I've crunched the numbers for everyone who has participated on this board so far - with some commentary.

(GR-1000)/50 revised with ROUNDDOWN..

Mad Bomber 705 -> 650  (-55)    -8% MB has 124 boards in the money)

Cona Chris 640 -> 448     (-192)   -30% (CC has 48 boards in the money)

Amidon37 451 -> 603      (+152)  +34% (A37 has 153 boards in the money!)

IRsmart 379 - > 403       (+24)    +6%

INAT 204 -> 242            (+38)    +19%

SquintGnome 103 -> 125 (+22)   +21% (SG only has 15 boards in the money, but over a third have GRs of 1500+)

Ratsy 79 -> 113              (+34)   +43% (Rasty has 46 boards in the money)

Ozyman 79 -> 91          (+12)     +15%

Babbalouie 68 -> 147     (+79)     +116% (Scores of 2968, 2705, and 2285 on Battle of Bladensburg, WGWF and Civil War respectively are handsomely rewarded)

AfroDaby 68 -> 80         (+12)    +18%

Thingol 56 -> 78            (+22)    +40%

Korrun 54 -> 62             (+8)     +15%

M57 49 -> 73               (+24)     +49%  (Full disclosure, I make out pretty well - Did not see that coming.)

Abishai 46 -> 40           (-6)        -13%

Btilly 43 -> 82               (+39)    +91% (Boards like Colossal Crusade, where previously he received 0 CPs, account for more than half of his gains)

Terminatorr 30 -> 54     (+24)     +80% (1952 GR for WGWF (worth 0 in the old system) accounts for almost all of his gains.)

Johasi Vidad 9 -> 24      (+15)     +267%

Litotes 2 -> 19               (+17)    +950% (Newer players get rewarded for their work quickly)

Generally speaking there looks to be about 20 to 30% inflation going on compared to the old. I think that will play well in Peoria. Most people would prefer more points, right?  As far as I can tell, the system generously rewards diverse play (A37's 150+ boards in the money is an exemplar), but players that specialize on the popular boards with lots of points do just fine.

I haven't said much, but want to chime in again since things might be actually going somewhere productive. (after having followed this post and in the past read through the massive other CP discussion threads. It would be cool to see something agreed upon and implemented even if it isn't the "Perfect" CP system as that most likely won't be agreed upon or implemented by Tom)

Thank you M57 for running that as I think it clearly shows (as you mentioned in your comments) most of the important things in a revised system we generally want.

>The majority get more points, but not drastically so.

>New players are rewarded with more earlier on such as Litotes and even myself.

(as I'll mention I never highly pursued CP as it just seemed to be difficult to achieve so I focused more on tourney play [which I've dropped in] and team-play which isn't rewarded [to discuss another time], but I enjoyed.)

>More players can gain from both popular and unpopular boards.

>Encourages diversity of play on more boards.

>And scales over time.

Only questions that come to mind for me would be:

How do the top players that take a hit feel?

Though even that shows just between MB and CC that diversity will be more rewarded as you head for the higher scores.

I know it would take a bit more time but what does the [(GR-1000)/50 revised with ROUNDDOWN + Old CP] look like just for comparison?

As I liked the sound of that one as it would have the + to board diversity as well as rewarding the champions of both popular and unpopular boards keeping the fight for the very top alive.

I really think one of these has the potential to work for all of us in general and maybe even get implemented.

Which would be awesome.

It would be nice to hear what the top players feel about it (especially CC who takes a big hit), but I think it is a good sign how the scores are working. Mad Bomber looks like he has a good chance of staying in the #1 spot and the #3 person with diversified play passes the #2 player who is the specialist. It seems like with our idea of what CP should be rewarding, that the person with good scores on 153 boards should have more CP than a person with good scores on 48 boards.

If M57 has time to crunch the numbers with the +OldCP I would love to see what that looks like as well. One advantage to the +OldCP is that even if orders switch no player would actually lose CP, which feels good psychologically.

Korrun wrote:

Ozyman wrote:

I like that, but feel the position bonus gets a bit minimized - maybe 100 for the denominator for simplicities sake?

So someone in 1st with 1500 gets 25 CP, and someone in 1st with 3000 gets 40 CP. 

This formula seems to hit all the desirements:

-) Never needs to be readjusted, since it continues to scale to infinity.

-) Still encourages play on multiple board

-) newbs can quickly get some CP - earn your first point at 1050 on any board.

-) Relatively simple to implement - not a huge deviation from current ranking.

Anything it's missing?

I'm fine with changing the denominator to 100, but that would earn your first point at 1100 instead of 1050. I just picked 50 for its consistency with the current system. Your above numbers sound like reasonable scores to have for those positions.

And yes, I think this system covers all of the bases.

I was thinking we'd round so (1050-1000)/100 = 0.5 rounds to one point.  Or maybe you need to get to 1051 => .51 => 1 point

I want to say that although MB & CC would take a hit, I bet that would be temporarily, as they have both played strategically, so of course any change to the rules would penalize them, but once they adjust they'd probably be right back where they were.

Johasi Vidad wrote (and Korron also asked):

I know it would take a bit more time but what does the [(GR-1000)/50 revised with ROUNDDOWN + Old CP] look like just for comparison?

It's not that hard to imagine, just add the two numbers in the chart.  Example:

Mad Bomber 705 -> 650 = 1365 For most people that means more than double the old score, or less than double the new.  The more I think about it, given what looks like parity in the (GR-1000)/C system, there's no need to tack on the old one - it just makes the whole thing that much harder to understand. 

Ozyman wrote:

I was thinking we'd round so (1050-1000)/100 = 0.5 rounds to one point.  Or maybe you need to get to 1051 => .51 => 1 point

I'd also be a fan of the idea of dividing by 100; it's dead simple, but most of the resulting scores would be lower and I think there would be blowback from people who would assume that the new system penalizes them. On balance, I think it's wiser to go with 50 as a denominator, but I'd be happy to go with whatever - it's just a denominator. It doesn't really change the metric functionally - It's just about perception.

Ozyman wrote:

I want to say that although MB & CC would take a hit, I bet that would be temporarily, as they have both played strategically, so of course any change to the rules would penalize them, but once they adjust they'd probably be right back where they were.

I agree.  Both MB and CC played the system masterfully, scoring just enough on their target boards to get the max amount of points.

In MB's case his 'hit' is marginal; he pretty much plays to the 1500 to 1700 mark depending what nets him the most points.  MB does pretty well in the transition because his strategy forced him to diversify and he ended up playing quite a few boards as a result.

CC, on the other hand simply did whatever it took to stay on top on a board by board basis. He is 1st 2nd or 3rd on every single one of the boards he plays that has a GR over 1500, and he's first on 2/3 of them.  The result of that strategy has him constantly working on the boards he already plays, and not having as much time to work new boards into the mix.

While it's easy to simply charge that MB and CC have been gaming the system, the reality is that they did not create 'that' game. They deserve credit for finding the best ways to 'play' it - and it's interesting to note that they both went about it differently. As I mentioned MB's (GR-1000)/50 score didn't really take that much of a hit because part of his strategy was to add new boards to his list as he capped out on others.

We'd have to ask him if this is true, but I'm going to speculate that Amidon37's meteoric rise is a result of him not really paying attention to CPs at all, and just playing a lot of boards. I.e, not playing the "CP game."

In any case, I don't doubt that MB and CC could easily adjust their strategies and enjoy the same benefits the proposed system offers as much as everyone else.

Ha, I think MB, CC and Amidon are going to end up with very high scores no matter the scoring system. The 'bastards' are just darn good players.  

Thingol wrote:

Ha, I think MB, CC and Amidon are going to end up with very high scores no matter the scoring system. The 'bastards' are just darn good players.  

Agree, you're never going to find a system that leaves then far down the list.

But it's easy to find a scoring system that sees them surpassed by berickf. 

How many focus on global ranking? I know several top players don't care about theirs at all, happily playing new boards to learn and sacrificing global points by the hundreds. 

I would prefer a straight round down method to the score. I think it makes it easier to eyeball. Once you get to 50 or 100 you get to the next point. Otherwise you'd have to remember/explain to new people the rounding system. Why would we want to get our first point at 1051? Maybe 1057... Of course I am fine with whatever is consensus.

Since it sounds like adding the old CP is sounding like it is not gaining popularity, what about something simpler? If you are in the top 3 or 10 you get your score doubled? Or a flat +10? +10 for top 10.

Korrun wrote:

I would prefer a straight round down method to the score. I think it makes it easier to eyeball. Once you get to 50 or 100 you get to the next point. Otherwise you'd have to remember/explain to new people the rounding system. Why would we want to get our first point at 1051? Maybe 1057... Of course I am fine with whatever is consensus.

The stats I posted are all ROUNDDOWN.  You don't score until you reach the threshold.  1050-1099 scores 1 CP, and 1100-1149 scores 2 CPs.

 If you are in the top 3 or 10 you get your score doubled?

MB would dominate on such a system - his entire score would almost double because his strategy was not to play many boards, but rather to dominate the ones he did play.  

Or a flat +10? +10 for top 10.

I think mixing 'goals' just makes things more confusing and risks players gaming the system.  If the 'goal' is to encourage play for the top positions, why not just have a separate CP system just for that? 10 points for #1, 9 for #2, 8 for #3 etc and be done with it.  No way to game that - you are either the top or you're not. Over the course of these discussions.  Actually, a straight CP race based on that metric sounds like fun and is more like what CPs should have been in the first place.

I believe calculating the amount of points awarded for top seats based on GR in any way is a recipe for imbalance and is more likely to encourage 'gaming.'  Consider the player who (not being concerned about the 'race') manages to get to 11th position on a higher than normal percentage of boards. His scores will be misleading for those of us who want to see a system that is a truer indicator of overall skill.

(GR-1000) admittedly creates more of an aggregate type of score, but it is a simple and pure summative system that equitably rewards high scores while encouraging diverse play.

What i like most about this (GR-1000)/50 system and not having bonuses for the top spots is this eliminates the effects of the "legacy" players in the individual board rankings. There are plenty of boards where the top spots are taken by players with high scores so you need 1300+ to earn a point on the board.  No rewards for playing those - and those tend to be the good/fun boards that got many of us hooked on the site.  (e.g. BattleUSA or Infection). 

This system would easily allow for a "top players of all time" ranking list versus a "top players who have played at least a game in the last year" type lists.  That's what I think would be the best outcome for this discussion.

Amidon37 wrote:

What i like most about this (GR-1000)/50 system..

..is that you make out like a bandit. ;))

But seriously - you would deserve it. You are a true ..if not exceptional journeyman player, and your bump clearly demonstrates how well the system values diverse play ..even as it values high scores. It's quite amazing to look at your stats.  You are above 1500 on only a half dozen boards, and two of those (WGWF and "World War") garner you 0 Points under the old system.

This system would easily allow for a "top players of all time" ranking list versus a "top players who have played at least a game in the last year" type lists. 

I'm not sure I understand what you're saying, but because the "top" spot doesn't really mean anything in the GR-1000 Aggregate, retiring from the site etches your score in stone - it won't recede over time ..the result of players overtaking you.

berickf  73 -> 154  (+81) +111%

berickf focuses more on tournament play if I'm not mistaken, and it pays to be more of a specialist in that venue, but his abilities are certainly much better recognized by GR-1000, and seeing that might encourage him to shift his focus a bit.

Since it sounds like adding the old CP is sounding like it is not gaining popularity, what about something simpler? If you are in the top 3 or 10 you get your score doubled? Or a flat +10? +10 for top 10.

I'm still a strong fan of including the old CP.   Couple that with a divide by 100, and scores will still be in the same ballpark.  

I don't like the idea of not rewarding position at all.  I think you should get something more for being #1 vs #2 on a board, even if your two scores are within 50 points of each other.

The 4 rank system is the best of all worlds. Players that want to continue to pursue CP's may do so. Players that want to pursue GR may do so.  Players that want to excel in team play may do so. And players that want to be top tournament players may do so. Whatever strikes your fancy. I personally do not care about CP's and it seems obvious that berickf (whom I believe is the best all around player on this site, who now seems to be semi-retired) does not either.

Perhaps some players want to be well rounded and pursue 2, 3, or all 4 areas. Cona Chris even suggested a 5th rank for tournament team play. This multi-rank system will renew competition and stimulate interest.

Tom wanted "Viper" to design the medals, but Viper had been a no show. This has delayed the implementation of the 4 Rank System. Is there someone that can take over for Viper so that this can finally be implemented? It has been a long time coming. Once up and running, modifications may be necessary. 

As regards to berickf possibly being the best all around player, I looked up H2H stats just for fun. 

I was astonished to see that his H2H with Mad Bomber is 2 win 0 tie 0 loss. They've only twice been in the same game? Amazing. Much the same with Amidon37, +2 -0 =1. Talk about pursuing different strategies.

He's +15 -1 =0 against Cona Chris.

Of those he's played at least 10 games with only Ceggon has more wins than losses against him -  6 to 5.

Babbalouie wrote:

The 4 rank system is the best of all worlds..

I would not disagree that the Star System is the best way of dealing with the current systems.  But those individual systems are flawed.

The GR is the cornerstone for all four, even CPs are based on the GR.  The GR by itself is notoriously fickle (mine has varied between 1200 and 2200), and even gameable for that matter. Something like TrueSkill is what the site would do well to replace it with but that has proven to be very complicated.  So we're stuck with what we have. That said, what GR does have going for it is that it is simple and scalable, and it's much less gameable when you spread it across multiple boards with metrics like CPs.

CPs on the other hand, which have been touted as the site's most coveted rank, are a mess. They are arbitrary by nature, gameable, and elitist, which is bad for the site on a number of levels. The irony is that because the CP is an aggregate GR-based score, it has the potential to be everything it has been touted to be.  It tames the fickle and highly variable nature of the Global Ranking score by spreading it across many boards.  GR-1000 is 100% based on the GR, and by spreading it across all boards to create the new CP system, there would be no need to include the GR itself in the Star system. GR simply becomes the engine that informs CPs.

If you take that paradigm to its logical end, the site should really have CPs as the metric for all of the Star categories, all of them being informed by the GRs of every board played in each respective category.

I'm a fan of the Star system.  But that's a blunt macro system. Again, it's only as good as the components that feed it, and if they were all derived with the same algorithm, it would be truly equitable.

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.

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

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