Page 6 of 6

Re: Ranked Draws & Rank-Based Scoring: Fibonacci-Diplo

PostPosted: 09 Jun 2018, 05:03
by NoPunIn10Did
boldblade wrote:Hmmmm I think you bullied him no pun. Way to go.

:evil:

Re: Ranked Draws & Rank-Based Scoring: Fibonacci-Diplo

PostPosted: 09 Jun 2018, 09:47
by super_dipsy
I think what V was pointing out was that a lot of the background justification and assumptions around this idea are from the Tournament / F2F world. It absolutely makes enormous sense to use this in Dip tournaments, because by definition most real life tourneys have a fixed (and small) number of rounds and limited time, so you have to have a way of scoring that does not require the game to actually end. Using drawbased scoring does not have enough definition in a short length time. But the reason we have resisted it in the online form of the game is that we do not have the artificial stop point. We can allow a game to play out.

However to be honest our resistance has always been based on the simplistic point that if you start playing a game of Diplomacy with a background plan to try to last as long as possible (thereby increasing your return if you use a time served system) or to ensure that win or lose you get to a decent SC count, you are bound to play differently. Just a a silly example, but it seems to me that this system encourages Carebears; carebears typically make long game alliances but always risk their 'friend' stabbing them at the end, but in this sytem that is cool because you still end up highly placed (provide people stop the solo of course). The key here though is that although I believe it is true that changing the scoring system will affect play, what is not necessarily so clear is whether it will be worse or better from an enjoyment point of view. It is easy to see holes in a different system, just as it is to see holes in the existing one. I confess I am still nervous because I am not sure I believe the assertions of its key values; more solos, less draw whittling, more enjoyment. But that does not mean it is not worth trying.

This is the tricky bit though. I completely understand the issues of trying to get something new tried out when people have to read up and volunteer; we alaready know only a tiny % of our players use the forum anyway. But making it an option that is included in the current rating system may be a step too far too. If as some suspect it majorly changes play/outcomes, there is a danger it will distort the ratings. But making it available as an option only for unranked games again has some of the same drawbacks as making it a forum-based recruitment thing; it can be hard to get a critical mass because for some reason most players want to play ranked games.

I think perhaps if we want to try this, it might be acceptable to the majority if we made it a game that counts to the site rating but not the Standard one. After all, the site rating includes such things as Gunboat, FOW, AoE, different maps, Stuff, all sorts of things which all could be argued to be changing the game play. Even if this option does offer some rating advantage/disadvantage, it is probably no more extreme than some of those other game options we have. So we would have an option to use this type of scoring and allocate it to the site rating but not the 'real' one (the standard rating).

How do people feel about that?

EDIT: BTW, I did not mean we just go straight ahead. I am saying perhaps this is a route to try it when we have thrashed it out. For me, I keep thinking of things we would need to sort out first. The latest is I don't know what happens if there is a tie (ie two or more countries on the same SC count). This is pretty important ;)

Re: Ranked Draws & Rank-Based Scoring: Fibonacci-Diplo

PostPosted: 09 Jun 2018, 14:04
by NoPunIn10Did
super_dipsy wrote:I think what V was pointing out was that a lot of the background justification and assumptions around this idea are from the Tournament / F2F world. It absolutely makes enormous sense to use this in Dip tournaments, because by definition most real life tourneys have a fixed (and small) number of rounds and limited time, so you have to have a way of scoring that does not require the game to actually end. Using drawbased scoring does not have enough definition in a short length time. But the reason we have resisted it in the online form of the game is that we do not have the artificial stop point. We can allow a game to play out.

However to be honest our resistance has always been based on the simplistic point that if you start playing a game of Diplomacy with a background plan to try to last as long as possible (thereby increasing your return if you use a time served system) or to ensure that win or lose you get to a decent SC count, you are bound to play differently. Just a a silly example, but it seems to me that this system encourages Carebears; carebears typically make long game alliances but always risk their 'friend' stabbing them at the end, but in this sytem that is cool because you still end up highly placed (provide people stop the solo of course). The key here though is that although I believe it is true that changing the scoring system will affect play, what is not necessarily so clear is whether it will be worse or better from an enjoyment point of view. It is easy to see holes in a different system, just as it is to see holes in the existing one. I confess I am still nervous because I am not sure I believe the assertions of its key values; more solos, less draw whittling, more enjoyment. But that does not mean it is not worth trying.

This is the tricky bit though. I completely understand the issues of trying to get something new tried out when people have to read up and volunteer; we alaready know only a tiny % of our players use the forum anyway. But making it an option that is included in the current rating system may be a step too far too. If as some suspect it majorly changes play/outcomes, there is a danger it will distort the ratings. But making it available as an option only for unranked games again has some of the same drawbacks as making it a forum-based recruitment thing; it can be hard to get a critical mass because for some reason most players want to play ranked games.

I think perhaps if we want to try this, it might be acceptable to the majority if we made it a game that counts to the site rating but not the Standard one. After all, the site rating includes such things as Gunboat, FOW, AoE, different maps, Stuff, all sorts of things which all could be argued to be changing the game play. Even if this option does offer some rating advantage/disadvantage, it is probably no more extreme than some of those other game options we have. So we would have an option to use this type of scoring and allocate it to the site rating but not the 'real' one (the standard rating).

How do people feel about that?

EDIT: BTW, I did not mean we just go straight ahead. I am saying perhaps this is a route to try it when we have thrashed it out. For me, I keep thinking of things we would need to sort out first. The latest is I don't know what happens if there is a tie (ie two or more countries on the same SC count). This is pretty important ;)


I think this is a good route. Ties are explained in the rules, and they are included in the examples; they simply split the points from all applicable ranks. Two players tying at first place thus receive the average of the points awarded to first and second place; the next lower player gets third. Three players tied for second split the points for 2nd, 3rd, and 4th; the next lower player gets fifth place.

Several of the graphs in the original post show the points for these and similar scenarios.

Re: Ranked Draws & Rank-Based Scoring: Fibonacci-Diplo

PostPosted: 11 Jun 2018, 19:32
by DavidMaletsky
Been following this thread, finally decided to bite the bullet and join both Google and this website. A few notes:

1. For DSS / rulebook purists. Please note the rules have never in any incarnation said surviving players share equally in a WIN. It says they share equally in a draw. Having spoken with Calhamer on multiple occasions before he passed, if original intent matters, the draw language was put in the first edition by an editor, not by him. His vision was that players would gather, and play until someone won; of course, he didn’t know there were stalemate lines either. He eventually came around to DSS because it was the only model being used, online or in FTF play, but again, for those purporting to be originalists, not his OG intent.

2. Non-DSS models have taken over the FTF hobby simply because DSS creates consequences that are BOTH artificial AND unpleasant. To be clear, any scoring model that assigns a particular value to players surviving in a draw is automatically a variant; for purists, the rules say “share equally in a draw”, but they don’t say WHAT value to share equally. Zero is available, for example; turns out 0/3 = 0/5, and under one consequence that falls out of the rules, draw size is irrelevant for scoring. That said, as you enter the late mid game under DSS formulations, play will more often than not boil down to the needless and unpleasant task of killing off small powers or forcing them to vote themselves out of the result, for no rational choice reason other than this is what DSS is making us do, sorry buddy.

3. Carnage has gotten demonstrably positive feedback since it was introduced, largely because it solves the aforementioned issue. When everyone is content with their result under Carnage scoring, players can simply and amicably call the game. Does this mean Carnage is in principle the best or most correct system? Certainly not. As a hobby leader and tournament director, I observed there was a problem with DSS. Not an a priori problem, because among variants, whichever suits your interests is a normative matter. Rather, there was an a posteriori issue, mentioned above, of artificially motivated play with unpleasant consequences. When you run events, you want the setting and experience to be as enjoyable as possible for attendees. And again, since its inception, there has been much feedback indicative of positive impact on player experience from playing under Carnage rather than DSS, which is why the North American hobby that was once dominated by DSS now only has one FTF holdout that uses any draw-size based element, DixieCon. Which was funnily the most liberal iteration when it first came out 30+ years ago.

4. Alex’ Fibonacci method really makes Carnage accessible to variant play, which it was certainly not before. Having tracked this conversation, I would add as a final point that in FTF, there is a saying that if you want to do or try something in the hobby, put in the work. Scoring system debates were going on decades before we were in the hobby, and will likely go on decades after. Ultimately, those who put in the work to create and test models are the ones with the dessert to have them actualized; it’s a simple thing to be a devil’s advocate, or a fist pounding traditionalist, requires little effort or ingenuity. Which party should be rewarded?

Re: Ranked Draws & Rank-Based Scoring: Fibonacci-Diplo

PostPosted: 11 Jun 2018, 19:54
by boldblade
DavidMaletsky wrote:Been following this thread, finally decided to bite the bullet and join both Google and this website. A few notes:

1. For DSS / rulebook purists. Please note the rules have never in any incarnation said surviving players share equally in a WIN. It says they share equally in a draw. Having spoken with Calhamer on multiple occasions before he passed, if original intent matters, the draw language was put in the first edition by an editor, not by him. His vision was that players would gather, and play until someone won; of course, he didn’t know there were stalemate lines either. He eventually came around to DSS because it was the only model being used, online or in FTF play, but again, for those purporting to be originalists, not his OG intent.

2. Non-DSS models have taken over the FTF hobby simply because DSS creates consequences that are BOTH artificial AND unpleasant. To be clear, any scoring model that assigns a particular value to players surviving in a draw is automatically a variant; for purists, the rules say “share equally in a draw”, but they don’t say WHAT value to share equally. Zero is available, for example; turns out 0/3 = 0/5, and under one consequence that falls out of the rules, draw size is irrelevant for scoring. That said, as you enter the late mid game under DSS formulations, play will more often than not boil down to the needless and unpleasant task of killing off small powers or forcing them to vote themselves out of the result, for no rational choice reason other than this is what DSS is making us do, sorry buddy.

3. Carnage has gotten demonstrably positive feedback since it was introduced, largely because it solves the aforementioned issue. When everyone is content with their result under Carnage scoring, players can simply and amicably call the game. Does this mean Carnage is in principle the best or most correct system? Certainly not. As a hobby leader and tournament director, I observed there was a problem with DSS. Not an a priori problem, because among variants, whichever suits your interests is a normative matter. Rather, there was an a posteriori issue, mentioned above, of artificially motivated play with unpleasant consequences. When you run events, you want the setting and experience to be as enjoyable as possible for attendees. And again, since its inception, there has been much feedback indicative of positive impact on player experience from playing under Carnage rather than DSS, which is why the North American hobby that was once dominated by DSS now only has one FTF holdout that uses any draw-size based element, DixieCon. Which was funnily the most liberal iteration when it first came out 30+ years ago.

4. Alex’ Fibonacci method really makes Carnage accessible to variant play, which it was certainly not before. Having tracked this conversation, I would add as a final point that in FTF, there is a saying that if you want to do or try something in the hobby, put in the work. Scoring system debates were going on decades before we were in the hobby, and will likely go on decades after. Ultimately, those who put in the work to create and test models are the ones with the dessert to have them actualized; it’s a simple thing to be a devil’s advocate, or a fist pounding traditionalist, requires little effort or ingenuity. Which party should be rewarded?


Thank you.

Re: Ranked Draws & Rank-Based Scoring: Fibonacci-Diplo

PostPosted: 12 Jun 2018, 04:11
by V
NoPunIn10Did wrote:
super_dipsy wrote:I think what V was pointing out was that a lot of the background justification and assumptions around this idea are from the Tournament / F2F world. It absolutely makes enormous sense to use this in Dip tournaments, because by definition most real life tourneys have a fixed (and small) number of rounds and limited time, so you have to have a way of scoring that does not require the game to actually end. Using drawbased scoring does not have enough definition in a short length time. But the reason we have resisted it in the online form of the game is that we do not have the artificial stop point. We can allow a game to play out.

However to be honest our resistance has always been based on the simplistic point that if you start playing a game of Diplomacy with a background plan to try to last as long as possible (thereby increasing your return if you use a time served system) or to ensure that win or lose you get to a decent SC count, you are bound to play differently. Just a a silly example, but it seems to me that this system encourages Carebears; carebears typically make long game alliances but always risk their 'friend' stabbing them at the end, but in this sytem that is cool because you still end up highly placed (provide people stop the solo of course). The key here though is that although I believe it is true that changing the scoring system will affect play, what is not necessarily so clear is whether it will be worse or better from an enjoyment point of view. It is easy to see holes in a different system, just as it is to see holes in the existing one. I confess I am still nervous because I am not sure I believe the assertions of its key values; more solos, less draw whittling, more enjoyment. But that does not mean it is not worth trying.

This is the tricky bit though. I completely understand the issues of trying to get something new tried out when people have to read up and volunteer; we alaready know only a tiny % of our players use the forum anyway. But making it an option that is included in the current rating system may be a step too far too. If as some suspect it majorly changes play/outcomes, there is a danger it will distort the ratings. But making it available as an option only for unranked games again has some of the same drawbacks as making it a forum-based recruitment thing; it can be hard to get a critical mass because for some reason most players want to play ranked games.

I think perhaps if we want to try this, it might be acceptable to the majority if we made it a game that counts to the site rating but not the Standard one. After all, the site rating includes such things as Gunboat, FOW, AoE, different maps, Stuff, all sorts of things which all could be argued to be changing the game play. Even if this option does offer some rating advantage/disadvantage, it is probably no more extreme than some of those other game options we have. So we would have an option to use this type of scoring and allocate it to the site rating but not the 'real' one (the standard rating).

How do people feel about that?

EDIT: BTW, I did not mean we just go straight ahead. I am saying perhaps this is a route to try it when we have thrashed it out. For me, I keep thinking of things we would need to sort out first. The latest is I don't know what happens if there is a tie (ie two or more countries on the same SC count). This is pretty important ;)


I think this is a good route. Ties are explained in the rules, and they are included in the examples; they simply split the points from all applicable ranks. Two players tying at first place thus receive the average of the points awarded to first and second place; the next lower player gets third. Three players tied for second split the points for 2nd, 3rd, and 4th; the next lower player gets fifth place.

Several of the graphs in the original post show the points for these and similar scenarios.


Although I “signed off” this thread, I see Super contributed with his usual wisdom & NoPun agreed, which looks like excellent consensus. Despite my apparent “devil’s advocacy” I’d definitely support this route forward as well.

Re: Ranked Draws & Rank-Based Scoring: Fibonacci-Diplo

PostPosted: 12 Jun 2018, 16:22
by DQ
Oh, Sure, David Maletsky shows up and everyone suddenly decides to be reasonable and amicable.

Swear to god he's channeling Jimmy Buffet. :lol:

Re: Ranked Draws & Rank-Based Scoring: Fibonacci-Diplo

PostPosted: 12 Jun 2018, 16:24
by Nanook
DQ wrote:Oh, Sure, David Maletsky shows up and everyone suddenly decides to be reasonable and amicable.

Swear to god he's channeling Jimmy Buffet. :lol:

He has gravitas.

Re: Ranked Draws & Rank-Based Scoring: Fibonacci-Diplo

PostPosted: 12 Jun 2018, 23:45
by asudevil
DQ wrote:Oh, Sure, David Maletsky shows up and everyone suddenly decides to be reasonable and amicable.

Swear to god he's channeling Jimmy Buffet. :lol:


He's just prettier than you