We're gonna die!

grarrgrarr · Post by **grarrgrarr** » Tue Jul 12, 2011 2:29 pm

btw TheCorsair, i play early morning now because it makes sense to sleep early and get up early, and thats where the action is at anyway.

Dibbley2 · Post by **Dibbley2** » Tue Jul 12, 2011 3:12 pm

Afk

Allegiance dead through the day ?

MrChaos · Post by **MrChaos** » Tue Jul 12, 2011 3:16 pm

Hi Ghost

congrats on your job.

All discussed already, all implemented to some degree talent available

@help was implemented to teach newbs but has basically died
was a member and found that most don't want help, few listened, and.I've got a job

Ranks phase I,
Worked on the beta and can tell you the intention then was to tweak once implemented... did not occur
...
Autobalance phase II
Worked on beta. Never implanted do to how to handle post launch balance and the little engine that could saying *meh*

Points for tasks phase III
Never started . Allegiance did ranking this way in the beginning. Huge balance issues from point scumming btw, people only doing things to max points. Oh! They @#(! canned it and implemented the current Allegiance ranking system for XBox Live. People want achievements tba but ranks are meant to rank not grow your epeen for time served

Servers give poor connections to client's
Server owner for R4 and R5. There was profound change in server performance around the release of R5. R5 dinked with connections. The server went from few complaints to a lynch mob outside my door with R5. Same story all servers, and all owners. Btw four people had access toy server trying to find the issue no one could. Hell Cable does it for living and to the best of my knowledge left Allegiance after owning a server. That's a $#@!ing no kidding shame

On my phone so your reply scrolled off. 'll stop there other then to say:

See bugs are bugs, fubared connection speeds are fubared connection speeds, and the same tired suggestions make Count cranky.

I'm frustrated and bummed out trying to light a torch under someone so don't mind me

Hugs for caring enough to bitch at WITH me . Also thanks for the shout out.

edit: phone typing make me even less understandable
MrChaos

Post by **pkk** » Tue Jul 12, 2011 3:48 pm

Ranks:
ATM Alleg just knows the rank of the player, it's unable to get Mu and Sigma from ASGS database. With release of CSS, work on autoblance can continue, CSS beta already contains that interface.

Server connections:
I personally haven't found a proof yet, that change of netcode of R5 produces worser connections. What I saw is that routing of packets to datacenters got worse (Planet (2005): ~180 ms, Planet (today): ~200 ms). Today we have tons of more traffic on the net than years ago. People always complained playing on GPZ, Planet or Stripe. Some things will never change.

Server owners can't do much to improve the performance of server in datacenters. The only way is to go to another datacenter, see GPZ/DocMach/Mach2.

Bard · Post by **Bard** » Tue Jul 12, 2011 3:59 pm

KGJV wrote:QUOTE (KGJV @ Jul 9 2011, 11:39 AM) I see that our discussion about the end of Allegiance as less interest for your than that picture...

nm j/k (or not?! ). try using 'search by image' in Google Image Search (click the little camera icon in the search box and paste the 11-11-11 link), you'll find a 1600x1200 version somewhere. (or use TinEye but Google is faster and better).

Ummm.

That mostly turns up @#(!ty movie posters.

You might want to plop a URL into your profile or signature or somewhere equally reachable if you want people to notice. You got scooped by Hollywood.

MrChaos · Post by **MrChaos** » Tue Jul 12, 2011 5:07 pm

It is the spikes that drive me bat @#(!. People clear out many times during the switch from EU to NA prime time during he arguing to switch servers. I'm not bashing you or the server I hope you know plus you are waaaay more qualified. Unsure maybe its coincidence. and a huge up tick in traffic about 1.5 years ago. Be that as it may the poor connections need to be addressed.

Other item, I know the details figured it was tmfi. There was a work around, well I think there was one. It was related to autobalance not tweaking ranks

Thanks for the info and reeling me back in

Spunkmeyer · Post by **Spunkmeyer** » Tue Jul 12, 2011 11:24 pm

pkk wrote:QUOTE (pkk @ Jul 12 2011, 10:48 AM) Ranks:
ATM Alleg just knows the rank of the player, it's unable to get Mu and Sigma from ASGS database. With release of CSS, work on autoblance can continue, CSS beta already contains that interface.

That's an improvement, but how is that going to help anything in any significant way? Are you talking about post-start autobalancing? I'm not even aware of the core issues surrounding that (as in, what's the big show-stopper?)

There are four issues with the whole ranking and autobalance issue that I can identify, off the top of my head:

1-The big improvement with autobalance will come when commanders are ranked and commander autobalance is enforced, and only games where commander autobalance is enforced count towards ranking of players.

What we have is garbage in, garbage out. We don't have a way to quantify the influence of the commander on the outcome, so we ignore it. The result is bogus changes in rank. It won't average out either since stacking means some players will always choose the better comm to fly for, and those who don't know any better ("I just wanna play") usually end up with the @#(!tier com.

2-The convergence needs to change to a running average. Right now any mistakes early on take an excessive amount of time to compensate for and the rank will not follow the increase (or decrease) in skill.

3-It should not be possible to (easily) game the system. Currently, the whole situation with duration in game being proportional to rank change makes it very easy for an observant player to drop out or hop in and reap maximum benefits, if they can identify the outcome, especially if the rank totals are close. So you can see how this is exacerbated by rank being GIGO in the first place.

4-The final autobalance system should NEVER, EVER stop anyone from playing on a public server. You should always be able to join a team, and commanders should not be able to stop you. This is the only way to AB acceptance by the community.

MrChaos · Post by **MrChaos** » Wed Jul 13, 2011 2:35 am

Spunkmeyer wrote:QUOTE (Spunkmeyer @ Jul 12 2011, 05:24 PM) That's an improvement, but wow is that going to help anything in any significant way? Are you talking about post-start autobalancing? I'm not even aware of the core issues surrounding that (as in, what's the big show-stopper?)

There are four issues with the whole ranking and autobalance issue that I can identify, off the top of my head:

1-The big improvement with autobalance will come when commanders are ranked and commander autobalance is enforced, and only games where commander autobalance is enforced count towards ranking of players.

What we have is garbage in, garbage out. We don't have a way to quantify the influence of the commander on the outcome, so we ignore it. The result is bogus changes in rank. It won't average out either since stacking means some players will always choose the better comm to fly for, and those who don't know any better ("I just wanna play") usually end up with the @#(!tier com.

2-The convergence needs to change to a running average. Right now any mistakes early on take an excessive amount of time to compensate for and the rank will not follow the increase (or decrease) in skill.

3-It should not be possible to (easily) game the system. Currently, the whole situation with duration in game being proportional to rank change makes it very easy for an observant player to drop out or hop in and reap maximum benefits, if they can identify the outcome, especially if the rank totals are close. So you can see how this is exacerbated by rank being GIGO in the first place.

4-The final autobalance system should NEVER, EVER stop anyone from playing on a public server. You should always be able to join a team, and commanders should not be able to stop you. This is the only way to AB acceptance by the community.

1 - Surprisingly the answer is: Nope. The evidence, and there are 10,000s of games of it, shows commanders, as a rule, mean $#@! all. Sure a zero against spidey is going to lose almost ever time but balanced teams and sane commander match ups means they are a non-factor. I noes you noes better and all but at the end of the day Spunk statistics and probability >>>>>>>>>>>>>>>>> Spunk's gut (btw this was the hardest thing for me too that I had to swallow while running the numbers). If you compare autobalance of teams vs commander (again sane commander choice) then the team is even more of a factor then Spunk's gut. The community can police itself from Jimmy vs noob com stacking is my position... and if they can't safe guards can be put in place. It will make Shiz angry but it could be done. Btw Spunk there are commander ranks... just scroll the leaderboard to the right

2 - If you mean last xx,xxx games count and they scroll against time you may have a valid point, my knee jerk reaction is I agree. Again ranking is about ranking and not about inflating someone's epeen for time served. That's what achievements *I heard the Xbox noise for completing an achievement in my head just now... scary* are for, to make thier epeen feel better. In principle we agree, there could be some tweaking done, especially for the long time players who's body of work weighs them down vs a newer player... although the algorithm says *meh* doesn't matter.... maybe... I'm simply not sure until I give it a whirl. Stats make you look silly unless you give them their due.

Honestly I just can't be assed to do the work (which starts at square one with getting the data again from TE. I deleted it all, as promised, when my work was finished on the project) given the bigger fish that need frying. There are mechanisms in place for AS to turn up the uncertainity fwiw. The down side is longer convergence but also allows those with more games to move to a different rank quicker. Its not something one should mess with it without some work first. Do I think it needs a tweak... yes, do I now what it is atm, nope.

3 - Right and until you FORCE autobalance on people it makes it harder for the system to converge. It will converge and be about as accurate as it can given all of the other factors fubarring things.

4 - Well not exactly in my book. It should always stop you from playing THAT particular game if it will cause an imbalance but it should never stop you from playing all together. See the problem is given the teeny, tiny player base there is only one game. Not insurmountable for clever folks, just needs to be done is all. It tends to skew things if you just start throwing crap at the wall by popular concensus.

Nice post Spunk

MrChaos

Spunkmeyer · Post by **Spunkmeyer** » Wed Jul 13, 2011 4:35 am

There is the whole correlation vs causation issue. Intuitively, a high ranking group of players (by definition of ranking, consistent winners) will generally not be on the same team with a weaker commander. That doesn't mean the commander is irrelevant, it just means AS is blind to the effect of the commander.

The overriding issue is the rank needs to be a better indication of "skill" as opposed to "the ability to win" so we can do away with things like sticking with certain commanders, stacking etc. that are very closely tied to "the ability to win".

MrChaos · Post by **MrChaos** » Wed Jul 13, 2011 5:38 am

Spunkmeyer wrote:QUOTE (Spunkmeyer @ Jul 12 2011, 11:35 PM) There is the whole correlation vs causation issue. Intuitively, a high ranking group of players (by definition of ranking, consistent winners) will generally not be on the same team with a weaker commander. That doesn't mean the commander is irrelevant, it just means AS is blind to the effect of the commander.

The overriding issue is the rank needs to be a better indication of "skill" as opposed to "the ability to win" so we can do away with things like sticking with certain commanders, stacking etc. that are very closely tied to "the ability to win".

Oops now you've stepped in it. This is your attempt to stack the deck to your position by using words without proof. If I allow you to do as you please, you most certainly can make stuff break and why your point about autbalance being mandatory makes sense. Is a drinking glass made of glass a poor design because your two year old smashes it on the ground then rolls around in it. Yes if we only look at from that perspective and discount what a poor parent you are for allowing them to have it in the first place. The world is big enough for Mommie's wine glass and junior's plastic sippy cups.

I know this reply infuriates guys like Cashto who think they are smarter then me by glancing at the endless links Ive posted on the subject but... Same answers as before and your opinion is backed by nothing but your gut without numbers. * points to TE* Get the data from him, then check th math, find the error, provide a better solution, then we can talk. Until then I sincerely hope you read all of my words to you cause I'm not dismissing your comments ,simply telling you it isn't worth the debate until you understand the math... no offense intended.

Can we still be friends?
MrChaos