Help me understand please

[Cepsy]

Hi, I was playing this table
https://boardgamearena.com/15/luckynumb ... =645042984

Can you just explain me why I was eliminated instead of the 2nd even if I had more remaining time and we had the exactly same points even in the rounds?
Is it because i lost more ELO than him/her?

Thanks

[nik592]

This looks like a true tie (with no secondary tiebreakers to break it), and there was a recent discussion on this where apparently who goes through at that point is random. So, basically, you were just unlucky.

[yodamus]

From what has happened to me quite a few times. Is you lost the coin flip. It is random as I have won that cojn flip many times also. It seems the time remaining comes into play on multi player tourneys that complelty run out of time. There are other situations where oldest account wins. I think multi players when game ends due to skips. Not positive. But I know oldest account does come into play somewhere.

[Meeplelowda]

even if I had more remaining time

In addition to nik's comment, the time remaining on your clock only factors in if the total time for the round runs out without the game finishing.

[nik592]

From what has happened to me quite a few times. Is you lost the coin flip. It is random as I have won that cojn flip many times also. It seems the time remaining comes into play on multi player tourneys that complelty run out of time. There are other situations where oldest account wins. I think multi players when game ends due to skips. Not positive. But I know oldest account does come into play somewhere.

Yes, there are three scenarios where "ties" occur (or perhaps 4, if you count my first one here):

1. There is a tie, but the game has specified one or more tiebreakers (eg. remaining leftover money, last player, fewest unplayed cards, etc.) which break the tie and decide a winner. Winner (or whoever is ahead in the case of ties for second or lower places) progresses.
2. There is a tie, but no other tiebreaker has been specified, or specified tiebreakers do not break the tie. A player is randomly chosen to win the tie and progress (what happened here).
3. A player is booted from the game due to running over their allocated time. That player "loses", all other players win. For games with more than 2p, if additional players need to be eliminated from among the "winners" (eg. 4p game but only 2 advance), the oldest players advance.
4. The game reaches the maximum time allotted for each game in that tournament (eg. 15 days, 30 days, etc. - defined by the tournament creator). Places are awarded in order of most time remaining and players advance accordingly (depending on how many are at the table, and how many advance).

ETA: Some of this has never been confirmed by anyone official, or I think has occasionally been confirmed and then contradicted by two different official people, so it's a bit sketchy. Scenarios 1 and 4 are definite, as they are confirmed by on-screen text when they happen. Scenario 2 is the one I think that has had confirmation by someone at BGA, but may also have been contradicted by someone else at another point. Scenario 3 is purely anecdotal, but widely believed.

[yodamus]

Thanks for that detailed explanation. From my experiences. And I play a lot of tourneys. Ties are exactly as your 4 scenarios describe. Or so it seems

[Meeplelowda]

Scenario 3 is purely anecdotal, but widely believed.

"Purely anecdotal" denigrates it, unless you are simply defining that as "not officially announced by BGA." Instances of this have been statistically analyzed and zero counter-examples have ever been given.

[nik592]

"Purely anecdotal" denigrates it, unless you are simply defining that as "not officially announced by BGA." Instances of this have been statistically analyzed and zero counter-examples have ever been given.

I haven't seen any statistical analysis, but I have seen the odd person claim they saw a counter-example (but one is never shown). As far as I knew, it's based on a bunch of people saying, "yeah, that's what happened in my game" which is how I would define anecdotal (just based on what individuals have observed). If it's been analysed in some way, I guess that gives it more weight? Although I'm curious what a statistical analysis would look like, since as far as I've heard, any examples have confirmed it (so that's 100% confirmed). I personally believe it's correct, but I know there are those who don't.

[Meeplelowda]

"Purely anecdotal" denigrates it, unless you are simply defining that as "not officially announced by BGA." Instances of this have been statistically analyzed and zero counter-examples have ever been given.

I haven't seen any statistical analysis, but I have seen the odd person claim they saw a counter-example (but one is never shown). As far as I knew, it's based on a bunch of people saying, "yeah, that's what happened in my game" which is how I would define anecdotal (just based on what individuals have observed). If it's been analysed in some way, I guess that gives it more weight? Although I'm curious what a statistical analysis would look like, since as far as I've heard, any examples have confirmed it (so that's 100% confirmed). I personally believe it's correct, but I know there are those who don't.

Well there's what I've copied below. I used to be on the fence about this because almost always when someone complains about this, it is a case when the tournament has to choose which of two people advance (as in the OP for this thread), so basically a coin flip. Of course, if the coin always lands in favor of the older account, you start to question whether it's a fair coin.

But we had an example below where 5 out of 9 people were advancing. This wasn't just a coin flip. So the chances that only the oldest accounts advanced by chance out of all the possible combinations of players that could have advanced became exceedingly low. Was it impossible? No, but if it's not account age, then it's something that correlates so strongly with account age that it is indistinguishable.

« it goes on something else, possibly age of account (newer accounts are eliminated). »

possibly ? or sure ?
so it would be too bad for newers?

how can we know exactly what the basis is for eliminating the other players if only one player is expelled for exceeding the time limit?

How could we possibly know unless BGA told us? BGA hasn't told us, so we can only go by what we observe.

What we observe is that for games that end by someone being expelled (which is handled differently from the game ending properly or time for a round expiring), the oldest BGA accounts advance. Your game was another example:
Advanced:
MaxLeffe 2013-05-01
kattat 2020-03-29
Raccoon974 2020-11-22
DoinkGo 2022-06-30
LeLuHe 2023-03-15
------------
Eliminated:
seugrof 2023-10-19
Lie_cat 2023-11-14
bot_A_nik 2024-02-28
Syuby05 2024-04-15

The newest account to advance was created on 2023-03-15. All of the players who were eliminated had accounts newer than that. The odds of none of the 4 newest accounts advancing if they were selected randomly is about 0.8%.* Yes, one could say older accounts are more experienced, so maybe they advanced based on ELO. Good thought, but no. The two lowest ELO accounts in this game advanced and the second highest ELO account didn't.

By the way, I'm never collecting this data for a tournament again because 1) it's cumbersome, especially for nine players (although I guess it was great procrastination from real work), and 2) this example in particular where there are so many random ways to select who advances yet only the oldest accounts advanced seals it in my mind that it is definitely account age that determines this. I reject the null hypothesis.

*Math appendix:
For none of the newest accounts to be selected randomly, you'd have to select a player from the nine who were not expelled five times, and each time choose one of the 5 oldest accounts in the entire group. Let's color the 5 oldest accounts green the the newer accounts red for simplicity (pretend we wrote their usernames on a green or red ball and jumbled them in a bag). On the first draw it is actually more likely than not to draw an old (green) ball, because 5 of the 9 balls are green. But if we draw a green ball on the first draw, then on the next draw only 4 of the remaining 8 balls are green. We draw another green. On the third draw, only 3 of the remaining 7 are green. Etc. etc. until on the last draw only 1 of the remaining 5 balls is green. The chance of only drawing green each time is given by 5/9 * 4/8 * 3/7 * 2/6 * 1/5.

[BananaHarambe]

This is strictly multi elimination tournament issue. There must be a Bug it it’s specific coding where it doesn’t randomize who to keep.

Tied games in single elimination function fine with a fair coin flip to decide who advances.