smoke wrote:

I dunno. It's an "aggregate" ranking that doesn't tell us who the strongest players are; that depends on how good your partner is for a quarter of your score; where losing a few points in a component where you're near the center of the bell curve could drop you a 100 spots in the ranking, swamping, say, a tough climb from 15th to 5th in another; that would provide essentially random and meaningless results to players outside the top ranks; that allows a "specialist" in a board or two to rank with or above someone who ranks in the top 4 on 20+ boards.

The benefit is that it encourages players to take part in different types of activities? Somehow, the site will benefit if Cona Chris plays more team games and Toto plays a greater variety of boards?   

I'm dubious. Combining scores by combining ranks seems a reasonable approach -- I thought about it for GR and CP when Atilla first launched this thread (really, I did. No, really) --  but it breaks down badly when you get below the top ranks and the scores start to bunch up. Some math might solve this. But, also implicit in this approach is equal weighting by component, and no weighting by number of games played. Doesn't matter for combining GR and CP, but here we're saying that 1 team game has the same impact as 100 tourney games and 1000 regular public games. 

I think this approach is just too simple. I'd be like 5th if we did it, though, so that's cool.

Hello Smoke!

So much to answer to here!

Regarding movement in the ranking.  When a player is lower in a ranking and you get this "clumping" effect, as you say, I was taking the highest rank of the tied value and not the rank that they were sitting at.  So for the multitude of players who have 5 championship points, for instance, I would give them all an equivalent rank of 225 (at the time of writing this), and not 225, 226, 227, etc as they sit.  Of course when you're that low on the aggregate you're part of the "rat race" and one's position would be susceptible to movement amongst the hundreds of other like players on WarGear as it should and 2 additional championship points or a teamplay victory/loss could make someone jump that individual ranking a lot and in aggregate as well.  But that is what it is to be a part of the masses at the center of the bell curve.  If you're a top ranked player, however, then no single loss; team, tourney or standard, is going to drop anyone by 100 positions of aggregate ranking, let alone within the component ranking.  Yeah it'll be hard to climb up the rankings from 15 to 5, like you say, but shouldn't it be?  I think that the results of such a ranking would be a pretty accurate list of the best of the best, for those who don't shy away from any aspect of the game.  For those who do shy away, if they care about ranking, and I think that most of the best who play on a site like this do, to one degree or another, then it should encourage them to play in other facets of the game...  Which is the primary intention of an aggregate ranking and is also part of being the best of the best.  Luieuil is my flag bearer for this ranking because he is the epitome of the type of player that this ranking is tailored towards, but it is not only for him.  The Snoochi Boochies', Djembe's and Davidny212's in this world all get their day in the sun as well by being highlighted for their "all around" game play.

As for the 1 map specialist vs the top 4 on numerous maps proposition.  I'm not reinventing the wheel here.  If one person rocks two commonly played boards and gets 40 points while dominating two highly competitive boards and another person goes looking for lesser played boards that they can collect 3+2+3+2... points on for top four finishes and accrues 39 points over 15 maps... It is what it is and the CP has already given them their rank.  I'm not here to debate the worthiness of how CP is ranked, just that a rank of one equals one aggregate point, a rank of two equals two aggregate points and so on.

Regarding team play.  If one is vain enough, and this includes me ;-), to care about where they sit in any of these categories in the first place and to diligently do whatever they can do to reduce any negative effects on their rankings, then I'm sure that such a person would choose their teammate(s) with a similar streak of vanity to protect and grow their ranking?  I play a lot of team games, but I try to choose my teammates wisely.  A couple months ago I saw a 3v3v3v3 that I wanted to win, so I went to that map's ranking and recruited a player to be my ally who knew how to play that map, and we won.  Since then I have played numerous 3v3v3's and 3v3v3v3's and a 3v3v3v3v3 with the same team and a WarGear relationship was born, and we have grown our team score together as it's easiest to do so in large games with a competent team.  2v2's are slow to grow your team points in!  That's part of the strategy of team games, you know (or learn) how your allies play so that you can compliment each other.  Like any sports team, you get into a groove with each other and learn to cover for one another to maximize on offense and defense.  Team play is a skill as well, and it starts with choosing a team who understands the principle of team work, or is willing to be coached to learn it!  Just as I would not jeopardize my GR by jumping into numerous 1v1's on maps which allow too much luck to be in play with opponents that have GR scores of 500-700, One need not take players of a low caliber as their teammates, or, if they do, then they had better be prepared to do a lot of coaching and teaching as can be a part of effective team work as well!  Team ranking is based on the combined score, so, there is a lot of potential gain by the greater ranked individual if someone with a team score of 700 lets themselves be coached to victory by a more strategically thinking player!  More Risk can equal more gain!

I think that due to the rigidity built into the top end of the rankings already, such a ranking would be mostly static and would be hard to move up on once entering into the upper echelon.  I know that it took Luieuil quite a while to pass Falker in GR and there was some back and forth for a while there.  In aggregate he would have passed Falker sooner and more permanently due to his all around game play style which includes playing in every type, and ranking high in all that WarGear has to offer.  In other words, and as it should be, it would have been more "static" once the change took place on the aggregate.  There was a time when I asked YuriZ to join a tournament and he told me he couldn't because he felt like it could risk his chances of catching Toto in tournament ranking, where he still chases Toto to this day.  The point being, yeah, it's hard to move up at the top and an aggregate ranking should be even more difficult and more static then its component rankings!  It's on this point that I start to get confused on how you could see how something so rigid at the top end could be considered "random" or "meaningless".  Perhaps you could explain that more precisely to me so that I can address your exact concern.

I also don't understand your equating 1 team game with 100 tournament games with 1000 standard games.  It's difficult to move at the top end of any of these categories and, standard is given the most "weight" in the aggregate.  Since teamplay is the most underutilized at this juncture on WarGear, it's probably the easiest to move up on at this moment in time, much to the advantage of Cona Chris, Falker and Mad Bomber seeing as that is their aggregate weakness, but, once the aggregate encourages cross-over play this would quickly close.

Finally, you ask why we'd want players to cross over to other areas.  Well, they don't have to, but, by having an aggregate ranking to encourage such, I think that it encourages more of a sense of community.  Especially for team play games.  It's hard to have effective teamwork if you don't communicate!  This would be in line with other community focused endeavors which could also be implemented on a site like WarGear.  For instance, maybe beyond the strategy forum, a full mentoring program could be implemented where players who were feeling frustrated and who might be considering leaving WarGear all together, but were open to coaching, could sign up for and be matched to higher ranked players to teach them the ropes of the game on a personal 1 on 1 level.  If someone did enough mentoring and grew the community then perhaps they could be rewarded with free premium, or, a free War Gear T-shirt or something.  I guess the easiest answer is that it comes down to what we're all hoping to achieve by being here and what tools we implement to get there.  An aggregate ranking system that encourages cross-over play is just the tool and it's up to all of us on how we use that tool and what we hope to achieve with it.

Your reply was full of a lot of points and I hope I got to everything you asked about, but feel free to point out what I may have missed?

Thanks a lot for your interest Smoke!

Erick

Team ranking should not be included. For starters, all of the other stats are individual achievements, but there's a more important reason.

Team ranking is perhaps the most game-able of the stats.  Just like how people who wish to accumulate master points in bridge often pay to play with a professional, a player who wishes to pad the win column in team games would simply go out of their way to play with strong partners.  There's no question that this is standard practice with many if not most players.

Unless you made it such that teams were randomly assigned, the stat is all but meaningless, and shouldn't be included in the aggregate. 

Edit: I wrote this before seeing the last 4 or 5 voluminous posts

After reading all of the above, I feel the need to strenuously reiterate my point.  Everything I've read confirms my concerns.

If Team scores are included in the aggregate, a culture is created where players only play with certain players, and more specifically, less skilled players would rarely, if ever, get to play with more skilled players.

Personally, I would rather see a completely independent set of scores for team play.. GP, Championship, etc..  I realize that some here are thinking in terms of a site-wide aggregate, where I am thinking much more in terms of an overall individual assessment of play.

..and what of tournament play?  Trophy count?  These are all thing that I also believe should NOT be included in the aggregate as well, but if you're looking for a site-wide stat, then shouldn't these be considered as well?

M57 wrote:

Personally, I would rather see a completely independent set of scores for team play.. GP, Championship, etc..

Agree with M here on this point in particular.

M57 wrote:

After reading all of the above, I feel the need to strenuously reiterate my point.  Everything I've read confirms my concerns.

If Team scores are included in the aggregate, a culture is created where players only play with certain players, and more specifically, less skilled players would rarely, if ever, get to play with more skilled players.

Personally, I would rather see a completely independent set of scores for team play.. GP, Championship, etc..  I realize that some here are thinking in terms of a site-wide aggregate, where I am thinking much more in terms of an overall individual assessment of play.

..and what of tournament play?  Trophy count?  These are all thing that I also believe should NOT be included in the aggregate as well, but if you're looking for a site-wide stat, then shouldn't these be considered as well?

Hey M57,

Sorry about the long replies.  I guess I can be a bit verbose.

You didn't give me a response with what I said in response to you about about H-rating.  You know that Tom likes a consensus, so if we're ever to have a true aggregate that includes all areas, then we'll need to tackle each issue until a consensus (or at least a great majority) can be reached

Onto your points about team play.

The greatest way to accrue points in team play is to find a low ranking team partner who is coachable.  When all the good players get too great a team score then it becomes counter productive to team with them and more productive to make new friends who'll absorb and play the strategy that they're taught.  When they're taught, they become better players, learn to think the game differently then perhaps they had before, and their game play in all their other areas might improve as well.  Communication and community is what I'm looking for here.  Even what you're referring to, about me picking my ally, when I first picked them their team score was below 1000, but I knew that they knew how to play the map.  So, I got the best of both worlds, someone that does not need coaching to be successful, but, also pulled down the combined team score so that I would gain more points.  So long as a low ranking person is willing to be communicative and talk strategy during the game and listen to advice if it needs to be given to improve their play, I have no problem playing with them as my ally as it's in my best interest to win with them as my ally.  So team play is not about the rank or excluding anybody, it is about the synergy.  It is not about picking the strongest players to play with, it is about picking the best communicator and someone keen to pay close attention to the strategy being deployed.  It is about communication.  It is about being able to count on each other to listen and adapt.  In 2v2 teamplay, it can be extremely strategic in how the game can be played, even more strategic then a 1v1 with it's extreme division of responsibilities which might need to be adhered to despite all the temptations to deviate.  I have encountered some teams that have watched my histories and brought my own strategy against me, and they beat me with it, and that is good!  Essentially as more team games are played then the jewel in the rough strategy for the highest ranked player to push their score higher would be to team with low ranked team players who can communicate and learn.  The lower rank player gets a teacher and becomes a stronger player in general and the higher ranked player gets more points.  Synergy.

The whole point of an aggregate is to include all areas of play.  If you had an independent set of categories for standard play, tournament play and team play, then their is no more incentive to have cross-over play, which is the benefit that I, at least, hope to gain from the aggregate in the first place.  I suppose one could make a standard aggregate, a tournament aggregate and a team aggregate, step one, and then have an overall aggregate which is a weighted combination of the three... But then again, we are deviating away from being simple and making a lot of work for Tom!

An aggregate ranking should be simple and intuitive.  It should consider how much weight that each area is given.  In my proposal I had a 50/25/25 Standard/Tournament/Team split, which, as the aggregate gets more people involved in all areas of the game, seems like a fair split to me, but, this is open to tweaking if the consensus feels strongly one way or the other.

Tournament trophies are like CP to a degree.  A bit hard to come by and easier to win with some map specialization skills being implemented.  The problem I have is that the whole point is to keep it simple and when we say "add this" "cancel that" "alter the weight of this", it starts to complicate what should be based on a simple relationship.  4 categories, equal weight based on rank position, simple.  I am open to any tweaks that need to be made to reach a consensus, but, in order to do so, how about you propose how you see things like H-rating and trophy scores actually being implemented in a simple a fair manner rather then just throwing the idea out there without a working solution for it's implementation.

As for what is, or is not considered game-able...  I could write an entire book on that (I'll try not to!).  Some players try to game the system by building up score on lesser played maps to accrue CP.  Some will watch histories religiously and scout the tendencies of their opponents.  Or, better yet, play a map where reviewing the history will make or break the game like in all the blind ants tournaments.  You can have 30 people in a tournament when it starts, but, you know who "games" the system and who it'll come down to before a single game is played!  Some will not enter games against bad players that they feel are not worth the risk given that a little bad luck would cost them too many points and that the victory would not bring enough to make it worth the risk.  Some might also avoid good players who consistently "own" them.  Gaming the system and picking who one does or does not play against is commonplace in all manners of game play, including standard play, so I ask, why should it be any different in teamplay for who you play against and with?  That's why it's called a game, because it's all game-able!

My viewpoint is that I want to create a all encompassing ranking system that will encourage players to play all areas of the game.  That is my end game.  So, my question to you.  Can you assist me to create a ranking system that is not based on the individual but the whole, standard, tournament and teamplay?  That is simple and intuitive?  That is inclusive to the WarGear community and encourages a greater sense of community via cross-over play?

Instead of saying... Well, this part... I think it can be gamed more then the others, lets scrap that part and build a non holistic ranking... How about we ask, how can we improve how that part is ranked, or how it's played, so that the aggregate can work as a whole?  So, how would you improve on my proposal while simultaneously addressing the weaknesses that you perceive it to have?

Sorry about my long responses.  I guess I'm trying to hash this out as I really see the potential here and don't want to see the idea die via over complication as it had earlier.

Thanks a lot for helping me tweak this M57,

Erick

I don't have too much to add, but want to say that I really like this idea and I think Erick is doing a good job presenting the argument for it.

berickf,

Ok, I'll try my hand point by point..

First of all I believe that it's calculated on standard games and therefor introduces additional weight to standard games over tournament or team play.  By including it, standard play with three categories would weigh in at 60% of the overall value of the aggregate. 

Standard is good.. but assuming your point is valid (and I'm still not sure it is.. I prefer it to be 'standard' weighted), I prefer H-R over GR any day of the week.  Over the last two years, my H-R has gone over 2000 to under 1000 and back to over 2000 (it looks to be on it's way down right now), but my H-R has been incredibly stable, remaining  between 60 and 62 during the roller coaster ride.  Either my play is streaky, or perhaps my GR dips when I'm trying out new boards. Regardless.. on any given day, my GR may or may not have any significant value.  (I've argued for a GR moving average before)

Secondly, it's not a "traditional" or "intuitive" category for most of the new players to follow, whereas, the other four are quite easy to understand right off the mark. 

I would counter that anyone who is interested in the mechanics of such a score would be inclined to take the time (and be able) to understand the value of the H-R.  Consider things like the WHIP and OBP in baseball. Most fans/fantasy players value these.  You didn't make the argument that simpler is better, but it is somewhat implied here, and I would disagree with that sentiment.

Thirdly, if someone wins their first 4, 5, 6... 11, 12 games their H-ratings could be artificially elevated for quite some time and would be the top of that ranking simply because of the small sample size of games played.

I can invert the same argument.  It is precisely because the other components take so much time to reach equilibrium, and they have to get there from the bottom, not the top, that the H-R can serve to equalize this problem.  If it's really a problem on the first-few-games end, a minimum requirement could be implemented to minimize the problem ..as you described in your post.  But as far as I'm concerned, if someone comes on the scene and wins their first 5 or 10 games, they should appear on the radar.

I'm not saying your points are not valid.  In fact, I'm willing to accept that mine is a minority position.

Don't worry, ..I'm use to it.

Don't worry, ..I'm use to it.

I actually have to point out that it's nice to see an eloquent post out there, and I appreciate a well constructed, and well written argument... that's refreshing and a nice change from the garbage I throw out... 

Yertle wrote:

Personally, I don't see the need for another score/calculation*.  New players aren't going to care/know/figure out what some crazy calculation based on a slew of data, logs, and exponents means and veteran players aren't going to need some crazy calculation because for the most part veteran players are able to knew who is good and who isn't.

I'd much more see Achievements type of system rather than another score.

* = I see the coolness of Team Rankings, so I would be in favor of that new score.

 

+1

M57 wrote:

berickf,

Ok, I'll try my hand point by point..

First of all I believe that it's calculated on standard games and therefor introduces additional weight to standard games over tournament or team play.  By including it, standard play with three categories would weigh in at 60% of the overall value of the aggregate. 

Standard is good.. but assuming your point is valid (and I'm still not sure it is.. I prefer it to be 'standard' weighted), I prefer H-R over GR any day of the week.  Over the last two years, my H-R has gone over 2000 to under 1000 and back to over 2000 (it looks to be on it's way down right now), but my H-R has been incredibly stable, remaining  between 60 and 62 during the roller coaster ride.  Either my play is streaky, or perhaps my GR dips when I'm trying out new boards. Regardless.. on any given day, my GR may or may not have any significant value.  (I've argued for a GR moving average before)

Secondly, it's not a "traditional" or "intuitive" category for most of the new players to follow, whereas, the other four are quite easy to understand right off the mark. 

I would counter that anyone who is interested in the mechanics of such a score would be inclined to take the time (and be able) to understand the value of the H-R.  Consider things like the WHIP and OBP in baseball. Most fans/fantasy players value these.  You didn't make the argument that simpler is better, but it is somewhat implied here, and I would disagree with that sentiment.

Thirdly, if someone wins their first 4, 5, 6... 11, 12 games their H-ratings could be artificially elevated for quite some time and would be the top of that ranking simply because of the small sample size of games played.

I can invert the same argument.  It is precisely because the other components take so much time to reach equilibrium, and they have to get there from the bottom, not the top, that the H-R can serve to equalize this problem.  If it's really a problem on the first-few-games end, a minimum requirement could be implemented to minimize the problem ..as you described in your post.  But as far as I'm concerned, if someone comes on the scene and wins their first 5 or 10 games, they should appear on the radar.

I'm not saying your points are not valid.  In fact, I'm willing to accept that mine is a minority position.

Don't worry, ..I'm use to it.

Hey M57,

This one is going to start with a bit of a deviation off topic here and focus on M57's comparison of H-rating to GR, so, for those following the original topic, you could skim this one if it interests you, but it's not so crucial to the greater conversation of having an aggregate and is more a comparison of GR to H-rating.

Here we go!

As for what you have said about H-ranking M57.  Let me frame a new debate about it based on your comments this way.  You present the debate for h-ranking by stating the stability of your own as your GR has oscillated.  And, obviously that is simply a fact with regard to how both rankings are created to operate.  H-ranking is basically an average of all your games played but weighs them such as winning 50% of 1v1's = winning 33% of 1v1v1's = winning 25% of 1v1v1v1's = winning 20% of 1v1v1v1v1's, of which all of these are equivalent to a 50% h-rating.  By design, this is not as much a ranking as much as it is a predictive value based on your past game play on how well you have performed and might also perform in the future.  So, for evaluating the performance of players, it can have some predictive value, but, as far as "ranking" goes, it does not take into consideration the quality of competition, as in it does not punish you any more or any less for losing to weak competition or beating difficult competition.  Because it is a growing average, moreover, it becomes increasingly difficult to budge off the point it's at as the number of games played that went into generating it grow, and hence why yours has sat statically between 60-62% as that's exactly as it's designed to do.  Because GR, meanwhile, incorporates quality of competition you are A) much more susceptible to oscillation as your rating goes above or below the average rating of the quality of competition that you play against and B) require players to adapt to their growing GR if they want to mitigate these oscillations.

Both of these stats, h-rating and GR, are, as you like to put it, "game-able".

Since h-rating is an accumulation off all one's games played, the best way to game your h-rating requires a lot of dedication straight from your first game played.  That player who wishes to game it the most successfully should study the way that large multiplayer maps are played and then rely on patience and strategy to win a large number of those games.  For instance, the highest h-rated players are well into the 80's and if one wants to still collect championship points and play a large variety of maps it should conceivably be easier for them to maintain a high h-rating more so by mastering large multiplayer games on each map then expecting to maintain a 85% winning percentage playing 1v1's on many maps with varying degrees of luck built into them.  So, maintaining a high h-rating is most conceivable via mastering the large multiplayer game method.  Another way might be to play medium sized games on maps that are new and not so intuitive such that newcomers to the map might feel frustrated and drop out such that 7 players quickly becomes 3 or 4, then because you've studied the map and it's histories more then the others you'd have an advantage and still be able to win 65-75% of such games till the map is learnt by a larger base.  These same strategy, of course, would also work well to maintaining a high GR, and the second would also be great for growing CP on the sly!

When a GR is lower (1000-1800), it's ok to play 1v1's with all the other players in the rat-race.  +20, +18, -22, +24, -16...  Your GR score can grow even though it might not treat your h-rating so nicely.  But, when your GR rating gets higher (1800-2500) and those numbers turn into +7, +6, -35, +11, -21... well, now, to reduce your GR being all over the place you have to change the way you play.  A) you can  do what's good for your h-rating and master large multiplayer games, win 45-55% of them so that your GR +/- might look more like +89, -30, +101, -25... B) You can specialize in medium sized games, but, you now it'd be best if you win 50-60% or more of them such that your scores go something like +45, +55, -35, +50, -25 or, C) you can get picky about who you play smaller games with so that your opponent's GR's are within a certain percentage of your own so that the +20, +18, -22, +24, -16 can be maintained (assuming that you can continue to beat your GR peers at the same rate as before), or D) you need to play maps and settings where you can maintain a 1v1 winning percentage of 75% or more, like blind ants, or specializing on how to play many of these 1v1 maps such that you can beat 95% of the "regular joes" that walk in to check it out and go 55% vs those who know what they're doing.  Then your GR would go something like +6, +5, +20, +12, -30, +14, +6, +18, +6, -25, +12, +15, +11, +16.  So choosing one, or combining these strategies, becomes important to reduce the GR oscillations that you have found yourself susceptible to and to keep growing your GR until it simply becomes unsustainably high.  The whole point of GR in the first place is to pull the top players down as at some point the math will no longer support growth in one's score no matter what one tries, and the only way to maintain such a high score, at some point, would be to quit playing or be forced into the mathematical oscillations once the strategies for mitigating them are phased out by the gains no longer outpacing the losses.  I don't know what that score might be, but probably close to 4000 right now on WarGear.  On some other more established Risk emulation sites it's closer to 5000 because of unsuccessful players having lost much of their points then, having left the site, the "average" for active players is no longer the 1000 that everyone started with, but closer to 1500.  Higher average = higher ceiling.

There is another strategy worth noting that I forgot to include but is commonly practiced.  In larger games frequently higher ranking players will target lower ranked players over equally ranked players when the game becomes more mature in order to hedge their odds.  If you end up losing the game, better to lose 23 points then 39 points.

Ok, now to make some kind of point of all of that that is more relevant to the original topic.  Standard is important, yes, and that's why I gave it half the weight in my original proposal.  Now, to add onto my original points about h-rating, I will add on that H-rating has predictive value, but, does not take into consideration the quality of competition.  Meanwhile to control oscillations in GR, strategies must be employed to do so as I outlined above.  So, in fact, you could say that the two systems have an inverse relationship in that regard to whom you might actually like to play against.  H-rating encourages high h-rate-interested parties to play against people with a lower predictive value, lower h-rated players, that they know they can beat at a rate above their current H-rating.  In other words if one was a stickler for h-rating, you'd want to play against people who lose a lot (have a low h-rating).  Whereas GR encourages strategies to be employed that will weed out the lower ranked players such that luck will be less of a factor or that they play GR and h-rated peers, in order to mitigate their point losses and increase their point gains.  I don't think that most players actually follow h-ratings to select for low h-rate-opposition, but many do follow GR and do employ such strategies to compete against high level competition, so, it does seem that GR is the more commonly utilized statistic for how most players try to develop new strategies?

Another point that could be made is that since the h-rating system works by tracking a collective average that it is susceptible to and dominated by ones earlier body of work.  As-of-such, anyone with a lot of games under their belt can't really hope that even by choosing a different and effective strategy that they'll be able to change their own h-rating by enough to displace many above them.  For instance, if someone has played 400 games and has an h rating of 60%, then even if they changed their strategy such that they were maintaining a new h-rating of 75%, their chances of actually catching up to and passing other 70% h-rated players, let alone the 75% peers that they're now playing to the level of are slim to none.  This is because even after playing 100 games at 75% their average will still be dominated by their earlier body of work and they'll still be at 63% overall.  It's just like you say how you can't seem to move up from 62%... You've played so many games that you're now stuck there and if you wanted to "rank" h-rating, you can no longer budge from there in meaningful fashion in a reasonable time frame even if you had an WarGear Revelation and were playing at a h-rating of 90% for your next 100 games! The only way that someone could achieve a greater h-rating and compete with a new peer set would be to scrap their original WarGear membership and start afresh playing at their new h-rating under a new style of play.  Due to the competitive oscillating nature of GR however, you can't keep a good dog down and everyone always, always!, has the ability to modify their strategies to catch up to, maintain, and compete with the leaders of GR or any of the other 3 categories I originally included.  No one can create an unattainable position for themselves as is possible with a strict adherence to an h-rating strategy from the get go which results in a non-competitive system.  I tend to think that anything that is included in an aggregate ranking should be competitive such that any player at any time can work on different strategies and climb in that ranking and improve their aggregate.  In regard to this, h-rating fails the "is it competitive" sniff test.

As for "reaching an equilibrium", they are both reaching a different kind of equilibrium.  GR never reaches an absolute equilibrium.  It's more evolutionary and is always adapting to it's environment, this being the strategies being deployed to create your rank, and to maintain it at any given level of equilibrium.  Some environments are chaotic and allow for wild fluctuations and have a lower ceiling.  Other environments are planned and regimented, require homework, studying histories and maps, becoming a specialist in some aspects and allow for a greater stability and a higher ceiling.  With h-rating, alternatively, you get a highly static equilibrium which becomes continuously more constricted in it's ability to move as one's number of games played grows.  It's actually more a predictive measure based on one's past results and if we were to call it a ranking, we could call it the quicksand of rankings because the more you try and move the more stuck you get!  So I'd say that the point of rankings are to be continuously reactive to one's strategies that they are deploying to effect their rank and should be in a state of mobility to those who want to experiment, learn, improve, and get better.

Anyway, sorry I wrote so much again M57, it's my nature, but thank you for really getting me to think about h-rating as I had never thought about it so much until you put me to the task.  I would love to hear your feedback on everything I wrote here and I would also love to hear how you see such a static statistic adding value to the tough, yet competitive aggregate system, that I dream of bringing into the mainstream fold here.

Take care,

Erick

berickf wrote:

So, in fact, you could say that the two systems have an inverse relationship in that regard to whom you might actually like to play against. 

In the sense that these things can be gamed, including both GR an H-Rating in an aggregate sounds like a good thing, especially if attempts to manipulate one adversely impacts the other.

M57 wrote:

berickf wrote:

So, in fact, you could say that the two systems have an inverse relationship in that regard to whom you might actually like to play against. 

In the sense that these things can be gamed, including both GR an H-Rating in an aggregate sounds like a good thing, especially if attempts to manipulate one adversely impacts the other.

The inverse relationship is only in regard to who you might want to play against if one were following either statistic individually.  It does not imply that they would "balance" each other in the aggregate as they simply don't because one is a competitive ranking and the other is a static predictive statistic based on each players past performance.  I would say that the most important thing that I have come to understand while I was learning about the differences between GR and h-rating were how their inbuilt structure relates to their ability to encourage competitiveness. 

Since the majority of players don't even try to game h-rating I'd say that this is because of the quicksand effect it has and the drawback of not being able to move one's h-rating once it has been established, and on the non-competitiveness of h-rating to be effected by new strategies that one might try to employ in order to become the "top" in the category.  Quite frankly, the statistic becomes a "stick in the mud" for anyone that has played 200+ games and increasingly gets stuck the more games played.  As I said before, h-rating is designed more to act as a quality of competition predictor and less as a true ranking as rankings are specifically designed to be mobile and to be able to change leadership, which h-ranking is not designed to do.  What's the endgame in including a category into an aggregate which is virtually immovable.  It's not competitive and doesn't encourage any new type of gameplay to compete for one's place on it.  It's crazy just to think of what you with your current 61% h-rating and 627 games played would  need to do if you ever decided that you wanted to move up in such a category and to compete for a top spot in it.  At this juncture you would have to play at a 90% h-rating for your next 8473 games in order to pull into an h-rating tie with the current top rated player...  Think about that for a minute, 90% h-rating for 8473 games!  What strategy could you employ at this point that would allow for that? 

To pass the top GR player, however, due to quality of competition being part of the equation there, you'd quickly climb at first and then slowly climb once your GR rating pulled past 2200 and then creep once passing 2600.  If deploying the correct strategies that would see you maintain a top level of play while mitigating your losses, however, it is conceivable for you and anyone else in the WarGear Community, to pull close to and then compete with GR leaders for a top GR rank in much fewer games played.  Keep in mind that while you're growing your GR, the top rated players are struggling to keep theirs up.   Since growing one's GR is based on your starting point and is not weighed down by your previous body of work, depending on the game size a player decides to create/join they could significantly grow their GR in as few as 30-50 games played (depending on number of players per game), and then with solid strategies, could enter the top echelon and start competing for a top ranking.  It's competitive, and anyone can compete for it!  And, isn't that the point of an aggregate and all of it's component parts, that they be competitive, and that they inspire the development of strategies and ideas to compete for them?  And for an aggregate to encourage cross over play and an amalgamation of what sometimes may be competing strategies into one that will allow you to perform well across the aggregate?

Thanks a lot for all your feedback M57, it's invaluable.  Maybe some others can weigh in on the GR vs H-rating debate too so that we can get a better idea of how the WarGear Community feels about these statistics and their place in an aggregate ranking?

Thanks so much everyone for your participation and your support!

Erick

berickf wrote:

Since the majority of players don't even try to game h-rating I'd say that this is because of the quicksand effect it has and the drawback of not being able to move one's h-rating once it has been established, and on the non-competitiveness of h-rating to be effected by new strategies that one might try to employ in order to become the "top" in the category.  Quite frankly, the statistic becomes a "stick in the mud" for anyone that has played 200+ games and increasingly gets stuck the more games played.  As I said before, h-rating is designed more to act as a quality of competition predictor and less as a true ranking as rankings are specifically designed to be mobile and to be able to change leadership, which h-ranking is not designed to do.  What's the endgame in including a category into an aggregate which is virtually immovable.  It's not competitive and doesn't encourage any new type of gameplay to compete for one's place on it.  It's crazy just to think of what you with your current 61% h-rating and 627 games played would  need to do if you ever decided that you wanted to move up in such a category and to compete for a top spot in it.  At this juncture you would have to play at a 90% h-rating for your next 8473 games in order to pull into an h-rating tie with the current top rated player...  Think about that for a minute, 90% h-rating for 8473 games!  What strategy could you employ at this point that would allow for that? 

This is a very good point.  I concede that using H-ratings from the beginning of history creates a 'stick in the mud' stat, making movement in that category difficult, or even impossible.  In fact, now that you've brought this aspect of the stat to light (I don't believe it's ever been noted before), I'm actually less a fan of it.  Now, and perhaps even more so than GR, I'd love to see something like a 50 game moving average on the H-rating.

M57 wrote:

This is a very good point.  I concede that using H-ratings from the beginning of history creates a 'stick in the mud' stat, making movement in that category difficult, or even impossible.  In fact, now that you've brought this aspect of the stat to light (I don't believe it's ever been noted before), I'm actually less a fan of it.  Now, and perhaps even more so than GR, I'd love to see something like a 50 game moving average on the H-rating.

Hey M57,

A little clarification for me if you don't mind.

A moving average is simply a mathematical function that dampens the oscillations and makes the mountains less high and the valleys less deep while the underlying statistic is the same only less jumpy.  Since H-rating is an absolute average against all games played, did you make a typo and mean to write that their should be a moving moving average on GR to reduce the oscillations that you dislike, or are you saying that you would like to replace the absolute average of the current H-rating model to only have a 50 game memory such that it becomes more competitive?

Thanks,

Erick

berickf wrote:

Hey M57,

A little clarification for me if you don't mind.

A moving average is simply a mathematical function that dampens the oscillations and makes the mountains less high and the valleys less deep while the underlying statistic is the same only less jumpy.  Since H-rating is an absolute average against all games played, did you make a typo and mean to write that their should be a moving moving average on GR to reduce the oscillations that you dislike, or are you saying that you would like to replace the absolute average of the current H-rating model to only have a 50 game memory such that it becomes more competitive?

Thanks,

Erick

The latter..

No, it wasn't a typo.. I've always been a champion of a moving average on the GR as a stand alone stat to reduce oscillations, but for purposes of the aggregate, where "movement" should be encouraged (I agree with you here), I see no need for a moving average on the GR.

On the other hand, the static nature of the ever-deeping pool of H-rating data renders it all but useless when trying to determine a player's current ability, and especially that of a developing player.  Here, an H-R moving average is perhaps the best indicator of current strength, even more so than any of the other stats under consideration for an aggregate, which are contingent on the complete history of a player's career.

M57 wrote:

The latter..

No, it wasn't a typo.. I've always been a champion of a moving average on the GR as a stand alone stat to reduce oscillations, but for purposes of the aggregate, where "movement" should be encouraged (I agree with you here), I see no need for a moving average on the GR.

On the other hand, the static nature of the ever-deeping pool of H-rating data renders it all but useless when trying to determine a player's current ability, and especially that of a developing player.  Here, an H-R moving average is perhaps the best indicator of current strength, even more so than any of the other stats under consideration for an aggregate, which are contingent on the complete history of a player's career.

Hello again M57,

I agree, it should be able to have some movement to it, so, a most recent 50 game sample for h-rating would allow it to be mobile and would allow any player to implement a strategy over their most recent 50 games such that their roving h-rating could now be a competitive measure against everyone else.  If such was done to h-rating, then, I can see some potential there.  It would still be susceptible to the sample size error of new players who are still in their first 10 games or so, but like I said previously, it does seem to get some stability after 20 games, so a 50 game average would capture a fair h-rate standing but still provide for competitive movement in the standing.

Obviously, GR and CP are a given in any aggregate.  If a roving-h was included that would now give standard three categories in the overall aggregate and I'm not going to nit-pick over 50% vs 60%.  A consensus is what is most important here.  I still don't like how to be fair to the person who has put in their 50 games to generate their "true" roving-h, to include new players who have only played 5 games and won them all.  Or to include players who have played 2 winning games and then abandoned their WarGear accounts such that they'd sit ranked 1st in perpetuity.  To include the roving-h into the aggregate I do believe that to be fair that a minimum ten games played would need to be implemented for the aggregate standing, and, the jury is still out in my mind regarding how that makes me feel to exclude new players as they pay their dues for their first 10 games?

Another thing to consider is that the inclusion of a roving-h, to a degree, benefits those who have already built up their CP.  That is because building GR and h-rating are more complimentary to each other in my mind then building up one's CP, so once someone is happy with their CP rank they are more capable of deploying a strategy to build their GR and h-rating up in tandem over their most recent 50 games while avoiding maps that might adversely effect all the work that they have put into building up their CP.

I still stand strong that the strength of an aggregate is to encourage cross over play, so a tournament and teamplay ranking are crucial to the aggregate.  So, for all of you who are following this, by having 3 standard categories, does that dilute the importance of tournament and teamplay too much in the aggregate, or do you all feel that a 60/20/20 split is a fair weighting?  And, do you feel that new players will take the time to understand the roving-h, or does this overly complicate things?

One final thing to say.  Adding a 5th category makes the divisions of weight per component to be 20% each, which is still simple and intuitive for how the aggregate is constructed.  I would be more hesitant to add a 6th category, however, as a 16.67 split is not quite as tidy for the casual viewer.

M57, I would like to give you a sincere and appreciative thank you for working with me in a way to brainstorm how to be inclusive to what you deemed as important.  Some people would not be as gracious as you have been to all my nit-picking at the construction of the h-rating and what that meant for it to be a ranking, so, thank you for listening and adapting your thoughts to my criticisms.

Sincerely,

Erick

B, Like I said.  You have changed my perception of the H-Rating. I am less fond of it as a current indicator of ability.  I still think it would be great to include a 50 Day H-R MA to the mix, but I acknowledge that it is a complicated stat to understand.

 the jury is still out in my mind regarding how that makes me feel to exclude new players as they pay their dues for their first 10 games?

How is this materially different from the way in which a single CP is likely to take quite a bit more than 10 games to accumulate?  Consider that your average noobie plays a lot of WarGear Warfare, if only to familiarize themselves with the site.

I see it as a strength; inclusion of an H-Rating MA is one of the only ways that a new player can accomplish much in the way of upward movement in the aggregate.  

M57 wrote:

B, Like I said.  You have changed my perception of the H-Rating. I am less fond of it as a current indicator of ability.  I still think it would be great to include a 50 Day H-R MA to the mix, but I acknowledge that it is a complicated stat to understand.

 the jury is still out in my mind regarding how that makes me feel to exclude new players as they pay their dues for their first 10 games?

How is this materially different from the way in which a single CP is likely to take quite a bit more than 10 games to accumulate?  Consider that your average noobie plays a lot of WarGear Warfare, if only to familiarize themselves with the site.

I see it as a strength; inclusion of an H-Rating MA is one of the only ways that a new player can accomplish much in the way of upward movement in the aggregate.  

Ok M57,

What about players who play and win one or two games then never return to the site?  They'd be stuck at first while perhaps not having any intention of returning to equalize their roving-h while growing their other rankings.  Enough of those types and suddenly everyone that truly plays on WarGear is fighting for... the 220th place h-rating as it stands now.  If a player goes to the extent of executing an effective strategy for 50 games to be at the top of that ranking, then shouldn't they actually be at the top of the ranking, or close to it at least?  If you want to include from game one as you're making the case for, then do you implement an player inactivity exemption from the roving h, or, how do you see it?

Erick

At one point I was thinking about a global player inactivity penalty, ..which could possibly be forgiven when activity resumes.  I still don't see how the H-R problem you describe is all that different from that of the player who steps away as they hit one of the peaks on their GR ride.