Skip to main content
Welcome guest. | Register | Login | Post

Organizing Armagetron Advanced tournament

45 replies [Last post]
Offline
Joined: 2007-09-10
I'm proposing a new award

I'm proposing a new award for the tournament:
Longest and best structured post on the nuxified forum. ^^

Offline
Joined: 2007-09-10
>> Then we split these 6

>> Then we split these 6 into groups of 3

>> But anyway, these are the number's we're working with if we have 12 players. They'd obviously increase if we have more. It's possible we'll have 16 or at best even 20 players in which case we'd have 4 or 5 groups of four which gives further opportunity for matches.

For the group of 3 scenario, I preliminarily propose we use 4, 2, and 1 point [below, I recommend something else]. This strikes a better balance than 4, 3, 2, which is what would apply using the 4-3-2-1 system. Otherwise, if we use 4-3-2, we are giving the AA LMS winner a significant *advantage* over the other tournament winners competiting for UW. In this last case the average is 3 (vs 2.5 for 1-1).

For the qualifier cases, we may not have an even multiple of 4 players to work with. I suggest we do an extra group of 3 (using the point variation proposed above or some comparable/better variation) if 3 is the remainder (ie, for 11, 15, 19, etc).

Otherwise, we can solve the problem (remainder of 1 or 2) by using 1 or 2 groups of 5. For points, we might leave things as they are (ie, 4, 3, 2, 1, 1). This case would result a small but nonnegligible disadvantage (2.2 vs 2.5). We can also try 4, 3, 2, 2, 1 or 4, 3, 3, 2, 1. The former gives only a slight disadvange (2.4 vs 2.5); the latter gives a slight advantage (2.6 vs 2.5).

Another option is to stay with the current point system. This means we would use 4-3-2 (ave: 3) for the case of 3 during the tournament rounds and use 4-3-2-1-1 (ave: 2.2) for the case of 5 during the qualifier rounds. During qualifiers, a disadvantage would accrue. During the tournament stage, a compensating advantage would erase some or all of the accumulated disadvantage.

BUT, there is no guarantee that groups of size 3 will be needed in the tournament rounds if we have a different number of starting players than 12. Similarly, 5 may not be used at all, or it may even be needed for the tournament rounds. Thus, I recommend we strive for parity with other tournament systems...

... at every step of the way. This means: use "4, 3, 2.5, 2, 1" (ave: 2.5) for groups of 5 and "4, 2.5, 1" (ave: 2.5) for groups of 3.. whenever we must resort to groupings of either of these sizes.

These last set of numbers is slightly uglier than normal but that may be deemed fine. OTOH, none of this analysis is "perfect" so we can just try to approximate the situation and hope for the best. [I recommend the ugly numbers though Smiling ]

>> Then the lower two and the higher two both play 1vs1. This gives us the winner, second place and third place.

I like the idea. We would use the 1-1 point system right? [Ie, 4-2-1-0 .. or simply 4-1 since a tie won't be acceptable.]

How would we deal with ties in tournament rounds (single, double.. elimination) in general?

[Reference also the comment I made at the bottom of the prior post. It relates to ties in AA LMS. The posting is http://www.nuxified.org/topic/organizing_armagetron_advanced_tournament#comment-11151 ]

>> Now if that seems low I suppose we can further adjust the numbers... After the tournament we'd then have total of 26 matches.

You may want to specify that the exact number of rounds, matches, etc, is subject to change depending on user participation, user demand, and organizational resources. Smiling

An almost final word: (from the prior post I made,) if one tournament has a significantly lower point average compared to other tournaments, it means that a player that is way above average in this particular tournament will still only have the same point average as a player that is average or slightly above average in those other tournaments. Since these average scores are essentially what determine UW, I think we should try and match them up as well as possible for the sake of fairness to above average tournament winners that want to win UW without unreasonable biases against them (as likely indicated by having won in a tournament with a low points/game average relative to other tournaments).

Final word:

I'll recap the averages of the various point systems mentioned above. I'll ignore ties and forfeits.

3/gp:
4-3-2 => 3
4-2-1 => 2.33
4-2.5-1 => 2.5 [my vote]

5/gp:
4-3-2-1-1 => 2.2
4-3-3-2-1 => 2.6
4-3-2-2-1 => 2.4
4-3-2.5-2-1 => 2.5 [my vote]

1-1:
4-1 => 2.5

4/gp:
4-3-2-1 => 2.5

libervisco's picture
Offline
Joined: 2006-05-04
Thanks, I'll refer to this

Thanks, I'll refer to this post depending on how many players we end up with and what numbers will those be. In any case, I see the bottom line, make the averages match up to other tourneys and so the golden number is 2.5. Your chart above helps.

Thanks

Offline
Joined: 2007-09-10
>> Wouldn't it be easier to

>> Wouldn't it be easier to just increase the number of rounds per match then? You initially said 10 rounds, but, guess what, the DS server on which I met with QBrain today has 100 rounds! Laughing out loud

>> The sky is the limit, if we want more play time let's just play longer matches.

I like this idea with the following two footnotes.

First is that unfortunately no reshuffling would be done (I suspect this in part motivates you Smiling ).

Second is that if we only play 3 matches, that takes away the ability for people that were caught unprepared to practice to come back and have a reasonable chance to win it all.

In consideration of this second situation, I say that we increase the number of matches to a number larger than three, ie, allow for more than three gathering of combatants.

Also, long matches may present some limiting situations for those that cannot readily play long matches.

Hence I will now provide us with "the perfect distro" [distro is short for distribution]

5 matches at 10 rounds followed by
3 matches at 30 rounds followed by
1 match at 100000 rounds!!

OK. The numbers were just grabbed from the air.

Perhaps we can present a few choices that can be voted on?

>> I know you say it's not too complicated to reshuffle

At least once I said, " ****** if ****** randomizing can be achieved with little extra effort...." so don't put words in my mouth .. even if they are tasty words... ummm.. tasty words.. gargle gargle drip drip shuffle shuffle.

>> We'll have to group everyone taking their location into account (so that we can easier deal with ping issues).

Good point.. those devilish pesky details!

>> But let's close that chapter. It's 4 players/teams per group (or as close to it as possible) everywhere, in every tournament, including AA.

I failed to read this before writing my last post.

"Can't umount 'that chapter' while files are held open."

So.. maybe for the next round of replies that chapter can be properly umounted.

[Note to self: make sure to hold all files on that filesystem open; I'd hate to close a good book. Sticking out tongue ]

>> I'd be interested to hear if AA servers ever recognize ties, although from what I can see the chances of that happening are very slim.

Actually, at first it seems you can't have ties, but we are talking about who was the LSM the most number of times during the match (as opposed to just finding the LSM/winner for the current round)? In this case, it's very conceivable to have the top two players in (eg) 10 rounds be those two that ended up with 3 victories each, thereby tying for first.

When I played, I saw constant references to my ranking on the ladder moving up and down. I also noticed I was gaining or losing points with fractional components (eg, 1.18 points). My guess is that the server keeps score and rankings but uses an elaborate system that takes into account perhaps things like how many other players crashed against your wall. It may even take into account the strength of your opponents based on their current or their "global" ranking.

>> If it would ever happen that there is a tie I'd say, give 4 points to both of the winners. Skip 3 points and give 2 and 1 to the rest. Basically then it's a case of one of the tied players, instead of settling with second place, stealing that one additional point with a shared first place.

I was about to agree and say how this is somewhat standard.. except that I couldn't shake the notion that the 1-1 point system (and other point systems) gives 2 points for a tie (not 4).

I think the most accurate would be to give 3.5 points to each of the two players tying for first. It would be like saying that they each got first and second place so get the average of those two point scores (in a similar way as 2 is roughly the average of 4 and 1).

How willing are we to move into fractional points in our tug-of-war between fairness and simplicity?

One way that will likely avoid fractions would be to take all our point systems and multiply the values by 10.

Win: 40 points
Second: 30 points ...

This doesn't look too bad and sort of keeps things fairly simple while allowing for a little more levelness among the tournaments.

In fact, 40 arguably looks cooler (for a fighting game) than 4. Why 4 anyway? ...Roses are roses.

This would fix the ugliness in the last post where I suggested we use 4, 3, 2.5, 2, 1 as points in a 5 group match (or 4 2.5 1 for a group os size 3). The reborn swan would now look like this: 40, 30, 25, 20, 10. ..Hubba hubba (whistle whistle). Mucho prettier, no?

If I could say one more thing on reshuffling.. It seems we will drop that, but let's try and work out a system or a criteria check to avoid having odd situations where say the best two players constantly meet each other. It might be extra fun for everyone, but it will hurt the goal of winning UW for whichever one of them ultimately wins AA.

In one approach to identifying a bad initial shuffle (ummm, shuffle), we might ask players to rank themselves 1-3 and keep that in mind as groups are decided.

Alternatively, we might do a few (nonscientific) preliminary matches and use those rankings as a guide.

The idea would be to avoid putting the top two players in the same group, and in general, not to put too many very good players in the same group. We should focus on protecting the better players since UW will only applies to the winner.

Of course, Murphy's law states that the 20th seed will win AA.

Offline
Joined: 2007-09-10
I don't have time right

I don't have time right now, but when I get back, I'll think about this some more and maybe reply more directly.

Related to a different discussion about the number of matches, I would say that there is room for various parameters (eg, rubber) to be varied througout the course of the qualifiers if there are enough matches in the qualifiers.

Participants may like that a particular group of matches was held under friendly conditions and another group under hostile conditions.

In battle, one must be prepared for anything!

Offline
Joined: 2007-11-11
/me doesnt want to read

/me doesnt want to read everything but he just wanted to explain a few things.

the 100 rounds on our server are there, because we dont want to end the match because of rounds. You simply cannot play 100 rounds when a player doesnt have 10 points or in the given ammount of time (a match has also a time yes). Its only possible if you really want to do it. But thats not the intention in the game is it?

about the point system: i think this is just a very small detail that gets high priority now. We shouldnt discuss this, and I personally think Libervisco is smart enough to find out the point system himself.

about the ties: Not possible in high rubber. thats a fact. its even this way: if 2 players get 10 points at the VERY same time (happend to me twice in 4 years...) it automaticly starts a new round and whoever gets 11 points is the winner then. This is not something we can set, this is something the server does automaticly... again, to summarize: Tie = out of the picture, its simply is not possible.

like i said, didnt read everything, any questions? feel free to ask Smiling

-Q

libervisco's picture
Offline
Joined: 2006-05-04
Well that clears up some

Well that clears up some things.

So basically, each match is played until someone hits 10 points. How many rounds can that last on average?

About the points system I'll refer to Jose's charts in order to make sure that the averages fit the ones in standard 1vs1 matches, for the sake of better balance between AA and the other tournaments.

Jose wrote:

At least once I said, " ****** if ****** randomizing can be achieved with little extra effort...." so don't put words in my mouth .. even if they are tasty words... ummm.. tasty words.. gargle gargle drip drip shuffle shuffle.

Sorry about that. Smiling

Jose wrote:

One way that will likely avoid fractions would be to take all our point systems and multiply the values by 10.

Interesting idea, but I'm not sure about changing it right now. It might be a bit too late to suddenly do this right before the tournaments.. Fractions of the .5 type aren't that terrible if we must have them..

If anyone else would like that change done, feel free to say so though, but the thing is, in this topic we're discussing AA rules. Other tournaments and the general rule set are pretty much set by now.

Jose wrote:

If I could say one more thing on reshuffling.. It seems we will drop that, but let's try and work out a system or a criteria check to avoid having odd situations where say the best two players constantly meet each other. It might be extra fun for everyone, but it will hurt the goal of winning UW for whichever one of them ultimately wins AA.

Who knows if two best players will end up in the same group? It's all random luck. But even if they do, they will play each other only these four times, and not just each other, but the other two from the group as well. Who is to say that each time they'll be scoring the same? If they're two best players, sometimes one will win and sometimes another..

Anyway, we could perhaps do some preliminary matches just to see who's who and try to group people in a sensible way, but that's about it.

Note that other tournaments will have same kinds of qualification groups that will also be determined randomly. In them too it could happen that two very good players end up being in the same group, which potentially evens the odds across tournaments.

Nobody denies that luck will have a part in selecting the winner of the game fest. If everything was to be determined with mathematical precision I think it would be a bit more boring and dull world. Also don't forget luck has mathematics of its own, just as many other things in the universe. What differs it from the math that we know is that it is a mystery and mystery has mysterious value of its own.. But I'm going on a tangent. Smiling

Cheers

Offline
Joined: 2007-09-10
I feel compelled to

I feel compelled to register my vote of sadness if this means that an unskilled player will be out of the competition entirely after 4 matches.

On the bright side, that will lead to a minimal amount of torturous game footage and us noobs will be able to get on with other projects all the quicker.

Having registered my disappointment, I'll now try and contribute something quasi-useful.

Here is an outline of the rest of this post:
-- Describe the games played from the pov of the players. The point is to highlight how many matches the various players will play. If anyone goes through this and thinks it disagrees with the parent post, please comment.
-- Ask about how the lower players in each group will be known.
-- Miscellaneous comment

****

Based on my understanding of the description given in the parent post, the following list shows a description of the matches from the pov of a participating player. I break down the players based on how well they did, grouping together players that ended up with the same end result. Twelve players enter ring. 3 emerge with a recognized ranking (1st, 2nd, or 3rd).

The key finding for the 12 player setup is:
6 players will play 4 games before elimination
2 players will play 7 games before elimination
2 players will play 8 games; one gets 3rd, the other is eliminated
2 players will play 8 games; one gets 1st, the other 2nd

For example, let's describe the journey of those making up the third line above. Each of these 2 played in 8 games total. Each played in 4 qualifier matches which included 3 other players, and each came out in either first or second place within the respective group. Each played 3 first round tournament matches against 2 other players, and each came out in second place within the respective group. They then played each other 1 time to determine who got third place.

Second example. The top two players followed a similar path as just described, except... they each came in first place in their group from the first round of the tournament. After which they played each other for the top spot.

I will take a guess at what might happen if we have 16 players instead of twelve. I assumed that the first round of the tournament would be played as 2 groups of 4 (since 8 qualify) with the lower two in each group being eliminated leading to a final 4 showdown as described in parent post.

12 players will play 4 games before elimination
4 players will play 8 games before elimination
2 players will play 9 games; one gets 3rd, the other is eliminated
2 players will play 9 games; one gets 1st, the other 2nd

Total number of matches in the case of 16 starting players would be: 4 qualifer matches times 4 groups = 16; plus 4 first round tournament matches times 2 groups = 8; plus 2 more games. Total = 16 + 8 + 2 = 26.

Is this correct? My intent was to make sure everyone understands that despite the large numbers (eg 20), that half the players will play less than a hand full of games and the rest about twice that many. Also, I wanted to fill in a few details missing from the parent post as I understood it. I know that these numbers may change later. Hopefully, players will cry out for more games Smiling

****

>> Then based on the charts we get

Is this something that is produced by the server? I am still learning how the server (or we) comes up with winner of a match. After reading Qbrain's Nov 22 reply and playing for a while longer online, I believe it's the first to 10(?) points and not the winner of ten matches. So does the server spit out the rankings? If not, how would we know who got second place in order to drop the lower ranked players?

****

>> regrouping doesn't sound very exciting

From organizational pov I suppose. From a player perspective it would mean more games and against different opponents. How could that not sound exciting?

A suggestion is that maybe all that want more games will help with the organization. It's easy for me to ask for more games, but I need to roll up my sleeve a bit more. Yes, I am talking to myself. These are thoughts many people wanting more games might have at one point or another.

Offline
Joined: 2007-09-10
I misunderstood how the

I misunderstood how the server determined winners. I see now that it isn't the best performer out of X matches but the first to Y points.

Offline
Joined: 2007-09-10
>> At least once I said, "

>> At least once I said, " ****** if ****** randomizing can be achieved with little extra effort...." so don't put words in my mouth .. even if they are tasty words... ummm.. tasty words.. gargle gargle drip drip shuffle shuffle.

>> Sorry about that. Smiling

That was fake anger. ... Maybe your apology was also fake!

Offline
Joined: 2007-09-10
[This comment is only

[This comment is only intended to clear up the record a bit.. it might be irrelevant to 99.9% of the people reading it.]

I believe in thoroughness because anything less leads to unnecessary mistakes and confusion. That's just me though, and there is a time and a place for practically everything (ie, here isn't the place for some things).

As an example of how the "gut" can fail, I think the "mistake" I mentioned above may not have been a mistake after all. I won't know without analyzing the situation well and putting in some numbers. I don't plan to do that, but I'll say this. I do think that putting in the top players in the same group throughout all gameplay is likely not to be good for those players' overall averages. What confused me above is that I mixed their averages with the average of the whole field. The whole field average is not changed by matching up top players against each other repeatedly, but the average of those top players will/should suffer. Another way to put this is that making sure the average points per game match across tournament point systems is not enough to "ensure" fairness to the top players. But how much further do we go? Anyway, I was trying to identify specific cases of "unfairness" that could hopefully be resolved easily. As is, things are pretty decent because most people basically have a similarly fair shot at the same points no matter where they are competing. For a tournament on a limit budget, we shouldn't expect Olympics precision. We should try and minimize people complaining that this or that game got much more favorable treatment for winning the Ultimate Winner title/prize. That is a lower standard than "Olympics precision."

Offline
Joined: 2007-09-10
>> Another problem I see

>> Another problem I see with highrubber is all the extra rules you need to make. For example, while I don't deny that camping exists in lowrubber, camping is a much bigger problem in highrubber because you only need a very limited space to be able to live 'forever'. Next to a no-camping rule, you'll also need to get a consensus about instants, closing, speeding, etc.

Just experienced what you mean here. It's true that you can go for a long time this way, but it's tiring and subject to mistakes if you are in a tight enclosure. Plus, don't people sometimes like to see endurance contests? I'd liken "rubber camping" to a tight rope walker staying standing on the line. It's a larger feat than merely "camping." At some point you even admire the skill and not mind that person winning.

Anyway, I am new enough to this that I wouldn't mind seeing both low and high rubber matches.

.. and I might have been wrong thinking that high rubber is probably more of a benefit to low skilled players (that have learned the basics) than to high skilled players. I haven't seen that many matches.

Offline
Joined: 2007-11-11
to clarify: most matches

to clarify: most matches (clan matches or normal friendly 'matches') are: "best of X matches from first to Y points"

Offline
Joined: 2007-12-02
first of all!

im the from CTF/watever clan AoS and i think we'd appretiate something that we can participate in.
mabye we can have a LMS tournie where the rubber is medium? around 10?
or a dogfite.
and about CTF being not popular...
its full almost all the time!

Offline
Joined: 2007-09-10
I would like to drop out of

I would like to drop out of the Armagetron Advanced, Nexuiz CTF and Open Arena CTF tournaments.
The main reason is that I probably won't have a permanent internet connection this month, so scheduling matches might get even more complicated than it is now.

Comment viewing options