Board Game Arena

When I look at the first 2 players of 7 Wonders for this season, I realize they have stopped playing for more than a week just after they reached the first and second place.
This post is absolutely not about criticizing the choice of these players, it's about how is it still possible it happens? The goal of the Arena was to adress competition, and we see again that top players stopped playing. It becomes thus impossible to be matched with them and to measure our skills with them..

Stopping playing for a week to maintain standings is preferable to stopping playing altogether. Arena will reset after a certain time, but that time is measured in months, not days.

Gunslinger99 wrote: ↑15 February 2020, 15:30 This post is absolutely not about criticizing the choice of these players, it's about how is it still possible it happens? The goal of the Arena was to adress competition, and we see again that top players stopped playing. It becomes thus impossible to be matched with them and to measure our skills with them..

The goal of arena is to determine the best player over a 3 month period...and is already very heavily weighted to those who play more. This maybe isn't so bad for a quick game as 50-100 plays are achievable, but to play that number of games and take a break is perfectly okay. In fact it seems pretty ridiculous to complain about 1 week of someone not playing. If they got to that rating it is because they played well enough to achieve it. Complaining about them not playing is implying that they are only keeping their ranking by doing so, which is just nonsense. They played 50+ games to get there, they should feel free to take a break.

The number of plays required for longer games to be competitive is already obscene to the point that competing without giving a significant portion of your life to a game (75-150 hours to be competitive over 3 months for longer games...and many more hours to actually reach your true EAS for the best players...absurd to contend that it takes that long to see who is strongest). Arena measures what strong players have the most free time.

sprockitz wrote: ↑18 February 2020, 01:17

Gunslinger99 wrote: ↑15 February 2020, 15:30 This post is absolutely not about criticizing the choice of these players, it's about how is it still possible it happens? The goal of the Arena was to adress competition, and we see again that top players stopped playing. It becomes thus impossible to be matched with them and to measure our skills with them..

The number of plays required for longer games to be competitive is already obscene to the point that competing without giving a significant portion of your life to a game (75-150 hours to be competitive over 3 months for longer games...and many more hours to actually reach your true EAS for the best players...absurd to contend that it takes that long to see who is strongest). Arena measures what strong players have the most free time.

Totally agree, especially for long games like Terra Mystica. Would there be a viable way for top players to advance more quickly for long games, to reduce the time commitment required ?

firexed wrote: ↑18 February 2020, 01:56 Would there be a viable way for top players to advance more quickly for long games, to reduce the time commitment required ?

Actually, yes, there is: https://boardgamearena.com/bug?id=14455&vote

But I think the main thing would be to improve the matchmaking so that strong players can play against each other. First, you would be more motivated to actually invest many hours and second, the EAS would not be cluttered with random noise against low-ranked players.

firexed wrote: ↑18 February 2020, 01:56
Totally agree, especially for long games like Terra Mystica. Would there be a viable way for top players to advance more quickly for long games, to reduce the time commitment required ?

Isn't that what swiss rounds and/or ELO do ?

But seriously now, I think the arena system today is a transformation of ELO calculation, with K = 6 and the division of the players acting as a constant ELO for 1 division. (@admin, am I not right ? ). Maybe they could follow the principles of ELO calculation till the end and use several values for K according to the number of games played during the season ? For example, for the 5 first games K = 10, the next 5 = 8 and finally applying K = 6 for the games after the 10th... Just a thought.

dschingis27 wrote: ↑18 February 2020, 10:34

firexed wrote: ↑18 February 2020, 01:56 Would there be a viable way for top players to advance more quickly for long games, to reduce the time commitment required ?

Actually, yes, there is: https://boardgamearena.com/bug?id=14455&vote

But I think the main thing would be to improve the matchmaking so that strong players can play against each other. First, you would be more motivated to actually invest many hours and second, the EAS would not be cluttered with random noise against low-ranked players.

That suggestion you linked has very good points. The current system in 3+ player games favors those who play more with higher ranked players. In 4-player games you can for example win 65% of games with weaker players and 38% with more higher ranked players but your excepted points per game are actually negative with lower ranked players as you get only 7 points for the win and -7 for the second place and -30+ for the lost.

So to gain very high points in current elo system you have to either hope that there are not many lesser ranked players playing with you or just grind the game and hope you get lucky streak going. At least the scheduled games gave top players an equal opportunity to get higher ranked players to play with you.

As a quick example if i play and win 10 straight 4p TM games all against average (1500) opponents my EAS will move up to ~1650. 1650 should mean i have a ~75% chance of against each average player individually and ~40% of beating all 3 and winning the game. So what are the odds of a 1650 skill player winning 10 straight to get to 1650 if that is their true skill level...1 in 10,000. (If they are truly 1500 it is less tan 1 in a million)

So let’s not be unrealistic with how slowly we are moving the EAS score and make it about who is best not who has the most time. Up the K factor (straight divide by number players is silly...that same 1650 player who got there winning 10 2p games has a 1/20 chance of doing it if 1650 is their true rating...and would have spent way less time as 2p games are much faster. Weight K by player number like in Elo. And yeah implement those other linked suggestions