webdiplomacy.net

Durga wrote: ↑

Tue Jun 25, 2019 1:51 pm

SimonMH wrote: ↑

Thu Jun 20, 2019 10:33 am

Fascinating. Thank you to everyone for putting in the work to make this a success.

A mild caveat: the players in the last two games know exactly what they need to do to go through; a luxury that was denied the speedier finishers. In Round 2 is there a case for keeping results secret until all are finished?

Yes I'm happy to enforce that. There's a reason I never posted the score for this round. Perhaps I should have been stricter when it came to other people doing it

ChippeRock wrote: ↑

Thu Jul 04, 2019 4:27 pm

It only makes the information harder to attain - not impossible to attain. I can still go and compile all of the scores and simply keep it for myself and not post them.

There's unfortunately no real way to eliminate this advantage of dragging out games longer through pauses and the like. You can suppress the distribution of the information, but the information still exists. The only other real solution would be to cut down on the amount of pause time given to players.

SimonMH wrote: ↑

Fri Jul 05, 2019 2:32 pm

ChippeRock wrote: ↑

Thu Jul 04, 2019 4:27 pm

It only makes the information harder to attain - not impossible to attain. I can still go and compile all of the scores and simply keep it for myself and not post them.

There's unfortunately no real way to eliminate this advantage of dragging out games longer through pauses and the like. You can suppress the distribution of the information, but the information still exists. The only other real solution would be to cut down on the amount of pause time given to players.

Is there not a way to make the games private so that only those who are playing can see them?

SimonMH wrote: ↑

Fri Jul 05, 2019 2:32 pm

Is there not a way to make the games private so that only those who are playing can see them?

SimonMH wrote: ↑

Sun Jul 07, 2019 2:39 pm

What limited mathematical abilities I ever had are now pretty much restricted to backgammon equities and bridge percentages. I would be interested to know what people think is the odds-on score to reach the final table.

I may even offer a small prize in gratitude to the responder with the most accurate answer.

• The players are assigned randomly to games.
• Each game has a random result, drawn from the (finished) results of the first round, except that I reduced the probability of solos to 1 in 6, because my understanding is that this should go down as the field gets stronger.
• The results are randomly distributed among the players in the game.

• To qualify 95% of the time, 51.1 points should do it.
• To miss out 95% of the time, aim to get less than 39.2 points.
• This threshold actually barely varies with solo probability; there's only a very slightly higher bar in a zero-solo world, and very slightly lower in a half-games-solo world.
• I also tried out a model where the players have varied skill, so win probabilities weren't uniform. This also has remarkably little effect on the points needed to qualify.

Claesar wrote: ↑

Sun Jul 07, 2019 11:26 pm

SimonMH wrote: ↑

Sun Jul 07, 2019 2:39 pm

What limited mathematical abilities I ever had are now pretty much restricted to backgammon equities and bridge percentages. I would be interested to know what people think is the odds-on score to reach the final table.

I may even offer a small prize in gratitude to the responder with the most accurate answer.

I'll calculate something for you if you'll play a few hands of Bridge with me online.

teccles wrote: ↑

Mon Jul 08, 2019 8:43 pm

Here's my crack at this; it has simulations and stuff! TL;DR - 47.2

I assumed that:

• The players are assigned randomly to games.
• Each game has a random result, drawn from the (finished) results of the first round, except that I reduced the probability of solos to 1 in 6, because my understanding is that this should go down as the field gets stronger.
• The results are randomly distributed among the players in the game.

I then simulated 1000 semi-finals. For each, I took the qualification threshold as halfway between the 7th and 8th best player. The final threshold - 47.2 - is the median of those thresholds.

Some other things:

• To qualify 95% of the time, 51.1 points should do it.
• To miss out 95% of the time, aim to get less than 39.2 points.
• This threshold actually barely varies with solo probability; there's only a very slightly higher bar in a zero-solo world, and very slightly lower in a half-games-solo world.
• I also tried out a model where the players have varied skill, so win probabilities weren't uniform. This also has remarkably little effect on the points needed to qualify.