With allegskill...
Yes. In a sense, AS is still only 50% implemented. The 'easy bit' was having ASGS handle the new system. The next, and regrettably orders of magnitude more awkward, task is to get rid of the plethora of vestigial Elo and Helo features. You've mentioned quite a few of them already:
0 - 30 ranks - Should be 0 - 50Pre-launch autobalance currently ignores 99% of AllegSkill. New algorithm developed, tested and awaiting implementation.Pre-launch AB balances all non-AFK players. This is silly. AB will be updated to only balance players who have already joined a team.
Post-launch AB is still a work in progress which, to be honest, I've neglected slightly of late. The reasoning here is that it's rather difficult to simulate human behaviour when testing the system. Does everyone just accept AB's decision? No. Do people accept generally except for when they dislike one commander? Would people be prepared to wait for another player to join such that AB could assign them proper teams as a pair? If so, for how long? Add in to this people dropping and rejoining etc and you have a wonderful soup of factors to simulate well before the system gets anywhere near a real-world test. (and about twelve other gnarly factors I can't be arsed to waffle on about here)
This leads me on to the next problem: Getting enough players to perform a real-world test will be difficult. Debugging a real-world test will be more difficult still. Releasing a borked Alleg server then realising something needs to be updated/corrected retrospectively will only cause mayhem, drama and lots of work for server admins and TE. All of these changes need to work the first time 'round. No exceptions.
As for 'ranks aren't important'. I don't know about you, but I remember the days before Elo. To say that a stack/whore/win fest was a regular occurrence is an understatement. This was readily recognised by the powers which had the community's best interests at heart. You can argue 'till you're blue in the face that ranks have no effect, but there is plentiful evidence to the contrary. The vast majority of comms in our current age do their best to maintain balanced teams. That alone is a good enough reason to have ranks.
And for the accuracy of AllegSkill - Can you correctly predict the outcome of 90% of games even when the teams are populated by players you don't even know? No? I can.
I'm sure you're all very comfortable in your own knowledge and superiority as veterans of this community. I'll leave you to consider the following image.
0 - 30 ranks - Should be 0 - 50Pre-launch autobalance currently ignores 99% of AllegSkill. New algorithm developed, tested and awaiting implementation.Pre-launch AB balances all non-AFK players. This is silly. AB will be updated to only balance players who have already joined a team.
Post-launch AB is still a work in progress which, to be honest, I've neglected slightly of late. The reasoning here is that it's rather difficult to simulate human behaviour when testing the system. Does everyone just accept AB's decision? No. Do people accept generally except for when they dislike one commander? Would people be prepared to wait for another player to join such that AB could assign them proper teams as a pair? If so, for how long? Add in to this people dropping and rejoining etc and you have a wonderful soup of factors to simulate well before the system gets anywhere near a real-world test. (and about twelve other gnarly factors I can't be arsed to waffle on about here)
This leads me on to the next problem: Getting enough players to perform a real-world test will be difficult. Debugging a real-world test will be more difficult still. Releasing a borked Alleg server then realising something needs to be updated/corrected retrospectively will only cause mayhem, drama and lots of work for server admins and TE. All of these changes need to work the first time 'round. No exceptions.
As for 'ranks aren't important'. I don't know about you, but I remember the days before Elo. To say that a stack/whore/win fest was a regular occurrence is an understatement. This was readily recognised by the powers which had the community's best interests at heart. You can argue 'till you're blue in the face that ranks have no effect, but there is plentiful evidence to the contrary. The vast majority of comms in our current age do their best to maintain balanced teams. That alone is a good enough reason to have ranks.
And for the accuracy of AllegSkill - Can you correctly predict the outcome of 90% of games even when the teams are populated by players you don't even know? No? I can.
I'm sure you're all very comfortable in your own knowledge and superiority as veterans of this community. I'll leave you to consider the following image.
Granary Sergeant Baker - Special Bread Service (Wurf - 13th Oct 2011)
-
takingarms1
- Posts: 3052
- Joined: Fri Feb 20, 2004 8:00 am
Well then AB is utter crap atm.
If I was smarter and liked Baker more I'd get on TS with him and work on the implementation of bits and pieces of newer AS features. Phewwww good thing I'm not the first, as for the second in spite of myself I do, must be that lovely Serbian accent of his
.
There will always be those who go on gut and feelings when numbers are the way to tackle an issue. It's why casinos exist and government policy is generally criminally asinine when trying to regulate science based things.
If somehow the ranks got implemented incorrectly then we would have a talking point. The fact that a few vets, like myself, have really low ranks has to do with an individual's in game behavior over a span of time AND the fact your skill compared to the community in general is across a range of values across a confidence interval of *insert number of your choice which effects the span*. In my case you remove the 15W vs 120L (or some such) span of game play where I was anti-stacking every game in an effort to encourage fair game play and you'll get a more realistic rank IMHO. *shrug* I wasn't good enough to change the outcome, obviously and I pay the price via my ego [I honestly don't give a $#@!]. There is a way to increase the volatility of things to allow my rise back to my "true" skill ranking faster but it increases the inaccuracy of the system as a whole. That's a discussion point surely but jiggling the handle of the system in this manner doesn't seem to me on the surface to be worth the increased system inaccuracy for a few "special" cases *points to his rotating avatar*.
Your Pal
Snack
If I was smarter and liked Baker more I'd get on TS with him and work on the implementation of bits and pieces of newer AS features. Phewwww good thing I'm not the first, as for the second in spite of myself I do, must be that lovely Serbian accent of his
.There will always be those who go on gut and feelings when numbers are the way to tackle an issue. It's why casinos exist and government policy is generally criminally asinine when trying to regulate science based things.
If somehow the ranks got implemented incorrectly then we would have a talking point. The fact that a few vets, like myself, have really low ranks has to do with an individual's in game behavior over a span of time AND the fact your skill compared to the community in general is across a range of values across a confidence interval of *insert number of your choice which effects the span*. In my case you remove the 15W vs 120L (or some such) span of game play where I was anti-stacking every game in an effort to encourage fair game play and you'll get a more realistic rank IMHO. *shrug* I wasn't good enough to change the outcome, obviously and I pay the price via my ego [I honestly don't give a $#@!]. There is a way to increase the volatility of things to allow my rise back to my "true" skill ranking faster but it increases the inaccuracy of the system as a whole. That's a discussion point surely but jiggling the handle of the system in this manner doesn't seem to me on the surface to be worth the increased system inaccuracy for a few "special" cases *points to his rotating avatar*.
Your Pal
Snack
Ssssh
It'd probably help if we got TE to link your old data to your current MemberID. Remember when the two database updates we got wouldn't merge? One of the problem players was you, sir. You're missing a good year to 18 months worth of game data.
Unfortunately, even if we get you re-linked, TE would have to recalc across the entire database. I doubt he'd do this even if I paid.
Edit: tyop
Unfortunately, even if we get you re-linked, TE would have to recalc across the entire database. I doubt he'd do this even if I paid.
Edit: tyop
Last edited by sgt_baker on Wed Jun 17, 2009 1:04 pm, edited 1 time in total.
Granary Sergeant Baker - Special Bread Service (Wurf - 13th Oct 2011)
-
Malicious Wraith
- Posts: 3170
- Joined: Sun Jul 16, 2006 12:51 am
If Sgt. Baker and I ever hated each other in the past, which I honestly don't remember (But I will assume is the case, as I think I have fought all these math types), I don't hate him anymore.
IG: Liquid_Mamba / FedmanUnknown wrote:[Just want] to play some games before Alleg dies for good.
I don't want that time to be a @#(!-storm of hate and schadenfreude.
sgt_baker wrote:QUOTE (sgt_baker @ Jun 17 2009, 07:58 AM) It'd probably help if we got TE to link your old data to your current MemberID. Remember when the two database updates we got wouldn't merge? One of the problem players was you, sir. You're missing a good year to 18 months worth of game data.
Unfortunately, even if we get you re-linked, TE would have to recalc across the entire database. I doubt he'd do this even if I paid.
Edit: tyop<--- ahh the sweet irony
Yeah well I've always been "special" in more ways then one. *meh* whatever I honestly don't care one way or the other and if it helps keep TE from doing more work then
. I'm in it for the rock ramming and fart jokes anywayedit: ahhh it seems as if the "hate" joke might have been a dud. So for the record: I think of Bakertine as one of my real friends in this world and hold the intellectual knife fights we had during the original go at this effort as some of the best times I had in Allegiance evah.
Last edited by MrChaos on Wed Jun 17, 2009 1:33 pm, edited 1 time in total.
Ssssh