7WD in Numbers

Forum rules
Please DO NOT POST BUGS on this forum. Please report (and vote) bugs on : https://boardgamearena.com/bugs
User avatar
Le007n
Posts: 64
Joined: 30 March 2020, 22:47

Re: 7WD in Numbers

Post by Le007n »

Hi All,

To respond to a few points:

@Pistol Star : You are right (as always :)) about the 2% each point deviation since its essential zero sum, i didn't really think about it like this!
I think we all know first player has an advantage so based on this smallish data set its interesting to put a number to this

@Exoone : I have more data around the affect of replay wonders on win % which I may write up when time allows based on this current data set among other things

I only recorded selection of Ra when looking at divinities as this was the one in most question for me
Last edited by Le007n on 27 April 2022, 16:11, edited 1 time in total.
User avatar
Silene
Posts: 778
Joined: 23 October 2013, 17:50

Re: 7WD in Numbers

Post by Silene »

Be careful on how you collect your data. This thread might be worth a read for you: https://boardgamearena.com/forum/viewto ... 90&t=22001
Hosting Allround-League: https://boardgamearena.com/group?id=7870115 --> a league where you have matches of random games vs. other players in your group - season 6 started in Nov. '23 with 128 participants.
User avatar
frankenStain
Posts: 79
Joined: 02 December 2020, 01:18

Re: 7WD in Numbers

Post by frankenStain »

Awesome stuff!! Thanks for your work! :)

are there any other ways to get usable data besides scraping btw? (since the statistics and the whole replay is pretty bad; but better than nothing I guess..)

...aaand, would you be willing to share your data set, or analyze some specific things (would have some ideas, just not too much time to make the short list of the most interesting stuff..) ?

Edit:
personally most interested in games between experts (mostly with both expansions, but vanilla would be fascinating too) - in practice this would mean (for both expansions): a max elo difference of 100, with both players having at least 450 elo (100 elo beacuse most top players have a max swing of +-50elo, exluding ones who seek to make some records [would just nominally assign for every player over 700, an elo worth of 700].
with this filters it would be interesting to know:
- how much affects Ra the chance to win (also interesting would be neptune!)
- how much affect specific wonders (like knossos, theatre, julia, artemis and marsoleum) the chacne to win
- (less important: which wonders choose players here (and it would be even more fun to know, which are chosen over which))
- which is the winning category (percentage distribution)
- how much being first in Age 1 affects the winning chances here (and being first in age 2) [this would also be interesting for vanilla; but I have no idea how to filter for top players with elo... maybe based on arena ranks/scores??]

final thought: maybe an max elo difference of 70 elo would be even better; 70 elo would mean, if this would function perfectly, that one player is x1.5 more likely to win (60% vs 40% win chance)
User avatar
Patrick of the Isles
Posts: 107
Joined: 30 August 2020, 13:20

Re: 7WD in Numbers

Post by Patrick of the Isles »

Thanks for taking the time to share these details. I only play the base game, so I'm of course most interested in that analysis.

Your wonder win rate is only provided in the aggregate (base + expansion), but assuming the win rates for the base wonders is similarly-ordered when taking out the expansion games, The Sphinx being ahead of the Temple of Artemis is quite interesting. I almost always take the Temple over the Sphinx, as the coins seem so much more flexible. So that was quite interesting to see!
silentProtest
Posts: 170
Joined: 26 March 2022, 20:12

Re: 7WD in Numbers

Post by silentProtest »

Edit: not relevant anymore - and cannot delete this post

original:
[Is my browser not loading some part or is starting player in age1/2/3 to win correlation gone ???]
Last edited by silentProtest on 23 August 2022, 23:00, edited 1 time in total.
User avatar
Lofski
Posts: 22
Joined: 15 March 2020, 12:02

Re: 7WD in Numbers

Post by Lofski »

Very interesting analysis!

The 1st player's advantage can also be seen through logic: Going first has to be worth at least 1 coin, because if starting 2nd with +1 coin would be advantageous, the 1st player could always just trash something for +2, effectively becoming the 2nd player with +1 coin. Obviously there are usually better options, so in most situations starting is worth more than 1 coin.

This and also rebalancing Ra would call for an unofficial variant, if the designer & publisher are not interested. I think it would be best to give the 2nd player just +1, because that would be a certain improvement over the original rules. +2 has a small chance of making the balance actually worse.
silentProtest
Posts: 170
Joined: 26 March 2022, 20:12

Re: 7WD in Numbers

Post by silentProtest »

Just thought about upping this topic, and sharing some information I noticed:
If playing with both expansions the "first player advantage" is equivalent to 22-23 elo. =aprox. 1.137 more likely to win
If playing without expansions the "first player advantage" is equivalent to 43-44 elo. =aprox. 1.257 more likely to win


Notes: It is based probably on a very brought elo destrebution and elo-difference destrebution, and is only based on "won if started in age 1".

PS: little table
70 elo = 1.5 more likely
120 elo = 2 times more likely
190 elo = 3 time more likely
240 elo = 4 time more likely
User avatar
Pistol Star
Posts: 99
Joined: 11 October 2016, 02:41

Re: 7WD in Numbers

Post by Pistol Star »

silentProtest wrote: 22 November 2022, 02:30 Just thought about upping this topic, and sharing some information I noticed:
If playing with both expansions the "first player advantage" is equivalent to 22-23 elo. =aprox. 1.137 more likely to win
If playing without expansions the "first player advantage" is equivalent to 43-44 elo. =aprox. 1.257 more likely to win


Notes: It is based probably on a very brought elo destrebution and elo-difference destrebution, and is only based on "won if started in age 1".

PS: little table
70 elo = 1.5 more likely
120 elo = 2 times more likely
190 elo = 3 time more likely
240 elo = 4 time more likely
Great work and very interesting statistics!
silentProtest
Posts: 170
Joined: 26 March 2022, 20:12

Re: 7WD in Numbers

Post by silentProtest »

there is no pin function (that functions automaticaly or can be used by majority) - So I will be bumping the most useful posts/discussion - hope that is fine.
User avatar
Jellby
Posts: 1343
Joined: 31 December 2013, 12:22

Re: 7WD in Numbers

Post by Jellby »

silentProtest wrote: 21 March 2023, 17:12 there is no pin function (that functions automaticaly or can be used by majority) - So I will be bumping the most useful posts/discussion - hope that is fine.
No, it's not.
Post Reply

Return to “7 Wonders Duel”