votes

The current system of NemePoints has the problem that players that play a lot will gain many more points than players less assiduous. In my gaming group we have many different players: there are three of us that play many games and are almost always present; and then there are some other players that do not play so much. I think that it is important however to have a system that allows to compare the ability of both frequent players and the least assiduous.

I think that this could be an issue for many other gaming group, but I'm not sure how to solve this. Perhaps using a system like the ELO system of Board Game Arena? I don't know... I trust in your competence and thank you so much for your wonderful work! =)

Author: Lorenzo, 26.12.2016, 05:40

nemestats, 01.01.2017, 21:29

I think one metric we are going to add is the "average percentile" which will basically show how you rank across many games. So if you were 2nd place in a 10-player game that would count as the 90th percentile. if you were 3rd in a 3-player game then you'd be 33 percentile.

Would this address the concern?

Also note there is an "Average Points Per Game" and "Average Points Per Player" which may also help. These are "NemePoints" BTW. Reply

Would this address the concern?

Also note there is an "Average Points Per Game" and "Average Points Per Player" which may also help. These are "NemePoints" BTW. Reply

I think this could be a nice addition for statistics, but it does not solve the problem. While I like to have a lot of statistic, I think that there should be a "privileged" one for calculating a player skill in a league, so that I can have a classification of the players of the league based just on one statistic. A unique score that sums up all the player skill at playing board games.

I think that this score should be NemePoints, that work really well in my opinion, but lack this fact of the more assiduous players and the less assiduous one. A score like the one I'm suggesting should have the following features:

- should be based on the "weight" of the game

- should be based on the number of players playing

- should take count of the players ranking, in a way that beating skilled players would award more points than beating bad players

- should not penalize players that miss some meetings

I guess that's all a scoring system for board games should do. And I think that neme points already satisfy the first two of these requirements. We should just implement them to satisfy the other two. I don't think this should be much difficult. I suggest to add to the current system of neme points the following ideas:

- use negative values for neme points. A player that loses in a game should lose neme points, not gain them. For example in a 4 player game the first and the second player would gain neme points, the third and the fourth would lose neme points. This way a player that plays a lot of games but loses them all would not be on the top of the league ranking just because it played a lot of games, and a player less assiduous but with a better percentile of winning would be over him in the ranking.

- calculate the amount of neme points available in a game relying on the current ranking in the league of every participant . This way a game against skilled players will award more points. Reply

I think that this score should be NemePoints, that work really well in my opinion, but lack this fact of the more assiduous players and the less assiduous one. A score like the one I'm suggesting should have the following features:

- should be based on the "weight" of the game

- should be based on the number of players playing

- should take count of the players ranking, in a way that beating skilled players would award more points than beating bad players

- should not penalize players that miss some meetings

I guess that's all a scoring system for board games should do. And I think that neme points already satisfy the first two of these requirements. We should just implement them to satisfy the other two. I don't think this should be much difficult. I suggest to add to the current system of neme points the following ideas:

- use negative values for neme points. A player that loses in a game should lose neme points, not gain them. For example in a 4 player game the first and the second player would gain neme points, the third and the fourth would lose neme points. This way a player that plays a lot of games but loses them all would not be on the top of the league ranking just because it played a lot of games, and a player less assiduous but with a better percentile of winning would be over him in the ranking.

- calculate the amount of neme points available in a game relying on the current ranking in the league of every participant . This way a game against skilled players will award more points. Reply

Any news on this matter?
Reply

nemestats, 06.02.2017, 21:38

Still digesting this one. Here are some thoughts:

* A player's skill/rank will change over time. If you go back and edit a game in the past the NemePoints calculation could change entirely. This could be weird as some users would expect/want this to happen and others would not

* What would the formula be for calculating points? It's already pretty complex: https://github.com/NemeStats/NemeStats/blob/ccf124af80c682dc84c222ac325ab415b115e1e5/Source/BusinessLogic/Logic/Points/PointsCalculator.cs

* I think the intention of NemePoints is not necessarily to measure the skill of one player vs. another. I.e. someone who just plays a crap-ton of games should still have a lot of NemePoints. That being said, I think there is absolutely a place for what you are describing, and maybe there is some overlap in how these points are calculated. I like the idea of having positive and negative points. Over time your distance from 0 would be a strong indicator of "skill"

I'm looking forward to other members of the community weighing on this one. It would be an interesting undertaking! Reply

* A player's skill/rank will change over time. If you go back and edit a game in the past the NemePoints calculation could change entirely. This could be weird as some users would expect/want this to happen and others would not

* What would the formula be for calculating points? It's already pretty complex: https://github.com/NemeStats/NemeStats/blob/ccf124af80c682dc84c222ac325ab415b115e1e5/Source/BusinessLogic/Logic/Points/PointsCalculator.cs

* I think the intention of NemePoints is not necessarily to measure the skill of one player vs. another. I.e. someone who just plays a crap-ton of games should still have a lot of NemePoints. That being said, I think there is absolutely a place for what you are describing, and maybe there is some overlap in how these points are calculated. I like the idea of having positive and negative points. Over time your distance from 0 would be a strong indicator of "skill"

I'm looking forward to other members of the community weighing on this one. It would be an interesting undertaking! Reply

Bahhh I just spent hours trying to figure out a Handicap system like they have in Golf but I got fed up with going nowhere and then realised it wouldn't really solve the issue anyways :(
Reply

nemestats, 07.02.2017, 07:49

@ Deadshot Omega Still, did you learn anything worth sharing so we don't re-invent the same square wheel? :)
Reply

Glad you like the idea and there is some brainstorming between the development crew.:) I'm going to try to help as I can:

- I think the skil/rank changing over time is not a problem. Of course, you will not gain more points for the game that you previously played just because right now one of your opponents is going uo in the rank. The points should always be static. One month ago you beated 2 skilled players? Good, you gained X points based on their skill in that precise moment. Doesn't matter if one month later they have become stronger or weaker. That particular game one month ago is always worth the same amount of points, because it is based on the skill of the players playing in that particular moment. However, I would like to concentrate more on the topic of the negative Neme points, which is more important.

- I think that improving NemePoints as the most important statistic to measue skill would be awesome first of all for the site, because it would become a strong instrument to really measure the skill in a playgroup. That's something that a player really searches in a site like NemeStats, IMHO.

- the formula. That's the hard part. I don't think I can understand all the things you wrote in the link. However, I would suggest to use the current system with some minor changes: you calculate basic nemepoints as usal (10*number of players) then you apply the Fibonacci series based on the number of players, giving to the last positions the same amount of nemepoints of the first ones, just with the negative symbol. For example, in a 5 player game you would have 50 Nemepoints to divide. The players would gain the following points:

1° ---> +A

2° ---> +B

3° ---> 0

4° ---> -B

5° ---> -A

The value of the fractions should be similar to the ones you currently use in the Fibonacci series.

Secondly, you apply all the bonus modificators as usual: lenght, weight, opponents skill perhaps ;) . Of course, these modificators are bonuses so they will all have the positive symbol. Reply

- I think the skil/rank changing over time is not a problem. Of course, you will not gain more points for the game that you previously played just because right now one of your opponents is going uo in the rank. The points should always be static. One month ago you beated 2 skilled players? Good, you gained X points based on their skill in that precise moment. Doesn't matter if one month later they have become stronger or weaker. That particular game one month ago is always worth the same amount of points, because it is based on the skill of the players playing in that particular moment. However, I would like to concentrate more on the topic of the negative Neme points, which is more important.

- I think that improving NemePoints as the most important statistic to measue skill would be awesome first of all for the site, because it would become a strong instrument to really measure the skill in a playgroup. That's something that a player really searches in a site like NemeStats, IMHO.

- the formula. That's the hard part. I don't think I can understand all the things you wrote in the link. However, I would suggest to use the current system with some minor changes: you calculate basic nemepoints as usal (10*number of players) then you apply the Fibonacci series based on the number of players, giving to the last positions the same amount of nemepoints of the first ones, just with the negative symbol. For example, in a 5 player game you would have 50 Nemepoints to divide. The players would gain the following points:

1° ---> +A

2° ---> +B

3° ---> 0

4° ---> -B

5° ---> -A

The value of the fractions should be similar to the ones you currently use in the Fibonacci series.

Secondly, you apply all the bonus modificators as usual: lenght, weight, opponents skill perhaps ;) . Of course, these modificators are bonuses so they will all have the positive symbol. Reply

Probably the best way to deal with this matter is to use the ELO system of chess. The site "boardgamearena", which is a site to play boardgames online, uses it with some minor changes to adapt it to boardgames. That's their formula: http://it.doc.boardgamearena.com/Rating

I studied it a little bit tonight and perhaps I've come to a good understanding of it and I think it could be useful to us, too. I just have to understand if it could still work with the weight and lenght bonuses, which is something I would not like to renounce. I will keep you updated. In the meantime, if you want to check the link I pasted, it is a pretty simple system and it's explained pretty clearly. Reply

I studied it a little bit tonight and perhaps I've come to a good understanding of it and I think it could be useful to us, too. I just have to understand if it could still work with the weight and lenght bonuses, which is something I would not like to renounce. I will keep you updated. In the meantime, if you want to check the link I pasted, it is a pretty simple system and it's explained pretty clearly. Reply

Alright, I asked to someone in the boardgamearena forum and he said me that the formula I found in their Wiki is wrong. So i studied a bit the ELO rating system on the right page, the one of Wikipedia: https://en.wikipedia.org/wiki/Elo_rating_system . I think that this system could work for our purpose, with some adjustments and additions. I suggest you to look at the page to understand how the system works. It's not so complicated and finally it is just two simple formulas to apply.

The Elo system is used in chess to calculate the skill of players. It takes count of the actual skill of the two opponents, so a weaker player that beats a stronger one will gain more points than if it would have happen the opposite. The fact is that ELO system works just for 1vs1 games, and not for multiplayers. To get rid of this problem, my suggestion is to "transform" every game of every boardgame in a 1vs1 game, where players fight against the table level, which is the simple average of the rating of every player at the table. This way we can simply apply the ELO rating system to every boardgame we could play. The table level could be even modified by the weight of the game and his duration, in my opinion, and the system should still work perfectly. It would be like if you would be playing against a particularly strong opponent. But this should be a little bit tested.

That's my basic idea and I could explain some details with an example, too, fi you say me that you are interested. What do you think? Could this work? Reply

The Elo system is used in chess to calculate the skill of players. It takes count of the actual skill of the two opponents, so a weaker player that beats a stronger one will gain more points than if it would have happen the opposite. The fact is that ELO system works just for 1vs1 games, and not for multiplayers. To get rid of this problem, my suggestion is to "transform" every game of every boardgame in a 1vs1 game, where players fight against the table level, which is the simple average of the rating of every player at the table. This way we can simply apply the ELO rating system to every boardgame we could play. The table level could be even modified by the weight of the game and his duration, in my opinion, and the system should still work perfectly. It would be like if you would be playing against a particularly strong opponent. But this should be a little bit tested.

That's my basic idea and I could explain some details with an example, too, fi you say me that you are interested. What do you think? Could this work? Reply

nemestats, 25.07.2017, 08:01

@Lorenzo -- we are ready to take on another big ticket enhancement or issue -- so we are revisiting this one again. I think it would be possible to implement something basic to give a minor modifier based on this -- but I'm worried that if we try to get too fancy it will be too complex and people will poke holes in the system.
Reply

I just would like to emphasis the following statement of Lorenzo:

"- I think that improving NemePoints as the most important statistic to measue skill would be awesome first of all for the site, because it would become a strong instrument to really measure the skill in a playgroup. That's something that a player really searches in a site like NemeStats, IMHO."

I switched

I don't know anything (yet) about the ELO system but negative nemepoints and a tabel level sounds really, really good (at first glance) and not too complex. Please, please revise this issue furthermore :-) Reply

"- I think that improving NemePoints as the most important statistic to measue skill would be awesome first of all for the site, because it would become a strong instrument to really measure the skill in a playgroup. That's something that a player really searches in a site like NemeStats, IMHO."

I switched

I don't know anything (yet) about the ELO system but negative nemepoints and a tabel level sounds really, really good (at first glance) and not too complex. Please, please revise this issue furthermore :-) Reply

I think the percentile option is a good one. I don't think we need to change the neme points. Maybe add another. At our group we use nemesis to rank people and also winning percentage for people playing more than 10 games in a month. We may change this to 15 not sure yet. We then publish each month's rankings.
Reply

Copyright - 2019 Informer Technologies, Inc. All Rights Reserved. Feedback system is used
Idea.Informer.com

Using an overall average for a group could still "disadvantage" rare players, but if each player individually received an average, only for games they actually played in, then that number may be a good way to represent and compare ability. Reply