job of showing you who was skilled. I'm sure that if MS supported the game for longer then a year we would have had many iterations of a ranking system that would have hopefully addressed the balance issues. Current developers have the opportunity to modify these stats with some of the work they have already done to help account for issues (like time played, or events performed per hour) to help get a more accurate ranking.

QUOTE (sgt_baker @ Jun 17 2009, 01:09 PM) For startes: Using something like MS's old system leads to a problem which is similar to that with curent core developement: How do we decide how many points to attribute to a certain in-game action? Then, when we've invariably given whores more points than nans (by a long shot), how do we correct for the error? Do we simply change the weights and factors from ground zero or do we recalulate all past games? If past cores weighted bombing over whoring, how do we account for that?

You can see how you suddenly become embroiled in a mess of numbers which have an extremely indestict eventual outcome. Ideally, there'd be a system which could account fairly for every single thing we did in-game. Currently we've got close to (if not more than) a million player-games in the ASGS database. Now imagine that every player performs 100 game-affecting actions in any given game. Suddenly you're faced with 100,000,000 events which you're charged to make sense of.

It'd require somebody to understand the intricate relationships between all of those actions before they could even begin to generate a system whereby decent rankings were generated.

Why is this a lost cause? The exact same problem is encountered by the core development folks. To date, they've been attempting to correct for unwanted effects given 100,000,000 inputs (and that's conservative). STILL to this day we don't have a core that people believe to be truly balanced.

The exact same problem would persist with a points-based ranking system.

Further to this, (I had a phonecall and have completely lost my thread.)

Another time maybe.

I'm not sure it has to be so complicated where we have to psycho analyze all 100,000,000 events.

I think putting support activities like keeping miners eyed, deprobing, probing and assigning them point values close to or above kill points would put you on the right track. I know that to truly value a support player properly you would need code changes to install hooks to capture events like I mentioned, but how out of the realm of possibility is it to do that? XboxLive has an achievement system that accounts for things that are not necessarily part of core gameplay. I would imagine that instead of giving somebody a one time achievement for keeping a miner eyed or deprobing a sector, you keep a running tally on a database and assign points for it.

Basically, don't search for perfection of ranking the first time out. Use broadstrokes and value support activities at or above what kills/bases killed-captured are currently valued and see what happens.

Look, I know I don't have much of a leg to stand on because i'm not a coder and I haven't donated much time to the issue. I also hate people criticizing me for things that i've done to try to help. So i'm gonna stop my input here. I'm not trying to down anybody for their hard work and i'm absolutely appreciative for the ranking system that you've implemented. I was just trying to offer a new perspective that I think could work.

And like you said earlier your ranking system isn't 100% implemented which I was not aware of before this thread. Hopefully we'll see a big change when it does get implemented fully.

[sgt_baker]

Negative.

AS is already aware of who has the best team. Joining that team and winning will only negate the effect of the win to a large extent. Losing whilst on that team, on the other hand, will bork your rank in ways which you will find displeasing.

Are you lot in total ignorance of what I did in-game between AS being approved and it actually going live? Ask around. Ranked 15th in the community. Blah Blah Blah.

The only reason I did that was to demonstrate what heppens under Helo when you constantly join the better team, then to further reveal what happens when AS gets its teeth into you.