We had another mega-thread on CP starting at http://www.wargear.net/forum/showthread/4631p1/Revisiting_CP.

Unlike past discussions, a clear consensus emerged among everyone who is still talking.  And that is the proposal to award per board CP's of (GR-1000)/50 (round down) for every board where your rating is at least 1050.  I would like to suggest that it be implemented.

Here is a quick summary of the change.  Most people with a lot of CPs will see moderate inflation (10-40% is common).  You can see http://www.wargear.net/forum/showthread/4631p5/Revisiting_CP for a chart for many specific players.  The system rewards a wider variety of styles of play than are currently recognized.  Many more people will get points in the CP system.  It hopefully will encourage people to play on a wider variety of boards.

There is, unfortunately, one big loser.  Cona Chris is a brilliant player who has played the current rules to perfection.  He would lose a lot of points from the change.  He will probably slip from #2 to #3 in the CP ranking.

Thingol, MB and I clearly stated that we are not in agreement with this change on that mentioned thread.  Just because we haven't posted in the last couple days doesn't mean our opinions have changed. I think it is a little decietful to imply that there is a clear consensus on that thread.  I for one found keeping up with all the updates on the new progressive system not worth my while because the objections that I previously stated still applied.

I think this should at least go to a vote, with a few of the suggestions getting voted on along with a "No change" option.

I think btilly misnamed this thread - It should be..

Replace GR with Amidon37's proposal

Silence does not equal agreement. In the interest of fairness to the whole community any system being proposed should come to a vote, leaving the thread open for at least a couple months so that any less active players will have the chance to vote. I also think that any new system should hold at least a majority of the votes to get implemented. I would also like my proposed system to be on the ballot in that it was a modification to the current system that some players found more agreeable.

M57 wrote:

I think btilly misnamed this thread - It should be..

Replace GR with Amidon37's proposal

Btilly says CPs multiple times in his post though.

Abishai wrote (in another thread):

I would say that the current GR score is pretty straightforward. It just shows how well you do overall on public games. What meaning would you like the GR score to have? Not trying to be a jerk, I'm really trying to understand.

The problem with the "Global" GR score is that it is too temporal.  My GR has shot up to 2200 and then down to 1200 and then bounced back up above 2000 over the course of just one year. This can happen for a number of reasons. It can go up as you work your way up a board you are focused on, and it can plummet as you try and learn new boards. It's one big roller coaster for most players and has very little meaning as indicator of overall achievement, has absolutely no value in terms of rewarding players for diversification In fact if anything, it's a deterrent to diversification.

That doesn't mean we want to get rid of the GR - on the contrary.  The GR is a very good indicator of players' abilities on a board by board basis.  We are simply suggesting that the new "Global" GR rating be an aggregate of all positive GRs on all boards.  This will make it MUCH more stable, and give it a tendency to move in one direction over time ....UP!

Abishai wrote:

Thingol, MB and I clearly stated that we are not in agreement with this change on that mentioned thread.  Just because we haven't posted in the last couple days doesn't mean our opinions have changed. I think it is a little decietful to imply that there is a clear consensus on that thread.  I for one found keeping up with all the updates on the new progressive system not worth my while because the objections that I previously stated still applied.

I think this should at least go to a vote, with a few of the suggestions getting voted on along with a "No change" option.

Yes - me too.  I don't particularly like this option, and have stopped following the thread carefully, because I have other more pressing issues IRL.

Abishai wrote:

M57 wrote:

I think btilly misnamed this thread - It should be..

Replace GR with Amidon37's proposal

Btilly says CPs multiple times in his post though.

Yes - you are right - and maybe he means CPs.  But he refers to Amidon37's proposal in the title and A37's proposal is for (GR-1000)/50 to replace the GR - not CPs.

My original proposal was for (GR-1000)/50 to replace CPs.  I think A37's idea is brilliant and I suggested we start a thread devoted to GR.

I know I'm a newb here, but GR should definitely be an indicator of a player's overall performance, not on how well he has specialised in a few maps. There's BR for that. Not to mention that, as has been said, exploring new maps should be encouraged, or at least, not stifled.

With that said, I'm not on par with the formulas used, just my 2 cents.

redshift wrote:

I know I'm a newb here, but GR should definitely be an indicator of a player's overall performance, not on how well he has specialised in a few maps. There's BR for that. Not to mention that, as has been said, exploring new maps should be encouraged, or at least, not stifled.

So how do you feel about taking the current GR, which is based on game by game performance, and replacing it with a (GR-1000)/50 aggregate?

I really like your idea of renaming each individual board's rating "BR."  It makes things much clearer. That would make it such that the new GR = SUM{(BR-1000/50)}

With that said, I'm not on par with the formulas used, just my 2 cents.

You mean the formula currently used for both GR and CPs, right?  ..I would agree.

I really like your idea of renaming each individual board's rating "BR."  It makes things much clearer. 

Yes!    If nothing else comes from these threads, I'm calling it a win.

Board Ranking not Board Rating. That is something else.

Replacing CP with (GR-1000)/50 (round down) was actually my proposal not Amidon37's. It should have more accurately been called (Score-1000)/50 (round down) as that is what it is called on the Board Rankings page or (BR-1000)/50 (round down) if you want to use that abbreviation. I apologize for the misleading variable name.

Amidon37's proposal was instead of replacing CP with (BR-1000)/50 (round down) that we instead replace GR with (BR-1000)/50 (round down).

Since the thread (http://www.wargear.net/forum/showthread/4631p1/Revisiting_CP) has primarily been discussing CP, all of the comments prior to 3/26 when Amidon37 brought up GR have been about replacing CP. Of those it appears that the (BR-1000)/50 (round down) has been the most popular.

The GR conversation, however, just started. So I don't think it would be safe to consider that to have any sort of consensus at this point.

Korrun wrote:

Replacing CP with (GR-1000)/50 (round down) was actually my proposal

@Korrun: ("CP OR BR"-1000)/(some constant) is actually my idea going quite a few years back (Option I), but I reintroduced it in the recent CP thread in an altered form.  You suggested that it essentially be considered in it's original form (of which you may not have been aware of), where it gained quite a bit of traction, but still runs into loggerheads with those who believe the score (CP) should disproportionally reward the top players of each board, hence suggestions such as (BR-100)/50+OLDCP.  I understand those sentiments - The word after all is "Championship."

Amidon37's proposal was instead of replacing CP with (BR-1000)/50 (round down) that we instead replace GR with (BR-1000)/50 (round down).

This is true - and I think that it's a great idea.  It beautifully addresses the flaws of the current GR system leaving the current CP system alone.  Basically, in trying to address and update/replace the CP system, we inadvertently came up with a really good idea for replacing the current GR system.

Since the thread (http://www.wargear.net/forum/showthread/4631p1/Revisiting_CP) has primarily been discussing CP, all of the comments prior to 3/26 when Amidon37 brought up GR have been about replacing CP. Of those it appears that the (BR-1000)/50 (round down) has been the most popular.

However, and as I mentioned, unable to reach a consensus because it doesn't really address what CPs are about.  And I now agree with that sentiment. I no longer think (BR-1000)/50 is an appropriate algorithm for determining CPs.  It is MUCH better suited for replacing the GR.

As an aside, I would comment that if (BR-1000)/50 does ever replace GR, I will have much less personal interest in the CP conversation as I think the new GR will be a very strong stat rewarding all aspects of play on the site. I think that because of the CPs disproportional nature it is a less relevant indicator for "most" players on this site. If it is 'scaled' or frankenstatted.. it will continue to be little more than a game inside the game.  I have never held it in high regard, but unfortunately I have had to consider that (flawed as it is) it has been one of the better indicators the site has.

The GR conversation, however, just started. So I don't think it would be safe to consider that to have any sort of consensus at this point.

I agree.  Consider (BR-1000)/50 (rounddown). ;)

@Korrun: ("CP OR BR"-1000)/(some constant) is actually my idea going quite a few years back (Option I), but I reintroduced it in the recent CP thread in an altered form.  You suggested that it essentially be considered in it's original form (of which you may not have been aware of), where it gained quite a bit of traction, but still runs into loggerheads with those who believe the score (CP) should disproportionally reward the top players of each board, hence suggestions such as (BR-100)/50+OLDCP.  I understand those sentiments - The word after all is "Championship."

True. I was not trying to lay claim to the initial idea of how to calculate CP, just trying to differentiate it from Amidon37's proposal. There seemed to be some confusion between the two.

However, and as I mentioned, unable to reach a consensus because it doesn't really address what CPs are about.  And I now agree with that sentiment. I no longer think (BR-1000)/50 is an appropriate algorithm for determining CPs.  It is MUCH better suited for replacing the GR.

As an aside, I would comment that if (BR-1000)/50 does ever replace GR, I will have much less personal interest in the CP conversation as I think the new GR will be a very strong stat rewarding all aspects of play on the site. I think that because of the CPs disproportional nature it is a less relevant indicator for "most" players on this site. If it is 'scaled' or frankenstatted.. it will continue to be little more than a game inside the game.  I have never held it in high regard, but unfortunately I have had to consider that (flawed as it is) it has been one of the better indicators the site has.

Replacing GR with a different formula should probably be done in another thread? It seems like something that would be worth its own discussion. One of the reasons that fixing CP makes sense is that is what rankings are currently based on. Are you proposing that rankings would be based on the new GR score? Or still CP?

Korrun wrote:

Replacing GR with a different formula should probably be done in another thread? It seems like something that would be worth its own discussion. One of the reasons that fixing CP makes sense is that is what rankings are currently based on. Are you proposing that rankings would be based on the new GR score? Or still CP?

Right now, I don't really have a strong opinion about rankings.  However, I'm pretty sure that if (BR-1000)/50 became the new GR, it would be my choice for Rankings (if the site was to base rankings on one number).

Optimally, I would love to see (BR-1000)/50 become the standard for all categories of games - Ranked - Team - Tournament, etc.  That way you could have a ranking system based on all categories that is fairly weighted.

CPs in my mind are a completely different category, in many ways incompatible with the above ranking system (though BR could be used in some ways to determine CPs - E.g. by setting a cut off, etc.).

If I was starting this site from the ground up, I might suggest that CPs should be a tournament category. "Champions" win tournaments!  I would suggest that they be awarded to the first, second and third place finishers and be based on the number of participants in each tournament.  Unfortunately, I think the site's hands are tied by the semantics of words like "Champion," as it has tied that term to board strength rather than the ability to win a tournament.

And lastly, I think it's silly that tournament play is not considered in determining BR.  It has made it such that some players use tournaments to learn a board because they know it doesn't 'count,' thereby inadvertently inflating the scores of players who play tournaments to win.

M57 wrote:

Korrun wrote:

Replacing GR with a different formula should probably be done in another thread? It seems like something that would be worth its own discussion. One of the reasons that fixing CP makes sense is that is what rankings are currently based on. Are you proposing that rankings would be based on the new GR score? Or still CP?

Right now, I don't really have a strong opinion about rankings.  However, I'm pretty sure that if (BR-1000)/50 became the new GR, it would be my choice for Rankings (if the site was to base rankings on one number).

Optimally, I would love to see (BR-1000)/50 become the standard for all categories of games - Ranked - Team - Tournament, etc.  That way you could have a ranking system based on all categories that is fairly weighted.

CPs in my mind are a completely different category, in many ways incompatible with the above ranking system (though BR could be used in some ways to determine CPs - E.g. by setting a cut off, etc.).

If I was starting this site from the ground up, I might suggest that CPs should be a tournament category. "Champions" win tournaments!  I would suggest that they be awarded to the first, second and third place finishers and be based on the number of participants in each tournament.  Unfortunately, I think the site's hands are tied by the semantics of words like "Champion," as it has tied that term to board strength rather than the ability to win a tournament.

And lastly, I think it's silly that tournament play is not considered in determining BR.  It has made it such that some players use tournaments to learn a board because they know it doesn't 'count,' thereby inadvertently inflating the scores of players who play tournaments to win.

Yeah, it's funny. I joined this game with a group of friends from a tennis forum and we started by playing private games, then slowly (some faster than others) started to play public games and then tournaments. In my view, tournaments were the holy grail of WarGear and playing regular games served as preparation for them. Turns out the system points in a different direction.

I think GS = SUM((BS-1000)/50) would be fine. Some boards are more played than others and harder to achieve top rank I guess, so maybe it wouldn't be fair to apply a function with diminishing returns. I guess a linear function is fine. How did you guys reached the 50 figure, btw?

As for CPs, as I have said before it should be an yearly competition from N best results in tournaments, maybe the best 12 which averages a tournament per month, maybe less, not sure. Maybe knowing how many tournaments Babbalouie plays per year could give us a clue. And should individual and team tournaments be counted seamlessly or have a Championship for each? Or maybe we could have both separated and combined.

redshift wrote:

I think GS = SUM((BS-1000)/50) would be fine. Some boards are more played than others and harder to achieve top rank I guess, so maybe it wouldn't be fair to apply a function with diminishing returns. I guess a linear function is fine. How did you guys reached the 50 figure, btw?

50 is simply a convenient constant that ends up creating numbers surprisingly close to current CP numbers, if slightly inflating them.

Going back a few years, the original "Option I" proposal was simply for SUM(GR-1000) for any and all boards >1000.  Applied now, this would result in high scores upwards of 50,000 for the top players.  I could care less what the constant is. It doesn't change the metric in any way.  A denominator of 1 makes the most sense to me.  Full disclosure, Option I was my proposal.

The reason i didn't propose boards under 1000 be included was because the original proposal was for Option I to be for CPs, not GS. But really a true GS should include play on ALL boards. Of course that would throw negative numbers into the mix ..and I'm guessing it would probably run up against some criticism for that reason.  Personally, I have no problem at all with it.

M57 wrote:

redshift wrote:

I think GS = SUM((BS-1000)/50) would be fine. Some boards are more played than others and harder to achieve top rank I guess, so maybe it wouldn't be fair to apply a function with diminishing returns. I guess a linear function is fine. How did you guys reached the 50 figure, btw?

50 is simply a convenient constant that ends up creating numbers surprisingly close to current CP numbers, if slightly inflating them.

Going back a few years, the original "Option I" proposal was simply for SUM(GR-1000) for any and all boards >1000.  Applied now, this would result in high scores upwards of 50,000 for the top players.  I could care less what the constant is. It doesn't change the metric in any way.  A denominator of 1 makes the most sense to me.  Full disclosure, Option I was my proposal.

The reason i didn't propose boards under 1000 be included was because the original proposal was for Option I to be for CPs, not GS. But really a true GS should include play on ALL boards. Of course that would throw negative numbers into the mix ..and I'm guessing it would probably run up against some criticism for that reason.  Personally, I have no problem at all with it.

Me neither. GS should definitely count all boards.

But this problem of some players not wanting to try more boards to not lose GS more and more tells me a Championship based on tournaments should be the system that ranks the best players.

Just recently read on the "Revisiting CP" thread that some top players have hardly ever played against each other. This is ridiculous, to say the least.

redshift wrote:

But this problem of some players not wanting to try more boards to not lose GS more and more tells me a Championship based on tournaments should be the system that ranks the best players.

I'm not sure if we're talking about the same thing here, but I would note that while there can be a GS cost associated with learning a new board, it would be very short lived in a 'true' SUM(GS-1000) system (including negative numbers).  Most players can quickly get their scores above 1000.  Sure, a few boards may fall into the low 900s or even 800s, but with the vast majority of players, these scores will easily be off-set by the number of boards that quickly make their way above 1200.