Has there ever been any analysis about how strong first player advantage is?

Forum rules
Please DO NOT POST BUGS on this forum. Please report (and vote) bugs on : https://boardgamearena.com/bugs
User avatar
MajorGamble
Posts: 3
Joined: 20 April 2011, 22:35

Re: Has there ever been any analysis about how strong first player advantage is?

Post by MajorGamble »

LargoLaGrande wrote: 22 December 2023, 16:15 I've checked for curiosity the last 5 losses for the first 10 players ranked in Arena, regardless of the strength of the opponent.

Code: Select all

1.  CuSO44ek       - 5/5 starting 2° - 100%
2.  Victhan        - 4/5 starting 2° -  80%
3.  predatorc      - 3/5 starting 2° -  60%
4.  slywww2020     - 4/5 starting 2° -  80%
5.  Wolog          - 2/5 starting 2° -  40%
6.  kapout         - 5/5 starting 2° - 100%
7.  viermaaldomi   - 5/5 starting 2° - 100%
8.  Almatziol      - 4/5 starting 2° -  80%
9.  ViciousCyclist - 3/5 starting 2° -  60%
10. SrkiBG         - 5/5 starting 2° - 100%
Well done @LargoLaGrande! This is EXCELLENT data. Very easy to confirm by any of us with most up-to-date stats at any time and quite compelling, considering it uses only the top 10 players.

I updated it with current data now that it's been a couple days since the original post (and i would encourage others to do the same in the future).

FWIW, I'm not a top player, but my own anecdotal experience completely follows this as well—all of my 5 most recent losses have been 2nd player. I have yet to see a single configuration in which i would wish to go 2nd. I will be quite surprised if we see any significant exceptions following this very strict metric: 5 most recent losses by a top-10-ranked player.
LargoLaGrande
Posts: 7
Joined: 27 April 2020, 01:48

Re: Has there ever been any analysis about how strong first player advantage is?

Post by LargoLaGrande »

The only big tournament played in december with a lot of the best players in the charts (https://boardgamearena.com/tournament?id=247728), analyzing the last matches with (in theory) the strongest players of the tournament:

Code: Select all

Quarter-finals: 4/4 wins of the player starting first. 100%
   Semi-finals: 2/2 wins of the player starting first. 100%
         Final: 1/1 wins of the player starting first. 100%
Probability of this in a (in theory) game with 50-50 chance of victory: 0,78 %
User avatar
tchobello
Posts: 525
Joined: 18 March 2012, 13:19

Re: Has there ever been any analysis about how strong first player advantage is?

Post by tchobello »

thanks for all your stats. It has convinced me to rewrite the whole random algorithm.

I hope this one suits you better...
LargoLaGrande
Posts: 7
Joined: 27 April 2020, 01:48

Re: Has there ever been any analysis about how strong first player advantage is?

Post by LargoLaGrande »

I wrote before in a private feedback, I'll rewrite here in public because I saw no change in your or the creators behaviour: people are pointing out a "problem" (real or not, is not the point) that they "feel" playing the game. A feel that can ruin the experience and I found your answers really arrogant.

I don't want to convince nobody.
You provided stats and motivations not really convincing me, and I'm doing my very little and incomplete research to convince myself, failing.
I don't want to convince you, I want to be convinced that the game is fair.

I really like the game, it's great, but sometimes it's starting to becaming frustrating.
When I start second I only play in the fear of "not to loose", constantly following and trying to block my opponent. An this is not fun.
On the opposite side, normally when I start first, I quickly get the control of the board even with strong players. And this is also not fun.
Maybe I'm doing something wrong, maybe I have to study more the game, but I'm not having fun.

I still have to find a single random game where the first player played well, and lost.

You don't need to change the game. On BGA I play for example "Connect four". I'm experienced, I managed to win the last Arena. In that game (expecially in the standard 6x7) the first player wins 100%, it's math. They resolved with a simple, not-changing-nothing rule: the pie rule.
It seems that you are taking this personally, that's not a fault of you or the game. Even really complex games like chess, with experienced players, have a strong first player advantage.

After 2 or more month on BGA there is no player with 500+ ELO, for an abstract fair 1vs1 game this is strange. You will expect to have at least a little bit of strong players that wins almost all the games that they play.

I hope that you will listen with a more "open" attitude.
User avatar
MajorGamble
Posts: 3
Joined: 20 April 2011, 22:35

Re: Has there ever been any analysis about how strong first player advantage is?

Post by MajorGamble »

You are handling this in a respectful, thoughtful manner, @LargoLaGrande. The problem in this case is that you're obviously correct about something that the creators and developer emphatically deny: that the game has an obvious advantage for the starting player. Until the creators acknowledge this, little is likely to be done to correct it.

It's a shame it can't be more competitive considering the game really is a lot of fun! I hope an expansion addresses some of these concerns in a meaningful way.
andreathejedy
Posts: 3
Joined: 26 September 2023, 23:32

Re: Has there ever been any analysis about how strong first player advantage is?

Post by andreathejedy »

Hi all again, I’m Andrea Mainini, one of the authors of the game.
Here's a concise update on the RANDOM SETUP, addressing concerns raised by players over time:
- in the "physical game", the issue of any advantage/disadvantage for the first player does not arise for many reasons already explained at length, and also because, during the setup phase, it is the two opponents who consciously place the galaxies and black holes so that their distribution makes "sense"; initial implementation of the RANDOM SETUP on BGA faced challenges in simulating human-like randomness, leading to occasional “imbalance” in galaxy and black hole distribution;
- so we continued working to find an algorithm capable of consistently providing a satisfying RANDOM SETUP, and at the end of January 2024 we introduced it on BGA; after weeks of observation, we're confident in the positive outcome and hope you'll continue enjoying ORION Duel!
Post Reply

Return to “Orion Duel”