Mad Bomber wrote: Why vs berickf does it show one of the two games we played, a tournament game?
http://www.wargear.net/games/view/129883

Tournament games counts for H2H. At least all of them do for me.

btilly wrote:

I would also love to see Cona Chris' opinion.

As for my opinion, changing CP to being based on GR works a lot better than I would have guessed.  (It also works out better for me than I would have guessed!)  And while it is natural to want to see a big reward for being #1, I'd like to vote for the simplest system that works.  And right now that looks like the (GR-1000)/50 system.

Unfortunately, CC hasn't visited the site since early Nov of last year and we might not ever hear from him. For all we know, he may be in favor of the change.  Yet, there is no question that his score is the most adversely affected.  Nevertheless, and GR-1000/50 fanboy that I am, I will still endeavor to argue a point against the proposed system on his behalf.  I looked at the top 14 current CP leaders.

The original purpose of the CP system was to reward the 'Champions' of each board. Looking at the above chart, it isn't too hard to see the inverse relationship between being #1 on boards and change in ranking. Note that players who have 6 or more #1 boards lose a net 8 places in the standings under GR-1000.  As a group, the also incur a -13% reduction in their scores.

Cona Chris hasn't visited since November? Must be a misunderstanding. I saw him entered in the Australian Beer tournament which hasn't started yet. His profile says he was latest online on Wednesday.

Litotes wrote:

Cona Chris hasn't visited since November? Must be a misunderstanding. I saw him entered in the Australian Beer tournament which hasn't started yet. His profile says he was latest online on Wednesday.

Hmm.. You're right - On his profile it says Last Visit: 23 March 2017.   I see where I got my information from though.   The Rankings page.

Right, the ranking page just has when they last won an open game, I think.

Let me float this - 

Suppose we left the BC points as is,

And replaced the Global Ranking with the (RS-1000)/50 thing.  (RS for "Ranking Score" for each board - it's not consistently named)

Then if we wanted to we could create another stat that added the two together -

I also was thinking of having two separate scores, but I kept trying to dissuade myself from even suggesting it - like all we need is another stat :/

I think Amidon37's idea of replacing GR with a (GR-1000)/50 aggregate is brilliant.

Scoring for individual boards simply remains in place - call them something like "Board Ranking."

But as I mentioned, I think the last thing we need is another stat, and I would caution against summing CPs and GR.  CPs are still messed up and need fixing - Summing just compounds the issue and creates a clusterfart of a Frankenstat.

One big advantage of establishing (GR-1000)/50 as the new GR standard is that it frees up CPs to truly be their own stat - much more independent of GR.  Here's an idea that I literally came up with in 30 seconds:

A board is not eligible to grant CPs until it passes the 1500 mark
Players cannot receive CPs until they individual pass the 1500 mark on a given board.
Each player whose score reaches 1500 or above is guaranteed at least 1 point.
The top score is dependent on how many players are above 1500 on a given board.

So if 5 players have scores above 1499, The top player receives 5 points, 2nd place receives 4, etc..

Of course there's room for all kinds of tweaking - Make the threshold 1250 if you want more points or add a coefficient to points awarded if you want the top scores to count for more, put restrictions on the number of places eligible based on GR, etc.

Well, maybe it's not a great CP idea, but regardless - I think we should start another thread proposing it that GR-1000/50 become the new GR. Actually, it's A37's idea - so I think he should start it.

Hey Everyone,

Love the conversation. Seems much more productive than 2 years ago.  I would ask that any posts not specifically about fixing CP be put in another thread so we don't get derailed.

I like the (GR-1000)/50 aggregate because it rewards Babba & Amidon for two separate but important reasons. Amidon for his breadth of game play and Babba for putting up some huge numbers that go relatively unrecognized CP wise. 

Again, I'm in favor of any simple reform that can actually be implemented whether 4x ing the current system or doing Amidon's proposal. I would argue against any special multipliers or anything that wouldn't K.I.S.S.

M57 wrote:

btilly wrote:

I would also love to see Cona Chris' opinion.

As for my opinion, changing CP to being based on GR works a lot better than I would have guessed.  (It also works out better for me than I would have guessed!)  And while it is natural to want to see a big reward for being #1, I'd like to vote for the simplest system that works.  And right now that looks like the (GR-1000)/50 system.

Unfortunately, CC hasn't visited the site since early Nov of last year and we might not ever hear from him. For all we know, he may be in favor of the change.  Yet, there is no question that his score is the most adversely affected.  Nevertheless, and GR-1000/50 fanboy that I am, I will still endeavor to argue a point against the proposed system on his behalf.  I looked at the top 14 current CP leaders.

	OLDCP	GR-1000/50	Differential	New Rank	#1 Boards
Mad Bomber	705	650	-8%	NC	4
Amidon37	451	603	34%	+2	2
Cona Chris	640	448	-30%	-1	22
IRsmart	379	403	6%	+1	13
Luieuil	335	398	19%	-1	4
Black Dog	449	389	-13%	-3	8
Sportlust	248	280	13%	NC	2
Hugh	234	269	15%	+1	4
Itsnotatumor	204	242	19%	+2	3
poloquebec	239	235	-2%	-2	9
AttilaTheHun	199	222	12%	+1	1
Andernut	155	206	33%	+2	2
smoke	213	195	-8%	-3	6
Yertle	161	175	9%	+1	1

The original purpose of the CP system was to reward the 'Champions' of each board. Looking at the above chart, it isn't too hard to see the inverse relationship between being #1 on boards and change in ranking. Note that players who have 6 or more #1 boards lose a net 8 places in the standings under GR-1000.  As a group, the also incur a -13% reduction in their scores.

I'm guessing you're playing devils advocate, but because it might move the conversation forward I'll bite.

First note, I have 6 #1's not 3, which might affect the analysis. 

Second Note, MB who has only 4 #1's is the current undisputed leader in CP, which is a pretty big outlier.

Third Note, that may have been the original intention, but how well has it actually played out?

Do the effects of the current system match the original intention?

How many of those #1's are just sitting because there is no real incentive for the board leaders to keep playing them.  I'm just sitting on 4 of my six (1500+) because I have no statistical incentive to play them. I've sat on one of them without a game for 3 years...  

How big a "champion" are you when no one bothers to even shoot for the title and the "champ" spends years in retirement.  Relatively few players will start a new game on a little played board, but many will join one, because why not.  I would argue the current system disincentives those people who would actually start those games from doing so to the overall detriment to the site.   

INAT, he just counted your 20-pointers. You have three of those. It takes a lot more time to check out if you're #1 on the board with less.

M57 wrote:

btilly wrote:

I would also love to see Cona Chris' opinion.

As for my opinion, changing CP to being based on GR works a lot better than I would have guessed.  (It also works out better for me than I would have guessed!)  And while it is natural to want to see a big reward for being #1, I'd like to vote for the simplest system that works.  And right now that looks like the (GR-1000)/50 system.

Unfortunately, CC hasn't visited the site since early Nov of last year and we might not ever hear from him. For all we know, he may be in favor of the change.  Yet, there is no question that his score is the most adversely affected.  Nevertheless, and GR-1000/50 fanboy that I am, I will still endeavor to argue a point against the proposed system on his behalf.  I looked at the top 14 current CP leaders.

	OLDCP	GR-1000/50	Differential	New Rank	#1 Boards
Mad Bomber	705	650	-8%	NC	4
Amidon37	451	603	34%	+2	2
Cona Chris	640	448	-30%	-1	22
IRsmart	379	403	6%	+1	13
Luieuil	335	398	19%	-1	4
Black Dog	449	389	-13%	-3	8
Sportlust	248	280	13%	NC	2
Hugh	234	269	15%	+1	4
Itsnotatumor	204	242	19%	+2	3
poloquebec	239	235	-2%	-2	9
AttilaTheHun	199	222	12%	+1	1
Andernut	155	206	33%	+2	2
smoke	213	195	-8%	-3	6
Yertle	161	175	9%	+1	1

The original purpose of the CP system was to reward the 'Champions' of each board. Looking at the above chart, it isn't too hard to see the inverse relationship between being #1 on boards and change in ranking. Note that players who have 6 or more #1 boards lose a net 8 places in the standings under GR-1000.  As a group, the also incur a -13% reduction in their scores.

I was conflicted about the same thing.  It feels reasonable that there should be a special reward for winning per board competitions.  But that is exactly the aspect of the old system that creates the most disagreement in any proposed system.   And is the feature that causes the problems with the most dissatisfaction.

I've therefore come to be convinced that we should consider this change a feature of the proposal.

I opened a new thread at http://www.wargear.net/forum/showthread/4640p1 to ask that this be implemented.

^ Which brings us back on topic.  Let's assume for argument's sake that the (GR-1000)/50 aggregate becomes the new GR and work on a more equitable CP system that can truly focus on Board Tops (Champions).

I think it's fair to say based on the number of tops Cona Chris has that he will at least stay at the #2 position if not move into #1 slot in the new or updated system.

There will still be some players who will not be able to get any GR points, but they will be very few. I feel that because the new GR system will be so inclusive and a much better indicator of overall participation and skill, it's much more acceptable if the CP system is exclusive.  Top 3 only can be on the table as far as I'm concerned.

For some reason, I'm enamored with the idea of using fibonacci.  1,1,2,3,5,8,13,21 makes for a nice top 8 scale, but from there I'm all out of ideas.

M57 wrote:

^ Which brings us back on topic.  Let's assume for argument's sake that the (GR-1000)/50 aggregate becomes the new GR and work on a more equitable CP system that can truly focus on Board Tops (Champions).

I think it's fair to say based on the number of tops Cona Chris has that he will at least stay at the #2 position if not move into #1 slot in the new or updated system.

There will still be some players who will not be able to get any GR points, but they will be very few. I feel that because the new GR system will be so inclusive and a much better indicator of overall participation and skill, it's much more acceptable if the CP system is exclusive.  Top 3 only can be on the table as far as I'm concerned.

For some reason, I'm enamored with the idea of using fibonacci.  1,1,2,3,5,8,13,21 makes for a nice top 8 scale, but from there I'm all out of ideas.

Why are you guys trting to change the current Global Ranking score? I assume that's what you mean by GR, or are you referring to the board score?

Abishai wrote:

Why are you guys trting to change the current Global Ranking score? I assume that's what you mean by GR, or are you referring to the board score?

Because the current GR score (the aggregate, not the individual board scores) is all but meaningless.  This is a question better asked on the (misnamed) Replace CP 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.

I stopped following this thread because I stated my objections and they still applied to the ever evolving proposed system for CPs. I was thinking it would eventually just come to a vote. Now that you guys have pivoted (dramatically) to wanting to change GR I would like to hear the reasons and what you guys are trying to achieve.

Abishai wrote:

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.

I understand, and I agree that it would be very premature to implement something like this after just a few days of posting.  I fully expect that it would and should take more than just a few months to let the the dialog play out before Tom would even consider implementing such a system.  Tom is VERY conservative with these kinds of decisions - which ultimately are his, not ours.  And trust me, our 'votes' are only considered in an advisory capacity by Tom. It's 100% his show.

I will address your question on the 'other' thread.

I was curious how GR/50 vs. GR/100+OLDCP stood up for WGWF.  Here's the table I came up with:

I still like the bigger spread between top positions that comes from adding the OLDCP back in.

I have  google sheet I can share if anyone wants to play with it.

Ozyman wrote:

I was curious how GR/50 vs. GR/100+OLDCP stood up for WGWF.

I still like the bigger spread between top positions that comes from adding the OLDCP back in.

I have  google sheet I can share if anyone wants to play with it.

@Ozyman, While I appreciate and agree with the notion that the top positions should enjoy the 'spread' in a category that includes the moniker "Championship,"  I would caution that (GR-1000)/50+OLDCP is a hybrid Frankenstat.

Consider that the (GR-1000)/50 aggregate may someday become the new GR (the Global Ranking in the main stats page).  Should this happen (and it would be a wonderful thing), there would be no need to have a CP rating that includes the "Global" GR ..and certainly as an addend.

With that in mind, it makes much more sense to update the OLDCP to something that is in synch with the present and scalable.  If you really want it to be a separate stat, it should avoid the inclusion of the GR stat at least to some extent (you could for instance use the GR to establish thresholds, but I would even try to avoid that).

The OLDCP is a mess. I've said it before, but it is arbitrary and scales poorly.   I challenge people to come up with something fresh and new that's scalable AND is largely independent of GR-1000/50.