BGNR and I are having a healthy discussion on Discord #events channel about the playoff format for Get2Gether.

Looking at Div1 Gold playoffs, these are the final seeds from the group stage:

And these are the brackets:

With this format:

*Seeds are based on trying to avoid rape games (keep "rank difference" equal, i.e. 1 vs 5, 2 vs 6, 3 vs 7)
*This means that the top seeded teams aren't rewarded by getting to play the worst teams
*For example, the team that finished in second position has a more difficult first round match than the team that finished in third position (looking at absolute rankings of who they face, not relative rank difference of matchups)
*Also as an example, the team that finished in 5th has a much more difficult first round (vs top seed) than the team that finished 7th or 9th (the winner plays 3rd seed)

The advantages of this format are that rape games might be less frequent. It might also make it more interesting for the some of the divisions to have more interesting matchups.

In my opinion, the problem with this format is that you're not rewarding teams for doing as well as possible during the group stage. A team that finished in the middle of the pack plays against a top seeded team, instead of playing against another "middle team" and one of them advancing to round 2.
Another problem in my opinion is that the system can be taken advantage of. It might be better to lose a match 2-0 than to try to win a map. For example:



What I am putting forward is the usual playoff seeding method:

*Use the regular season rankings and do 1 vs (9 or 10), 2 vs (7 or 8), 3 vs 6, and 4 vs 5
*This might lead to more rape games in the early rounds, but it correctly rewards teams that won more in the regular season
*Anyway, teams that didn't do well in the regular season get to keep playing in the "silver" playoffs
*There is no incentive to lose during regular season, winning more typically means better playoff position

The above is my point of view and heavily biased. BGNR has his own thoughts.

What do you think?!

I think anything else other than traditional system of playing highest placed teams vs lowest placed teams in first playoff rounds is trying to reinvent the bicycle?

What is the point of "avoiding rape games" anyway?

LOL, our team has no business in Div1 because we can't play tb3 for shit (well, at least me and gloomy can't) and yet we are neither suprised nor put off by losing heavily to other teams on maps we're vastly inferior on or to players who are simply better than us. We learn from that or least try to.

I see no point whatsoever in reducing the quality of latter stages of competition. Who would want that? If you want to play more games - just practice more often...

Blood_Dog: *Seeds are based on trying to avoid rape games (keep "rank difference" equal, i.e. 1 vs 5, 2 vs 6, 3 vs 7)

Bgnr: I already explained that was not the only reason.
Overall games will be more competitive, so the games design of bracket phase was also made to stimulate the competition, and provide more interesting games to follow for the community.


-------

Blood_Dog: "*This means that the top seeded teams aren't rewarded by getting to play the worst teams"

Bgnr: Why they should be rewarded easy games?
If they are good, shouldnt they be demanded the same as the others?
Why are easying it up to the ones that are good already?
What promotion of improvement are we doing with your suggestion?
Yet,and to avoid confusions, please note That
No1 and No2 are on the opposite ramifications of the bracket
Aswell no3 and 4 arent playing vs no1, no2 or with each other on 1st round


------

Blood_Dog: "*For example, the team that finished in second position has a more difficult first round match than the team that finished in third position (looking at absolute rankings of who they face, not relative rank difference of matchups)"

Bgnr: you Refering to group rank?
The 3rd in a group will go to Silver bracket, so he will not be in same competition.


-----

Blood_Dog: *Also as an example, the team that finished in 5th has a much more difficult first round (vs top seed) than the team that finished 7th or 9th (the winner plays 3rd seed)

Bgnr: Can you explain me why 7th vs 3th is diff in rank difference of 1st vs 5th?
7-3 = 4
5-1 =4
As already explained before.


-------

Blood_Dog: "In my opinion, the problem with this format is that you're not rewarding teams for doing as well as possible during the group stage. A team that finished in the middle of the pack plays against a top seeded team, instead of playing against another "middle team" and one of them advancing to round 2."

Bgnr: Was not choosed for the motives stated above


---------

Blood_Dog"Another problem in my opinion is that the system can be taken advantage of. It might be better to lose a match 2-0 than to try to win a map. For example:
4-team groups in regular season stage, 2 of the teams each won their first matches 2-0. For the last match, they play each other. They know they will both make it to Gold playoffs because no matter who wins that last match, they will finish top 2 in their group.
During this last match, one of the teams wins the first map. In the current playoff format, why would the team that lost the first map try and win the second map?
They would finish in a better playoff position (lower seed) if they lose the match 2-0 instead of trying hard but losing 2-1.
In fact, after losing the first map it's better to just /suicide all game in the next map to get a worse frag differential, just to be lower seeded (still within gold) and have easier playoffs."

Bgnr: In this model who finishes 1th to 4th will have easier games
So they are rewarded that way
Still the matches are kept the closest possible
And equal to every team pair
can you explain me how finishing 5th to 8th can give you an easier game?

------
------

Drake: I think anything else other than traditional system of playing highest placed teams vs lowest placed teams in first playoff rounds is trying to reinvent the bicycle

Bgnr: This model is not new, is just one adapted to use by me, considering the current specificities of the qw community


------

Drake: What is the point of "avoiding rape games" anyway?

Bgnr: I think this should be quite simple to grasp
Its refered vaguely above as specificities of qw community

-----

Drake: I see no point whatsoever in reducing the quality of latter stages of competition. Who would want that? If you want to play more games - just practice more often...
Bgnr: Can you explain me why within the present model, the quality of games is reduced??

It should be quite simple to grasp.
If you put better teams (those who did better in the groups) against each other, instead of putting them against worse teams, then you have chance of less 'better' teams in the latter stages of the playoffs and more 'worse' ones. Thus, instead of facing those other good teams in latter stages - good teams will face worse teams - and produce lesser quality of games. In other words: you are making the first round perhaps somewhat more even and with less 'rape games' - but it comes at the cost of less interesting further stages - because some of the better teams will be out, because they had to play vs the very best in the first round.

When you place someone against the 1st team - you "throw away" that team, almost for sure, unless it's 2nd or 3rd one who have some chances. Isn't it better to 'throw away' the team that performed the worst?

What I mean is that 1st team will have a guaranteed win vs 8th team, sure. 5th team though, might slip and actually lose vs 8th one (it's not probable, but suppose it's possible). And thus you have probability that in the second round you'll have 1st vs 8th. A 'rape game' that you've tried to avoid in the first round.

I don't think the difference is very dramatic, but I also don't think that any alteration to a standard 1-8 2-7 3-6 4-5 formula is needed. That formula is designed specifically to make latter stages of competitions as 'elite' as possible. Obviously, your way isn't THAT much different, it just introduces some possibility of having less interesting games further on.

I also don't understand what is meant by "current specificities of the qw community". I'm fairly sure anyone would define those in their own way and some definitions may border on being directly opposite .

But whatever, it's your tournament, experiment away. Maybe part of the problem of having rather sporadic tournaments is that people are naturally afraid of experimenting and thus finding good formats and practices - they're afraid to scare participants away. But we're at the stage when it should no longer matter, imo, so do whatever you want.

You will make some games more competitive, but other games less competitive. For example, in the first round, 4 vs 5 and 3 vs 6 should be very competitive because they are both "middle" teams playing each other, and anyone can win. Now, with 1 vs 5 and 2 vs 6, those will just be rape and the middle teams will be out in the first round.


No I'm referring to overall rankings after the group stage. Team #2 plays against team #6, but team #3 plays against a "worse" opponent in team #7. How come team #3 has an easier opponent than team #2? Why should a team care to get the best possible frag differential and finish in second place, when they now have a more difficult opponent than team #3? The reward for team #2 to do well in regular season is to play against a better opponent?
It's not about comparing 6-2 = 7-3, it's about how good the individual opponent is.


This is the point that I don't think I'm explaining well. It's not about comparing 2 different matches (7-3=4 vs 5-1=4). It's about comparing a single match. Team #5 usually plays against team #4. That's a difference of only 1. That's a close match and anybody can win! In your system, team #5 plays against team #1. That's a difference of 4, and basically impossible to win. -1 vs -4 is how I look at it (unfair for individual teams), but 4 vs 4 is how you look at it (fair between different matches).

I'm not sure if you understand my point better now? If so, do you see that there is now less motivation to win regular season games in some scenarios? Instead of finishing 5th, we could have easily finished 7th by "throwing" the last game 0-2 instead of trying to win and losing 1-2. Finishing 7th means we have much easier first 2 rounds of playoffs (vs team 9, then vs team 3), instead of vs team 1.


Yes I know that wasn't the reason it was chosen, but this is a side effect. Teams in the middle will lose in round 1, so round 2 will have rape games. Normally, round 1 has some rape games (1 vs 8, 2 vs 7), but middle teams have close matches against each other (3 vs 6, 4 vs 5). Then, in round 2 you should have top teams vs middle teams.


Team 4 vs team 5 (normal playoff bracket) is a closest possible match. Team 1 vs team 5 (your playoff bracket) is not close.


Again I think I'm not explaining myself well. It's not about "5th to 8th gives easier game". It's about finishing 7th gives easier opponent than finishing 5th. It is EASIER to beat team #3 than to beat team #1, isn't it? So, if you finish 7th, you play against an easier team. If you finish higher (like 5th) because you did better in regular season, you play against the best team?! Again, it's not about 5-1 = 4 versus 7 - 3 = 4 being "equal", it's about 5-4 = 1 versus 5 - 1 = 4 being "unfair".

If you make rank difference of 4 for every first round match, then every top seeded team will win in probably rape games (7 has no chance to win against 3, 6 has no chance to win against 2, 5 has no chance to win against 1). Of course I'm talking about Div1 Gold, maybe it's different in other divisions where teams are more even.
If you do it the "normal" way, some middle teams will definitely win first round in close games (either 4 or 5 will win in close match, and maybe even 6 can beat 3).

Does it make a bit more sense now? Basically, in normal playoff brackets you get some "big rape" games (1 vs 8, 2 vs 7), and some close games (3 vs 6, 4 vs 5). In your system, you get all "medium rape" games.


I agree with Drake and I hope I'm not offending you I'm only trying to explain why the "normal" system is almost universally used and why every other system has more problems, but I don't think I'm doing a good job.

Thanks,
BD

Drake in this format we got
1st vs 5th --
|
-
|
3rd vs 7th --|

4th vs 8th --
|
-
|
2nd vs 6th --|

I would say games in 1st round can be of good quality, contrary to what you've said

If the best seeded teams win
We've got

1st vs 3rd
and
2nd vs 4th

Those high quality games you talked about
And if a inferior seeded team wins, its because it deserves it, and we can consider it beeing in the end, equal or better in strenght to the highest seeded team it won

Do you still think we can have lesser quality games with this model?
Well i agree when you say the differences shouldnt be dramatic between the 2 models.
I avoided to the maximum experiements in the tournament, thats why map pool was TB5, even tough i would like to introduce Burialv2 by foogs or an finished version of Burial.
The decision of this model was to implement a better methodology to qw community actual configuration
In the end i will assess the result of the measures taken

And another factor that was already pointed out for this decisions, is for in every game a team "dream" that is possible to win
Not pre-determined games was the waay to go
Thats why i also picked on eql random pickup of map choice, to add some more unpredictability to who can have an advantage in picking his own map, and in detriment of the Thunderdome used model wich is the highest seeded team picks
This also comes to increase a little bit more the odds of the underdogs and make the best teams pull their sleeves to win

Thus increasing the game quality

that is not contrary to what I said at all .

Nevermind. BD's post is also explaining the very same thing I did in a much more detailed way.

bgnr wrote:
Blood_Dog: *Seeds are based on trying to avoid rape games (keep "rank difference" equal, i.e. 1 vs 5, 2 vs 6, 3 vs 7)

Bgnr: I already explained that was not the only reason.
Overall games will be more competitive, so the games design of bracket phase was also made to stimulate the competition, and provide more interesting games to follow for the community.

Blood_Dog:You will make some games more competitive, but other games less competitive. For example, in the first round, 4 vs 5 and 3 vs 6 should be very competitive because they are both "middle" teams playing each other, and anyone can win. Now, with 1 vs 5 and 2 vs 6, those will just be rape and the middle teams will be out in the first round.

Bgnr: In your suggestion we got 3 vs 6 that is almost same rank difference as the current model
And only one game, that is really hypothetically close.
One thing you should take into consideration is that teams werent all in same group, so a lot of randomness can appear due to that, wich could even make that 5vs4 diff than the predictable outcome.

-------------------


bgnr wrote:
Blood_Dog: "*For example, the team that finished in second position has a more difficult first round match than the team that finished in third position (looking at absolute rankings of who they face, not relative rank difference of matchups)"

Bgnr: you Refering to group rank?
The 3rd in a group will go to Silver bracket, so he will not be in same competition.

Blood_Dog:No I'm referring to overall rankings after the group stage. Team #2 plays against team #6, but team #3 plays against a "worse" opponent in team #7. How come team #3 has an easier opponent than team #2? Why should a team care to get the best possible frag differential and finish in second place, when they now have a more difficult opponent than team #3? The reward for team #2 to do well in regular season is to play against a better opponent?
It's not about comparing 6-2 = 7-3, it's about how good the individual opponent is.

Bgnr: Team 3 plays against a worse opponent than 6, its the 7th
But as i stated before same rank diff
Hypothetically same probabilities of winning on both matches, wich in reality is never like that. Just more probable.
Going to your initial statement
There are no easier games for any team

------------------------

bgnr wrote:
Blood_Dog: *Also as an example, the team that finished in 5th has a much more difficult first round (vs top seed) than the team that finished 7th or 9th (the winner plays 3rd seed)

Bgnr: Can you explain me why 7th vs 3th is diff in rank difference of 1st vs 5th?
7-3 = 4
5-1 =4
As already explained before.

Blood_Dog:This is the point that I don't think I'm explaining well. It's not about comparing 2 different matches (7-3=4 vs 5-1=4). It's about comparing a single match. Team #5 usually plays against team #4. That's a difference of only 1. That's a close match and anybody can win! In your system, team #5 plays against team #1. That's a difference of 4, and basically impossible to win. -1 vs -4 is how I look at it (unfair for individual teams), but 4 vs 4 is how you look at it (fair between different matches).

I'm not sure if you understand my point better now? If so, do you see that there is now less motivation to win regular season games in some scenarios? Instead of finishing 5th, we could have easily finished 7th by "throwing" the last game 0-2 instead of trying to win and losing 1-2. Finishing 7th means we have much easier first 2 rounds of playoffs (vs team 9, then vs team 3), instead of vs team 1.

Bgnr: Going to your initial statement
There are no easier games for any team

--------------


bgnr wrote:
Blood_Dog: "In my opinion, the problem with this format is that you're not rewarding teams for doing as well as possible during the group stage. A team that finished in the middle of the pack plays against a top seeded team, instead of playing against another "middle team" and one of them advancing to round 2."

Bgnr: Was not choosed for the motives stated above

Blood_Dog: Yes I know that wasn't the reason it was chosen, but this is a side effect. Teams in the middle will lose in round 1, so round 2 will have rape games. Normally, round 1 has some rape games (1 vs 8, 2 vs 7), but middle teams have close matches against each other (3 vs 6, 4 vs 5). Then, in round 2 you should have top teams vs middle teams.

Bgnr: I dont think team in middle will necesarily loose, but its not the same for them as is in your model
If you consider 3rd and 4th midlle teams like you are saying.
Going trough the most probable outcome, those middle teams will be anyways in the 1/2 final.
Your model can work for a very big players pool
And a motivated players pool
I dont think it fits best the qw scene
And still I have doubts even on a bigger motivated pool wich could be the best bracket format.
-----------------------


bgnr wrote:
Still the matches are kept the closest possible
And equal to every team pair

Blood_Dog:Team 4 vs team 5 (normal playoff bracket) is a closest possible match. Team 1 vs team 5 (your playoff bracket) is not close.

bgnr: Thats the so called "short blanket" model
If you have cold in the head and pull it up, you will uncover your feet
so to have some more close matches you are unbalancing others
Like i said before, that model could be used, but i dont think it fits best the qw scene


----------



bgnr wrote:
can you explain me how finishing 5th to 8th can give you an easier game?

Blood_Dog: Again I think I'm not explaining myself well. It's not about "5th to 8th gives easier game". It's about finishing 7th gives easier opponent than finishing 5th. It is EASIER to beat team #3 than to beat team #1, isn't it? So, if you finish 7th, you play against an easier team. If you finish higher (like 5th) because you did better in regular season, you play against the best team?! Again, it's not about 5-1 = 4 versus 7 - 3 = 4 being "equal", it's about 5-4 = 1 versus 5 - 1 = 4 being "unfair".

If you make rank difference of 4 for every first round match, then every top seeded team will win in probably rape games (7 has no chance to win against 3, 6 has no chance to win against 2, 5 has no chance to win against 1). Of course I'm talking about Div1 Gold, maybe it's different in other divisions where teams are more even.
If you do it the "normal" way, some middle teams will definitely win first round in close games (either 4 or 5 will win in close match, and maybe even 6 can beat 3).

Does it make a bit more sense now? Basically, in normal playoff brackets you get some "big rape" games (1 vs 8, 2 vs 7), and some close games (3 vs 6, 4 vs 5). In your system, you get all "medium rape" games.

Bgnr:
I think in end of tournament we will see how things went
the limits of a population have always its "outlyers"
So in a normal U populationn curve, its expected to have more unblance in a sample of this extremities
Talking about Div1 (Maybe 3 teams) Gold and Div2 Silver (MAybe 1 team)
Imo that even further validates my choice to avoid those hard games but can make like harder for the middle
Yet the principle of protecting the teams value is present here
Thats why teams 1 to 4th were "rewarded" with less strong opponents and avoided meating each other
Still i think the overall rank distance can make those inferiour ranked teams to dream
And we always have randomness
Like WO, and having for example a Rune team in 8th place
And other randomness coming from different overall strenght of groups quality, not every group is equal in reality
Aswell some randomness coming from the superation of teams, like TL beating BFK, due to their superation

(Edited 2018-02-22, 18:18)

Drake the phrase was took out of context
Or i didnt explained the context very well
From the argument you took that phrase i tryed to show round 1 and roun2 games are of good quality
That is contrary to having less quality games as you stated

You say that because you look at "rank difference = 4" for every match, so for you every team has the same difficulty in their matches. That's why "there are no easier games for any team".

BUT, that is only if comparing relative difficulty between different matches. For me it doesn't matter that Aeronauts' match is "as difficult" as TTP's match. Aeronauts only cares about their match, not the relative difficult of "their match compared to the other matches".

What I'm saying is that Aeronauts has to play a more difficult opponent than HNSM does. A more difficult opponent is a more difficult match than it should be, based on both teams' efforts in the regular season.

That's pretty much the reason we disagree. You are looking at comparing different matches to each other. I'm looking at comparing what each individual match should be. In your case every match is equal difficulty compared to other matches. In my case an individual match is more difficult than it should be. Otherwise, why try hard to finish higher in the standings during the group stage?

To classify yourself as 1st or 2nd in your group wich can be an uncertainty until last match.
And if you took attention to the group phase games schedule no1 vs no2 seed games was the last, ending 1st in your group would probbaly qualify you between 1st and 4th, in overall table for the bracket, so,

To finish 1st to 4th in your brackets table, as others wlll render you in an hypothetical situation of an superior opponent
And for the obvious reason of show your best, improve during the tournament itself to aim a bit higher than you were able to before the tournament started

So what you are saying of no trying to make an effort was also taken into consideration to the group phase games calendar

Put it more practical, lets imagine 2 teams are the no1 and no2 seed in their group, if that holds true and they win their 2 first games vs no3 and no4 seeded teams, their last game will be between themselves, wich will render one team with most points and probably classifying this team in the 1st to 4th place in the overall bracket table, giving them the advantage of playing a bit low seeded team.
So with this design, and in this specific case, you should make all your effeorts until your last game to qualify as 1st in your group

And if some of the no3 and no4 team gets in the middle of the first places fight, maybe you will still be trying to win them to guarantee your spot in the gold bracket and not get demoted to Silver

As a side note, and as a proposal from another player
Maybe making a pool of 1st place teams, spread them randomly across the bracket without facing each other's
And then picking the 2nd place teams and randomly spread them across the available spots can be a solution

Random as it is
It less guarantees that effort of the teams till the end in frag making
But still it seems a model better than the 1 vs 8 kind of bracket seeding

Actually this model proposed by a player is used aswell in professional sports like Tennis or Soccer
Yet even in these models you can have Admin interference, like:
- Setting heads of bracket
- Wildcards

Edit
With the following image im trying to explain the text above.
If the no1 and no2 seeded teams (Team A and B) win their round1 and round2 matches, wich hypothetically is the most probable, as they enter the last round they will be playing each other.
In this situation every team would still want to win their last matches, even the 3rd and 4th teams.
Because a win in this situation probably will make into no1 to 4 in the overall bracket table
And as we have seen before, game pairings are 1 vs 5, 2 vs 6, 3 vs 7 and 4 vs 8
So any team would like to finish 1st to 4th and for that to be a reality hey would like to win last match

In this another example, were a lowest seeded team steal a game to a highest seeded team
Still on last round everybody wants to win
Team A with 4 points dont want to loose to team B, because team B could finish 1st place
Team B obviously wants to win because of wanting to go Gold bracket
Team C wants to win because can go to Gold in case B looses
Team D wants to win, because it wants to be no 3rd in group to be in the 1st to 4th places in Silver bracket

Other scenarios can happen, still it seems to me every team will still want to win in group phase to reach the best scenario for them


Certainly there arent perfect scenarios and I will take my conlcusions about the righteousness about this model i picked based on conscious decisions

(Edited 2018-02-22, 21:35)

Not exactly, but this is easier to explain in person than in text so it's not worth discussing