Free Allegiance Forums

Wasp wrote:QUOTE (Wasp @ Nov 27 2006, 10:24 AM) Was the dropping of the probes more indicative of skill or is what I did with the sum of ALL the elements utilized in my attack the real measurement of skill?

Thank you, Wasp, for summarizing neatly the entire problem with action-based scoring. Abstracting the individual actions out so that ranking players is done by overall contribution to a team is the underlying assumption of Elo and every other statistical system, and I'm glad to see you agree with it.

Terralthra wrote:QUOTE (Terralthra @ Nov 26 2006, 05:34 PM) Thank you, Wasp, for summarizing neatly the entire problem with action-based scoring. Abstracting the individual actions out so that ranking players is done by overall contribution to a team is the underlying assumption of Elo and every other statistical system, and I'm glad to see you agree with it.

You're welcome, however, I believe you are mistaken. I do not agree with the assumption ELO uses to determine a players overall value. Nor did I summarize enough the true value of action based scoring. You would know this already had you read anything but your own type.

ELO's failing is that is uses a flawed logic.

I'm glad to see you're finally getting the point. When I see you begin to redirect this thread into a personal debate, I know you're grasping straws in a failed effort to support your belief. If you can offer any logic behind your assumptions (or ELO's), I'd love to hear it.

Wasp wrote:QUOTE (Wasp @ Nov 27 2006, 11:36 AM) You're welcome, however, I believe you are mistaken. I do not agree with the assumption ELO uses to determine a players overall value. Nor did I summarize enough the true value of action based scoring. You would know this already had you read anything but your own type.

ELO's failing is that is uses a flawed logic.

I'm glad to see you're finally getting the point. When I see you begin to redirect this thread into a personal debate, I know you're grasping straws in a failed effort to support your belief. If you can offer any logic behind your assumptions (or ELO's), I'd love to hear it.

Did you not say
QUOTE Here is an example of strategic probing that in my opinion still requires little effort.......Suppose I want to stealth bomb the enemy's tele and it's a long way from home passing through occupied sectors. I first scout the route and destroy enemy probes along my expected path but I don't destroy all their probes as this might warn them of my planned attack. Meanwhile I drop probes where the enemy isn't likely to spot them yet my probes will likely spot the enemy in case I need to alter my chosen path. I return to get my stealth bomber, study the route once more and carry out my attack.

Was the dropping of the probes more indicative of skill or is what I did with the sum of ALL the elements utilized in my attack the real measurement of skill?[/quote]?

Was your point not that the dropping of probes was fairly skill-less in and of itself, but rather the sum of all of your actions, planning, and foresight that was of value to the team, that contributed to the team?

Terralthra wrote:QUOTE (Terralthra @ Nov 26 2006, 06:51 PM) Was your point not that the dropping of probes was fairly skill-less in and of itself, but rather the sum of all of your actions, planning, and foresight that was of value to the team, that contributed to the team?

You have that correct.

But your assumption is that those actions of that pilot was directly tied to the game's outcome. Using this flawed logic that a better player will be on the winning team more often than a lesser cappable pilot is the problem with the ranking system. This is not a 1 vs 1 game.

Please tell me how you're going to weed out all the other factors that influence a game's outcome and determine a single players rank. Until you can control (isolate) all the other variables....commanders influence on game outcome, different maps, rock placements, available factions, techs, changing players.....all the other variables that alone influence game outcome, you'll never in our lifetime isolate the player's individual rank and therefore render autobalance useless.

I think that a simple drug test should be conducted and applied to the ranks of all SRM members.

cause our skill can drastically change depending on what has been recently smoked/ drank/ injected/ freebased / consumed/ invented.

/tongue.gif" style="vertical-align:middle" emoid=":P" border="0" alt="tongue.gif" />

Bottom line is this: for team based games ( group win/loses ) MS Research ( the makers of this very game ) has published results showing great success predicting game outcome by using win/lose as the sole criteria. This is across a wide range and type of games.

Professor Glicko uses something VERY similar for ranking chess and has a modified version to deal with rank whores (sitting on your high rank without playing). This approach will almost certainly work for team based games since MS Research used his work extensively.

Ive read a bewildering array of "facts", and ill conceived notions regarding statical rankings. Conservatively the amount of drivel approaches 75%

Im sure this will change no closed mind, nor the rhetoric from the bewildering array of factions regarding ranking.

I remind all camps that until Management is presented a rigorous numbers based proof the future plans remain fixed. Wail or rejoyous as you see fit.

MrChaos <--- no dog in the fight just a wee bit sick of it all

_SRM_PsycHosis wrote:QUOTE (_SRM_PsycHosis @ Nov 26 2006, 10:35 PM) I think that a simple drug test should be conducted and applied to the ranks of all SRM members.

cause our skill can drastically change depending on what has been recently smoked/ drank/ injected/ freebased / consumed/ invented.

I agree Psy,

It will certainly be more accurate than ELO can ever produce.

I was under the impression that one of the reasons the squad system was created was to give the player base an easier way to play unstacked games.

My whole point is so simple.

Join SQUADS to stop stacking. Play squad games.

/wub.gif" style="vertical-align:middle" emoid=":iluv:" border="0" alt="wub.gif" />

MrChaos wrote:QUOTE (MrChaos @ Nov 27 2006, 07:59 AM) Bottom line is this: for team based games ( group win/loses ) MS Research ( the makers of this very game ) has published results showing great success predicting game outcome by using win/lose as the sole criteria. This is across a wide range and type of games.

Then provide us with a specific example of which game SIMILAR to allgiance that calculates accurately a player's individual value. We're all dying to hear what you have to offer.

QUOTE Professor Glicko uses something VERY similar for ranking chess and has a modified version to deal with rank whores (sitting on your high rank without playing). This approach will almost certainly work for team based games since MS Research used his work extensively.[/quote]Unfortunately, you can't produce any details of how this is to be achieved. Simply stating that someone's researched it, especially MS, is extreemly naive thinking. What "approach" are you talking about? "It will almost certainly work "SINCE" MS Research used his work"?!?!....Are you kidding? Do you hear what you're saying? Does that conclusion use the same "logic" you're applying to this ranking system?

Comparing the ranking of chess players to the ranking of allegiance players just goes to show, you're comparing apples to oranges. Chess is a 1v1 game! The chess game uses the process of elimination to arrive at the value of the piece/player. The chess board doesn't change. There are no factions. There are no tech trees. The pieces don't change. The pieces don't play multiple roles. Each chess player starts the game with exactly the same pieces. The pieces don't think for themselves with the freedom to move about the chess board of their own free will. When you lose your queen, a pawn doen't dock and launch in a new queen ship!

Allegiance imposes far too many variables onto each game that the likelyhood of duplication or having two games similar enough for comparison is nil. Unless you get enough samplings of games where these variables are constant enough to eliminate them as the reason for game outcome leaving the pilots as the only remaining variable to abstract, you'll never EVER find rank.

In the case of chess, you're measuring only two things...the SINGLE individual's skill against a SINGLE opponent and the value of each chess piece. Since all other matters in chess remain constant (board layout, pieces used, pieces functions and abilities), we can over time determine the value of both. But if you were to allow the chess pieces to do what Allegiance players can do and allow the chess board never to be the same and impose the infinate combination of variables Allegiance has upon the game, you'll never be able to derive the value of each individual chess piece.

QUOTE Ive read a bewildering array of "facts", and ill conceived notions regarding statical rankings. Conservatively the amount of drivel approaches 75%[/quote]
So then why don't you state specifically where the flaw is in our logic? If you would do this you could end this discussion once and for all and everyone can see for themselves the logic in your thinking. We've stated over and over again the flaw in your thinking and you refuse to validate your assumptions. Again, instead of validating the logical process of how rank will be determined accurately using the process of elimination/abstraction of all the other variables, you offer an attempted redirection of discussion from this system's flaw to one of "drivel". Stay focused if you can.

QUOTE Im sure this will change no closed mind, nor the rhetoric from the bewildering array of factions regarding ranking.[/quote]How can you expect to change the viewpoint of another if you don't offer a viewpoint in the first place? You've offered nothing to describe the relationship of "MS Research" or "Glicko's" work to our ranking system. Nor have you offered any explanation/validation of your irrational conclusions.

QUOTE I remind all camps that until Management is presented a rigorous numbers based proof the future plans remain fixed. Wail or rejoyous as you see fit.[/quote]

I remind all Developers that until Players are presented a rigorous numbers based PROOF that such a system has the capability of measuring something that cannot be measured from simple game outcome, it will be ignored, never used, laughed at, and exposed as a misrepresentation of fact.