Forged Alliance Forever

Re: Differential Rating System in Team Games

2013-11-02T04:37:21+02:00

No, definitely don't go picking within a team. Then you get the same stuff as in Battlefield 2/3 where others revive you five times instead of killing the opponent for the sheer number of points they get for that... such rewards have as unintended consequence the focus on individual "teamplay" while losing the game...

Statistics: Posted by E8400-CV — 02 Nov 2013, 04:37

Re: Differential Rating System in Team Games

2013-10-25T07:00:56+02:00

Mafa wrote:
First, it is of course impossible to make an algorithm or a software that absolutely reliably categorizes one's performance in a game like FA. Second, it is obviously possible to make an algorithm or a software that more or less reliably categorizes one's performance in a game like FA. Such software exist e.g. for poker.

Well we already have a rating system. so it is obvious that one can exist just not one as complex as the one being suggested. also poker is not a valid example to compare FA to. something like Starcraft could be a valid example(did I just say that?) but not poker.

my opinion of this idea is that even if it could be done i dont think it should be done. reason being is that its a team game so if people win or lose rating differently on the same team then it promotes individualism. thats not something that should be in a team game you should win or lose rating as a team.because the whole point of a team is to work together.

Statistics: Posted by Reaper Zwei — 25 Oct 2013, 07:00

Re: Differential Rating System in Team Games

2013-10-24T23:50:17+02:00

Mafa wrote:
I didn't read this very carefully, but a couple of points. First, it is of course impossible to make an algorithm or a software that absolutely reliably categorizes one's performance in a game like FA. Second, it is obviously possible to make an algorithm or a software that more or less reliably categorizes one's performance in a game like FA. Such software exist e.g. for poker.

For the guy proposing this system for FA, I have only one comment: you have written huge walls of text, but you have not given a single exact numerical formula or algorithm that should be used to make this estimate of one's performance in a team game. In the absence of these, anything you have written has absolutely no value at all.

Give the formulas for determining the performance based on the data, and I am sure someone will write the code to execute it. If you can not do this, then it simply means that you yourself do not know what you want to evaluate.

The reason is that this idea works, until you start assigning numbers, and definitive values, then it begins to get subjective and falls apart.

The other problem is that the targeting priorities of units change over to course of the game, due to time and positioning.

I agree and think it COULD be possible to get some kind of rough system going... maybe. But I would rather performace was measured by a non-subjective approach. Winning.

Statistics: Posted by Nombringer — 24 Oct 2013, 23:50

Re: Differential Rating System in Team Games

2013-10-24T13:30:06+02:00

I didn't read this very carefully, but a couple of points. First, it is of course impossible to make an algorithm or a software that absolutely reliably categorizes one's performance in a game like FA. Second, it is obviously possible to make an algorithm or a software that more or less reliably categorizes one's performance in a game like FA. Such software exist e.g. for poker.

For the guy proposing this system for FA, I have only one comment: you have written huge walls of text, but you have not given a single exact numerical formula or algorithm that should be used to make this estimate of one's performance in a team game. In the absence of these, anything you have written has absolutely no value at all.

Give the formulas for determining the performance based on the data, and I am sure someone will write the code to execute it. If you can not do this, then it simply means that you yourself do not know what you want to evaluate.

Statistics: Posted by Mafa — 24 Oct 2013, 13:30

Re: Differential Rating System in Team Games

2013-10-22T23:36:53+02:00

Yes, the other problem with characterisation is that sometimes I like to rush, sometimes I like to ecowhore, sometimes I want to find a mid ground..... ect... People will change their play styles, not only as they evolve as players, but also by choice

However, I really do like the idea of this positive reccomendation system and I think it would be a nice thing to see, but of course you need likeminded people to code it

Statistics: Posted by Nombringer — 22 Oct 2013, 23:36

Re: Differential Rating System in Team Games

2013-10-22T22:35:50+02:00

its neither supposed to be a feedback, nor to be a characterization of your style of playing.
its purpose is that you might see "oh this guy gave me a helpful recommendation" and then you want to help more people to get more of them.
and thats why there is nothing negative.
i think this decreases hate and supports being kind and helping other people.
Maybe people with 100 helpful recommendations will be as admired as pro's

Statistics: Posted by Golol — 22 Oct 2013, 22:35

Re: Differential Rating System in Team Games

2013-10-22T20:19:53+02:00

Golol wrote:
i agree that this isnt really possible.
and it could propably be exploited easily.
how about we start with something else.
recommendations (or sth. like this)
this system is in dota 2
you can give other players diferent kinds of positive recommendations.
in dota there are:
helpful (most important one)
friendly
leader
forgiving
maybe some more can be made that have to do with supcom
and i dont mean charakteristicd like "ecowhore" or "turtle"
and there also shouldnt be negative things to prevent hate.
a player could only give a recommendation once to another player to prevent two guys giving them eachother a thousand times

Thanks for the interesting idea. It is probably good to have this anyway, even though it is somewhat subjective, because where the subjectivities of many agree and intersect, the voting on a player's performance would tend to converge toward a consensus that reflected, at least partially and significantly, the reality of what it is like to play with that person. It would also become effective as a datum over time for correlating certain types of play with certain damage patterns and resource gifting patterns. "Ecowhore" seems to be a backhanded compliment, really. I'm not sure ruling out all forms of negative play is a good idea. Feedback is feedback. It should be as objective as possible, however. "Ecowhore" doesn't seem objective, or even clear, for example. I find turtling to be a useful strategy from time to time, or in certain phases of a battle, or in certain parts of a map. I think players should be able to recommend, or disrecommend no more or less than once per game, per each other player on their own team.

Statistics: Posted by MetaOntosis — 22 Oct 2013, 20:19

Re: Differential Rating System in Team Games

2013-10-22T18:50:57+02:00

I've mentioned this before, i think this is a nice idea, but as has been mentioned before it will be hard to find someone who will be willing to do it, so you may have to do this yourself.

i believe a much more useful alternative would be a stats page, which would display many numbers about the player - this would give and overall impression of what kind of person they are, and, more importantly it will feed into your work (of a better rating system), as you need to have a method of gathering and storing these numbers, and the rating and the stats both need a similar (or even the same) framework to function, which i believe is a large portion of the work.

-additionally, it will be easier to get people to do this for you, as it has more benefits.
-the remaining work may have to be done by you, but it will be putting all these numbers into an equation that works
-testing this would be the hardest component - the more realistic your system is, the harder it is to simulate it, and testing it will require many, many games played (could be achieved by adding the new rating system in parallel to the current one)

i know you think your text doesn't need changing, which may or may not be right, but i think its safe to assume that the proportion of people who read it is inversely proportional to its length, its not a case of whether your text is good or not, but more of a case of how much it suits your needs - and i believe your needs are to get as many people to agree with you, and therefore read it, so getting more people to read it is what you want, and for that you have to adapt your text, to make it as convenient for them as possible.

from what i can say:
- formatting goes a long way
- lists
- try to keep it short, but with detail
- split your paragraphs up, i can manage many short one better than one huge monster
- forums are not books, where they're split by pages, here all your (lovely and informative) dribble is in one huge chunk
- this post isn't a good example

I'm not making this post to criticize you; that would be a waste of my time; i value your work/ideas, and i think that if you do what i say (obviously, its what i think) you have a better chance of achieving something useful

so, forget about this for a bit and make a stats page first, then you can (relatively) easily implement this, this way you get 2 improvements in one

also i think there's a good chance that if you want to get anywhere you're going to have to do a significant part yourself, once you get going you can ask people to help you with specific things which they're much more likely to help you with.

and write a book! you know you want to! (i think i would actually read it)

Statistics: Posted by Exotic_Retard — 22 Oct 2013, 18:50

Re: Differential Rating System in Team Games

2013-10-22T10:44:22+02:00

how about we start with something else.
recommendations (or sth. like this)
...
there also shouldnt be negative things to prevent hate

I like this idea!

All we'd need to get this ball rolling would be someone capable of writing in the script/programming code w

I am a programmer and i know that almost anything is possible, but the resources (time/skill/motivation) are limited. So if Ze_PilOt or others say it's impossible they dont mean it cant be done, they mean you won't find anyone that is willing to do it.

In my oppinion the best we can do is to realize first the simple things that have a great impact on game play or organisation. There are so many things that should be done before.
- just spread the word about faf ( this has the most impact and is doable by anyone). Buy your friends a copy of Forged Alliance and force them to play
- donate some money for the servers
- complete and bugfix coop missions (thats what Ze_PilOt is currently doing)
- complete Galactic War (thats what i think is Ze_PilOt long term plan)
- team chat in lobby
- stream replays for casts among a group of players
- make casts
- better homepage
- update / make new maps

you see there is a lot of work to do! And now think to yourself what is more logical: to spend 5000 hours to realise your idea, which is wanted perhaps by 5 - 40 people or just to spend all this time for the other things I listed above?

Statistics: Posted by RoLa — 22 Oct 2013, 10:44

Re: Differential Rating System in Team Games

2013-10-21T16:38:46+02:00

i agree that this isnt really possible.
and it could propably be exploited easily.
how about we start with something else.
recommendations (or sth. like this)
this system is in dota 2
you can give other players diferent kinds of positive recommendations.
in dota there are:
helpful (most important one)
friendly
leader
forgiving
maybe some more can be made that have to do with supcom
and i dont mean charakteristicd like "ecowhore" or "turtle"
and there also shouldnt be negative things to prevent hate.
a player could only give a recommendation once to another player to prevent two guys giving them eachother a thousand times

Statistics: Posted by Golol — 21 Oct 2013, 16:38

Re: Differential Rating System in Team Games

2013-10-19T13:51:07+02:00

All we'd need to get this ball rolling would be someone capable of writing in the script/programming code which enables a program to look at the commands in a game and classify that data into predetermined categories. Then we'd need equations set up which examine the raw data and give a prelimenary summary of points for and against based on that raw data. Simultaneously a subroutine would activate which would take that data and check it for certain "anomolies" which indicate better play by some over others, or of seveal over one in particular, or play that was nominal, or play that was worse than nominal (we'd obviously clearly and objectively define nominal, not impossible as some think), and then it would re-evaluate key data according to any anomolies found, tallying up a final score. for each player as nominal wins or losses for their own particular rating in the group vis-a-vis their opponents ratings (already done it would seem), but modify those according to whether a player's team play was found to be above or below nominal, and in direct proportion to the degree of that deviation.

The hardest parts of this are three areas:

It is hard, and would be a challenge for anyone who undertook it, but a team of likeminded individuals with a good set of skills in areas of programming, mathematics, and conceptual analysis in some type of theoretical capacity would be able to tackle it without too much diffculty. Even just laymen with some talent in each of these or related areas could pull it off admirably well I think.

I'm not sure when or if such individuals will express their interest in such a project, but I'll always be happy to provide my own input into it help it in any way that I can. That such a thing might be possible at all it had first needed to be brought foward for others to examine, some of whom might have the ability and interest to veture forth to make an attempt at some point, but only if their minds were properly stimulated with the general idea, and properly motivated to see it as a worthwhile endeavor. That was the least I wanted to accomplish by this discussion. But also in discussing it I opened it to challenges, some of which were reasonable, but most of which unfortunately were not. That's how it is. But the reasonable ones were valuable and helped me further think out and reinvent my own ideas, and test them and see if they really were sound or if my thinking really was valid. So in that sense also this was a worthwhile endeavor in my view. Thanks for helping me with this and I'm looking foward to yours and anyone else's constructive criticisms or other ideas and input.

Statistics: Posted by MetaOntosis — 19 Oct 2013, 13:51

Re: Differential Rating System in Team Games

2013-10-19T13:53:41+02:00

Nombringer wrote:
I still, dispite what you already said, do not think it is possible to create an unbiased statistical analysis that measures the how "well" somebody plays within a team.

As soon as you start putting "numbers" in to value variable x of variable y your system immediately becomes subjective, if you want to it to or not. You define primary and secondary targets due to the mirror relationship between two players, and you said that how they are weighted will be determined contextually. You simply connote desighn an algorithm to do this as the amount and type of interaction with a players mirror will depend on:

The size of the map

The type of the map (starting positions of players) Have a look at a 2v2 on haven reef, these provide a great example of a map where helping your teammate, as the back player is essential, and where being able to soak up damage as the front player, is also essential, both of these things are rewarded my your system. However, depending on the design of another map, the same two players will interact less, and thus get less points, pureley because of the map design limited their interaction.

The type of game each player is playing (One player might spam, one might eco, neither of these are bad things, just different strategies, in some cases, one of the other might be objectively worse, however these two will always be rewarded DIFFERENTLY by your system.)

At the end of the day, the current system encourages players to do WHATEVER they can to win, if they follow the system, then they will always try to do what is necessary to win, if that is team play, it will be team play, if it is eco, it will be eco. Your problem is with players who DO NOT follow what the rating system sets out to do, (trolling) if they think they can win by doing idiotic and selfish things, then it is they are more likely to lose rating and they will go down. Yes it is unfortunate that sometimes you do all you can do, and you still lose and are not rewarded, but the fact remains that winning and losing is the only unbiased way to determine what side played better.

Yes, occasionally you will get shafted, but I would prefer to get unlucky occasionally, and have it all balance out in the end, rather than get SOME numerical gratification from those times, with no guarantee that it will eventually balance out correctly.

I do want to add to this and mention that, to clarify about such differences in maps as you mention, it is still a NET INCREASE if they win, and a NET LOSS if they lose, regardless of what type of map it is. If they are not able to cooperate much, then it begs the question of how it is a "team game" in essence. of course they can cooperate, maybe not as soon or as directly or as quickly or as often or whatever as in other maps, but there is still a difference between each player defeating their opponents alone, as if there were two 1v1 games on a map that had nothing to do with each other, and instead helping each other when opportunities arise. Besides, if they cannot help each other much, is it really likely that their respective opponents can more easily coordinate attacks on just one of them? If they cooperate when it is possible, in some of several various ways, rather than allow an enemy to coordinate such an attack if it does occur, that is a significant factor in their repective scores. LIkewise, if they lose because a coordinated attack against one player occured, but the other player did not squat to help with units he had available which can be EASILY algorthmically determined to be relevant to a proper defense, then it would mitigage the loss of score for the first-defeated player, letting the second-defeated player take that share of the loss in points (not all of them, obviously, but a duely determined proportion).

Statistics: Posted by MetaOntosis — 19 Oct 2013, 13:25

Re: Differential Rating System in Team Games

2013-10-19T11:50:44+02:00

da_monstr wrote:
Can you please make a TL;DR for 1st post?

Btw, I took a shot at a TL:DR of the OP, in spite of my resistance to this and my stubborn refusal to think of my writing as needing any changing. But still, it is a good challenge to condense things and see if they can't be polished. It's probably still long even for one of those, but it's way shorter than the OP.

Statistics: Posted by MetaOntosis — 19 Oct 2013, 11:50

Re: Differential Rating System in Team Games

2013-10-19T04:58:28+02:00

so i suppose im considered one of those close minded dogmatic people. you have an opinion about something in this great game of ours. thats great. you are entitled to that opinion. what i have been "dogmatic" in my rejection of is 2 things.

1. you seem to think your opinion is better than everyone elses. calling people who agree with you openminded and imaginative and those that dont closeminded and dogmatic about their closemindedness.

2. the much more important disagreement: you have talked a lot about what you think is possible and what you would like to see but you have not talked any about implementation. collecting the stats for energy, mass, kills and losses would i believe be easy because the game already does this. you however want a much more comprehensive list of stats. both require a program to do so. the list of stats that you have would be quite difficult to code. the reason for this is much of what you want is based on evaluation of the game. not only do you want a program to tally kills but you want it to then determine which kills were more valuable than others? how would a program go about doing that? now maybe if you were just talking about commander kills then maybe but as Zep already pointed out getting a commander kill is not always the right move. is it the right move most of the time? sure but not always. you think a program is capable of deciphering that? the code for that would have an enormous list of exceptions and that's just for that one case. you would have to go through all of those stats and code for tons of exceptions and you would also have to do it for every map.

what you want may sound innocent enough to you or others. but it would require a massive amount of work to be done.

Statistics: Posted by Reaper Zwei — 19 Oct 2013, 04:58

Re: Differential Rating System in Team Games

2013-10-19T12:16:31+02:00

Nombringer wrote:
I still, dispite what you already said, do not think it is possible to create an unbiased statistical analysis that measures the how "well" somebody plays within a team.

As soon as you start putting "numbers" in to value variable x of variable y your system immediately becomes subjective, if you want to it to or not.

We disagree on this, possibly because of differing views on where the line is drawn between subjective and objective, or what those terms denote.

Nombringer wrote:
You define primary and secondary targets due to the mirror relationship between two players, and you said that how they are weighted will be determined contextually. You simply connote desighn an algorithm to do this as the amount and type of interaction with a players mirror will depend on:

The size of the map

The type of the map (starting positions of players) Have a look at a 2v2 on haven reef, these provide a great example of a map where helping your teammate, as the back player is essential, and where being able to soak up damage as the front player, is also essential, both of these things are rewarded my your system. However, depending on the design of another map, the same two players will interact less, and thus get less points, pureley because of the map design limited their interaction.

The type of game each player is playing (One player might spam, one might eco, neither of these are bad things, just different strategies, in some cases, one of the other might be objectively worse, however these two will always be rewarded DIFFERENTLY by your system.)

This also struck me as a possible difficulty at the outset unless I handled the issue of target quality (primary or secondary) in a proper way. It is a good point that you raise here. I would clarify the meaning of the idea of these target prioritizations. In cases where matches are between mirrors, it would seem to me that there is no problem with using the criteria for targeting qualification as a factor for point adjustment for several reasons.

1) If the relationship between the opposing forces are homogenized in certain ways it could be for several reasons. One would be the map has not differentiating terrain sufficiently like to cause a different emphasis in unit creation of any kind between players in a given team. Therefore power struggles between mirrors will be more flexible and can drift to non-mirror opponents more easily. That simply means that any variables for team play scoring which would be dependent upon the likelihood of been locked into battle with one's mirror would be muted relative to a map where they weren't. But it would seem that in this case they are muted in a well distributed manner, so this offsets any biases. There could still be plenty of room to demonstrate team play according to the mechanics I suggested. Focused attacks and defenses still can easily happen, and still be found to more the stripe of some players than others, and those deserve better rewards in point value than merely fending for oneself and going it alone on attacks. Those are still rewarded to the extent that they are successful. The intent is to find a clear and objective way to locate the data which identfies those that are clearly harmful to the team, or self-neutralizing, or helpful, diffferentiate them, and reward them differently. Some coordinated efforts will be more successful than others, some harmful to the team, some just weaker than the opposing teams. All that is still rewarded or penalized by a net gain or loss depending on if the group wins or loses as a hole. Let's not forget that I haven't dismissed that or thown that out or denied its value and relevance. I am proposing a sensible modification to it to account for individual variances of team play, and simply individual quality of play plain and simple, which are masked within the group win/loss evaluation as it stands bluntly.

2) There may even be asymmetries such that one side of a "mirror" can better use sea forces (because can often reach sea much sooner than mirror can), or something like that. Well, in that case, the "proper" mirror is the opponent on the other team who has the same strategic access to a symmetrically placed ocean venue, though not directly accross from his functional "mirror". In this case it is really a hidden symmetry. In most maps it is quite clear who one's "main oppenent" would be! There are not SO MANY maps that this cannot be determined objectively by a cursory look at high-level play and seeing how the opposing forces tend to clinch, and with what forms of units primarily. Those maps for which this has been accomplished, Setons is a relatively easy one to use as an example, would be available for a "team rating". Indeed, they make the perfect battlegrounds in which to determine such teampla mettle. That having been so determined, the characteristic is determined which enables that rating to solidly established via those maps. Even if other maps were not amenable to this, they would still be ratable, but the team play variables involved which would be brought into focus might be otherly stressed, or simply muted to some degree. Even if maps are highly asymmetric, or spawn points are, that still leaves plenty of room for demonstrating individual quality of play as part of a team which needn't be muted by a single, flat rate win or loss score for the group! I'm not saying the score is decided merely by any particular one of these kill/damage variables, nor necessarily by all of them without any room for emphasis appropriate to a given map or playing condition. Not at all. No inconsistency of application of the system due to maps being different as far as I can tell.

If in these two points I haven't understood what you are getting at, please detail where I've missed the idea. I think it works fine.

Nombringer wrote:
At the end of the day, the current system encourages players to do WHATEVER they can to win, if they follow the system, then they will always try to do what is necessary to win, if that is team play, it will be team play, if it is eco, it will be eco. Your problem is with players who DO NOT follow what the rating system sets out to do, (trolling) if they think they can win by doing idiotic and selfish things, then it is they are more likely to lose rating and they will go down. Yes it is unfortunate that sometimes you do all you can do, and you still lose and are not rewarded, but the fact remains that winning and losing is the only unbiased way to determine what side played better.

Yes, occasionally you will get shafted, but I would prefer to get unlucky occasionally, and have it all balance out in the end, rather than get SOME numerical gratification from those times, with no guarantee that it will eventually balance out correctly.

My system doesn't discourage a player doing whatever they can to win, not in the slightest way. I don't see how you come to that conclusion. It doesn't affect that aspect of play in any way. The fact of a team win will always result in an increase of everyone's score, and the fact of a loss will result in a loss to everyone's score. That is always true. Therefore there is no alteration in the motivational incentives that would discourage inventiveness or cleverness in any way, shape, or form. The question is simply whether or not a player's actions can be found to be, BY THE END OF THE GAME, to have had some net result toward, or away from the end result for the team. The thrust of the effects of a player's actions can be determined accurately by the ways I describe, which are not subjective anymore than any other system devised by someone to evaluate something is subjective. It is certainly not exceptionally subjective. I'm sticking close to some basic ideas and I'm maintaining a lot of flex in how those ideas should be nailed down into numerical parameters. I'm not saying they MUST be the exact way that I describe them, and I'm not prescribing them to be any preset system, nor proscribing any. I gave a general outline that cannot be properly dismissed out of hand with vague assertions of subjectivity, or claims that they are unworkable by those who really seem not to be inclined to even imagine possibilities for them to work anyway. Whether or not it is ever seriously developed or implemented in any way, this exploratory discussion has certainly borne out a few things. It seems that those who like it are openminded without being dogmatic, but those who don't are closeminded AND dogmatic about it.

Regardless of people's feelings about it, its doesn't seem that any truly objective criticisms have been raised. I'm having to ignore a lot of rhetoric instead. Your criticism on the way that differing maps may affect the consistent application of this rating system was the only real exception to this. I think I have given a fair reply to them. I explained how the differences in how the ratings would shift the weighting of score increases from wins or losses differentially per player in a team, as an adjustment on the flat rate increase they received (or decrease) due to their individual pregame rating relative to the average team ratings (their deviation I suppose, from the mean) wouldn't be a problem because it would be well distributed in every case that I can see, and it would be non-biased over a career of games because a "real" rating for a player wouldn't be clear anyway, no matter what system you used, if they only kept playing in one position or on one map! Those factors would have the same variances and leveling out in the system I'm proposing, which is only a modification really, as in the basic "flat rate" rating system, which is really not a "system" at all. It is barely an indication of anything specific to the player, because it is divided among all members of the team, each of which played quite differently, and is never, or in 99% of the cases, not consistently or even often the same team. Certainly not enough to create a valid "group based" rating. The specific players that make up the teams varies widely, and we don't really get a picture of any particular player's performance in a team this way. Even after hundreds of games, his own performance as a team player is blurred. In the current system we only know from his rating increase over time that he is playing better in some ways than before, not which ways, nor how much, because in fact it could easily be that he is just getting into better groups, or groups that win more because they each had an opposing team who gave the win away by sucking badly, sometimes due to ONE PLAYER. Maybe his eco got better and he got lucky a lot because it turned out that lower level games are simply big messes anyway. It's the same in chess. We don't consider a chessplayer's games meaningful indications of "skill" until he's gotten up into the upper club, frequent competition level. He and his lower rated opponents make so many mistakes that give their games away that it is really a coin toss if any given game is one or lost. The idea is that a certain better plan was put into motion which had a nuanced superiority to the opponents, or that their plans were so even that it came down to one of them blinking first and slipping. The same here. Lower rated "progress" is about things like learning how to get your mexes upgraded. That has very little if anything to do with team play per se in the way I'm saying matter ON ALL LEVELS, from lowest to highest. I'm talking about a modification to a rating system for teams such that those differences are, to the extent possible, rewarded differentially in objectively consistent ways that are clearly valid, not matter whether a player is a noob who barely knows that he should build a factory asap to a pro who wins the world cup. That's a ameasure of objectivity right there, and of theoretical power, because it has consistent application over a wide variety of player skills and knowledge bases.

I say again that It clearly matters not only THAT a team wins or THAT a team loses, but HOW WELL EACH ONE PLAYED REGARDLESS if the team won or lost. And that CAN be objectively determined to a SIGNIFICANT extent, and I gave objective definitions of how that is approached in the OP.

But I have to get to something right now and can't give a plethora of examples. I'll get some ready by the time I post again. They will be examples of things that should be differentiallly rewarded for sure, or differentially penalized. Mostly what I'm into here is finding GOOD play and rewarding it MORE, and rewarding nominal play less, and rewarding everything less than that, least. When a team wins that means more points gained in that order. When a team loses that means less points lost, in that order. Couldn't be simpler, saner and more objective than that.

Statistics: Posted by MetaOntosis — 19 Oct 2013, 03:08