MAME Knockout Championships 5

Discussion about the MAME Championships
http://sawys.ifrance.com/K7.htm

Moderator: Knockout Coordinators

User avatar
The TJT
MARPaholic
MARPaholic
Posts: 2479
Joined: Wed Mar 06, 2002 10:56 am
Location: 20 Grand Palace

Post by The TJT »

Notman,
I'm absolutely not trying to stall the tournament or to be a nuisance in any way. I was just saying my opinion on what would be the best for the tournament.
I feel if we go with current final format, loserside is actually not fighting for the win, but 2nd place: winning side finalist could focus 99% on the other game, since he needs only one gamevictory...while loser side finalist should, logically, divide hes focus 50-50 between the two games, since he can't afford to lose in either of the games and cannot guess how much other player focuses on either game. You guys might have originally missed this tactical finesse?

If you stay with your original plan, then so be it. I'm sure game selections will be the crucial part in these kind of tournaments anyway.
---------

P.S. Saw this at some banner, now if you want loads of gamers to take part...
Attachments
Knock out advertising
Knock out advertising
gamers.JPG (20.62 KiB) Viewed 10608 times
User avatar
NotMan
MARP Knight
MARP Knight
Posts: 280
Joined: Fri Feb 14, 2003 12:00 pm

Post by NotMan »

Well....let say if the winner of the winner bracket could beat the hell out of the loser of winner brackets 2 games in row to prove that he doesn't need one game to prove to be a champion, but the truest fighting champion ever. IMHO. :D It could happen, cuz you never know.

Nah! We don't need that banners, but thanks anyways. All we need is a banner for Zlk for his next bounty. TEENAGE LUST SHUMPS. lol!

NotMan
User avatar
TRB_MetroidTeam
MARPaholic
MARPaholic
Posts: 1055
Joined: Fri Oct 22, 2004 4:50 am
Location: PR - Brazil

Post by TRB_MetroidTeam »

Well... my points...
1 - I didn't liked this rule to the final;
2 - At my knowledge (at least during ALL the CounterStrike tournaments ;))the "essence" of Double-Elimination is to have the 1st LOSER vs WINNER Final Match/Round. IF the LOSER have success here, so we run the 2nd Final Match/Round... and who win will be the champion!
3 - If used "2" rule, the match/round need to be also "7 days" each match/round (To avoid this situation... player play MAINLY or ONLY at one week day)!

The current "final rule" don't seems good IMO, even so if I am the Winner's Bracket player. I agree with changes on it, if possible.
Valter "TRB" - The Real Battle
Metroid Team, from Brazil
Crazy Taxi and Crazy Taxi 2 world champions
User avatar
TRB_MetroidTeam
MARPaholic
MARPaholic
Posts: 1055
Joined: Fri Oct 22, 2004 4:50 am
Location: PR - Brazil

Post by TRB_MetroidTeam »

Hmmm... other suggestion...
As we changed a lot of the "essence" of the Knockout competition... I think we would change the name of this new competition to "Head-To-Head"... and leave the "old" "Knockout" format as is... because it is very good too... but different. And we can use Knockout format again in the future...

What do you think about???

PS.: The HEAD-TO-HEAD format is very good and exciting IMO.
Valter "TRB" - The Real Battle
Metroid Team, from Brazil
Crazy Taxi and Crazy Taxi 2 world champions
User avatar
TRB_MetroidTeam
MARPaholic
MARPaholic
Posts: 1055
Joined: Fri Oct 22, 2004 4:50 am
Location: PR - Brazil

Post by TRB_MetroidTeam »

A good "Double-Elimination" sheet/table. It is from olds Counter-Strike tournaments. I know this exact sheet can be found on internet, to be filled... however I dont know now where.

"Perdedor" means "Looser". And "Vencedor" means "Winner".

(BBH edit: large images that stretch out the page make baby Jesus cry. Let's try this instead.)

Image
Valter "TRB" - The Real Battle
Metroid Team, from Brazil
Crazy Taxi and Crazy Taxi 2 world champions
User avatar
Zhorik
MARP Seer
MARP Seer
Posts: 788
Joined: Mon Sep 27, 2004 8:16 pm

Post by Zhorik »

TRB_MetroidTeam wrote:Hmmm... other suggestion...
As we changed a lot of the "essence" of the Knockout competition... I think we would change the name of this new competition to "Head-To-Head"... and leave the "old" "Knockout" format as is... because it is very good too... but different. And we can use Knockout format again in the future...

What do you think about???

PS.: The HEAD-TO-HEAD format is very good and exciting IMO.
Some old knockouts were head-to-head after a certain round as well. For example, K3 was head-to-head single elimination from the round of 8 (see viewtopic.php?t=10466 ). At least with the proposed double elimination format you have a chance of winning after a loss; with single elimination, you are out. If you don't want to play in the loser's bracket because its too hard, then DON'T LOSE! Even if you have the scenario where the player from the winners bracket only plays one game in the final, then if you think the other player is doing that and you are from losers bracket, pick only one of the games to play yourself and play the other one for a half day. If you guess right (50%) and can beat the player at the game you pick (50%), you have about a 25% chance of winning! If instead you use the rule that the losers bracket champion must beat the winners bracket champion twice, this is also 50% for the first match and 50% for the second match: again 25%.
User avatar
MJS
MARP Knight
MARP Knight
Posts: 367
Joined: Mon Jul 29, 2002 10:07 pm
Location: Buenos Aires, Argentina

Post by MJS »

Yeah, to tell the truth making the final a "two games at once" approach was a last minute idea and maybe we didn't put a lot of thinking into it. At least I didn't think of the tactic you describe.

However, I don't think it makes things much harder for the loser bracket player to win the tournament (than the "one game at a time" approach). It just introduces this new option for the winner bracket player to decide if he wants to focus on just one game or both, which I don't think it is such a big deal, because if he goes with this tactic, he would be betting everything in one of the games which is a bit risky IMO... at least I wouldn't do that!

And keep in mind we are choosing games more reliant on skill than easily-copyable patterns or subject to memorization, so that the champion isn't the one who had the most time to play but someone who knows his shit in a good variety of games! (sorry for the language) :P

So our proposal is: let's try this new method and see how it goes... or would it prevent someone entering the tournament???
User avatar
BBH
Editor
Posts: 1582
Joined: Wed Feb 06, 2002 7:06 am
Location: Portland, Oregon
Contact:

Post by BBH »

TRB_MetroidTeam wrote:Hmmm... other suggestion...
As we changed a lot of the "essence" of the Knockout competition... I think we would change the name of this new competition to "Head-To-Head"... and leave the "old" "Knockout" format as is... because it is very good too... but different. And we can use Knockout format again in the future...

What do you think about???

PS.: The HEAD-TO-HEAD format is very good and exciting IMO.
as Zhorik said, the head-to-head format has been in place in K3 (and K2 as well). But both of those only had bracket play for the top 8, and single elimination. Imagine getting to the round of 8, then getting a game genre you really suck at, and you have to play against someone who's really good at games of that type. It can still happen here, but with double elimination, even if you're unlucky enough to get placed in a situation like that, you're not out of the tournament completely. Besides, the old "knockout" elimination rounds will still be in place at the beginning, assuming more than 16 people decide to play in this, of course.

Now, getting back to the whole issue of the finals. It might be an interesting idea, it might be a bad idea. But we'll never know for sure unless we actually try it out and see what happens. And no, I don't agree with the basis that all the Winner's Bracket has to do is focus on one game and they'll win... what if it just so happens that the person in the Loser's Bracket is better at the game in question, and ends up winning that game? There's a lot of wasted effort that the Winner's Bracket player could have spent on the other game, and just like that, it's a possible upset and they could lose. It could very well happen. But nobody knows for sure at this point in time.
User avatar
TRB_MetroidTeam
MARPaholic
MARPaholic
Posts: 1055
Joined: Fri Oct 22, 2004 4:50 am
Location: PR - Brazil

Post by TRB_MetroidTeam »

BBH wrote:(BBH edit: large images that stretch out the page make baby Jesus cry. Let's try this instead.)
Pardon! And I will pray more next time...

And George can be right in the math, however it yet will be very less fair and very less exciting than if not used this "incredible new" finals method! And I am sure that everyone here, in that situation will use it (concentrate a lot in one game) in your advantage... and yes, yes... I know the "Winner's" need to be some advantage.

Just my opinion!
Valter "TRB" - The Real Battle
Metroid Team, from Brazil
Crazy Taxi and Crazy Taxi 2 world champions
User avatar
destructor
MARPaholic
MARPaholic
Posts: 1972
Joined: Fri Sep 20, 2002 9:38 am
Location: Poland

Post by destructor »

Because rules are long and somebody don't understand them, I prepared rules in Polish language.

-------Polski/Polish---------

Aby wystartować w turnieju najpierw trzeba zarejestrować się na MARP:
http://marp.retrogames.com/index.cgi?mo ... te&tourn=0
Przestrzegać zasad MARP:
http://marp.retrogames.com/rules/

Strona turnieju:
http://bbh.marpirc.net/k5/
Forum turnieju:
viewforum.php?f=7

PODSTAWOWE ZASADY

- w turnieju będziemy używać emulatora WolfMAMEPlus-106. Żadne inne wersje WolfMAME nie będą akceptowane. W przypadku problemów z odtwarzaniem recinputów lub innych problemów wcześniejsze wersje WolfMAME mogą być dopuszczone. Gry do turnieju zostały przetestowane w celu uniknięcia wszelkich problemów.

- wrzucając wynik należy pamiętać aby oprócz pliku .inp dołączyć też plik .wlf, który tworzy się też podczas nagrania w tym samym katalogu co plik .inp. Wyniki wrzucamy na MARP przez zakładkę Submit

- zalecane jest stosowanie nazw plików .inp/.wlf zgodnych z nazwą romu w celu ułatwienia pracy sprawdzającym wyniki.

- zgodnie z zasadami MARP gramy tylko jednym kredytem, nie używamy kontynuacji ani nie gramy ponownie na tym samym nagraniu po pierwszym straconym kredycie.

- recinputy muszą być wrzucane na MARP w ciągu 48 godzin od nagrania z wyjątkiem ostatniego dnia każdej rundy gdzie czas ten wynosi 24 godziny.

- w grach używamy domyślnych (default) ustawień. Jeśli ktoś nie jest pewien swoich ustawień dla którejś z gier może wykasować odpowiedni plik .cfg z katalogu Cfg. Pomimo że na MARP można używać trudniejszych ustawień w tym turnieju nie jest to dozwolone. W grach które używają biosów należy upewnić się że bios jest ustawiony na domyślny. Użycie innego biosu może spowodować problemy w odtwarzaniu recinputów.

- ostateczną godziną w ostatnim dniu każdej rundy do której można wrzucać wyniki jest godzina 23:59 GMT. Jeśli wystąpiłyby problemy ze stroną MARP recinputy można wrzucać na forum. Nie należy wrzucać wyników na forum jeśli strona MARP jest dostępna.

- nie ma ograniczenia w ilości wrzucanych wyników, ważne aby każdy następny był lepszy od poprzedniego.

- średnia prędkość nagrania nie może być mniejsza niż 95%.

- jeżeli żaden z trzech koordynatorów turnieju nie może odtworzyć twojego recinputa twój wynik nie będzie zaakceptowany. Należy się upewnić przed rozpoczęciem nagrania aby wykasować/wyłączyć katalog NVRAM, dla gier NEOGEO należy pamiętać o tym również dla katalogu MEMCARD.

RUNDA KWALIFIKACYJNA


Runda kwlifikacyjna trwa od 20 sierpnia 00:00 GMT do 29 sierpnia 23:59 GMT (10 dni). Zakwalifikowanie do turnieju nie będzie trudne, trzeba będzie zrobić naprawdę mały wynik w pierwszej grze w celu zakwalifikowania się do następnej rundy. Oczywiście w zależności od zrobionego wyniku zdobywa się odpowiednią ilość punktów do pozycji w rundach head-to-head (jeden na jeden) (o czym będzie dalej).

RUNDY ELIMINACYJNE

Od tej chwili każda runda będzie trwać 7 dni z 24-godzinną przerwą pomiędzy nimi.

Po rundzie kwalifikacyjne koordynatorzy mają 2 dni czasu na zdecydowanie jak dużo będzie rund eliminacyjnych. Wszystko zależy od liczby zakwalifikowanych graczy. Pierwsza runda eliminacyjna rozpocznie się 1 września.

W rundach eliminacyjnych gracze z dolnych (ostatnich) pozycji będą eliminowani i nie będą przechodzić do następnych rund.

Należy pamiętać że punkty z rund eliminacyjnych również liczą się do pozycji w rundach head-to-head.

Jeśli mniej niż 16 graczy zakwalifikuje się do turnieju nie będzie rund eliminacyjnych, ale raczej jest to wątpliwa sytuacja.

DRABINKA 16 GRACZY - HEAD-TO-HEAD


Po nieobecności w poprzednim turnieju rund jeden na jeden (head-to-head) w tym turnieju będą one ponownie. Będzie to jednak odmienne niż w poprzednich turniejach Knockout. Trzeba będzie PRZEGRAĆ 2 RAZY aby zostać wyeliminownym.

Przykład drabinki 16 graczy
W rundzie tej będą drabinika bedzie podzielona na 2 części, jedna graczy przegranych (Loser's Bracket), druga graczy wygranych (Winner's Bracket). Przegrani w rundach head-to-head przechodzą na drabinkę przegranych, wygrani na drabinkę wygranych. Jeśli na drabince przegranych przegra ktoś ponownie jest wyeliminowany z turnieju.

Po 5 takich rundach zostaną 3-ej gracze. 2-óch z drabinki przegranej i 1 z drabinki wygranej. Więc 6 runda będzie pomiędzy dwoma graczami z drabinki przegranych a gracz z drabinki wygranych ma 7 dni na odpoczynek.

FINAŁ

Zamiast jednej gry w finale będą 2 gry przez 10 dni. Jeśli gracz z drabinki wygranych wygra w obydwu lub jednej z dwóch gier zostaje zwycięzcą turnieju. Aby zwycięzcą turnieju został gracz z drabinki przegranych musi wygrać w OBYDWU grach.
Jest to zgodne z zasadą 2 przegranych w rozgrywkach head-to-head aby zostać wyeliminowanym.

POZYCJE W RUNDACH HEAD-TO-HEAD

Do rund head-to-head gracze będą dobierani w pary według punktów które zdobyli we wcześniejszych rundach.

W zależności od wyników gracze w rundach otrzymują:

za 1 miejsce - 0 punktów
2 - 1 punkt
3 - 2 punkty
4 - 3 punkty
...

Gracz z najmniejszą ilością punktów będzie ustawiony na 1 pozycji, następny na 2, .... Gracze z mniejszą ilością punktów będą parowani z graczami z większą ilością punktów.


Turniej jest dla rozrywki, nie ma żadnych nagród. Jest to turniej z najciekawszymi zasadami. No ale kto wie może w niektórych rundach znajdą się jakieś nagrody jak to czasami bywa.
Last edited by destructor on Wed Aug 30, 2006 6:06 am, edited 1 time in total.
User avatar
TRB_MetroidTeam
MARPaholic
MARPaholic
Posts: 1055
Joined: Fri Oct 22, 2004 4:50 am
Location: PR - Brazil

Post by TRB_MetroidTeam »

destructor wrote:Because rules are long and somebody don't understand them, I prepared rules in Polish language.
Thank you Destructor. Now the mother of my grandmother can participate on K5! :D
Valter "TRB" - The Real Battle
Metroid Team, from Brazil
Crazy Taxi and Crazy Taxi 2 world champions
User avatar
destructor
MARPaholic
MARPaholic
Posts: 1972
Joined: Fri Sep 20, 2002 9:38 am
Location: Poland

Post by destructor »

It's not for any grandmother or mother or mother grandmother :D
It's for Polish players, ex. kuba, DieseL, asn007.
Koza have no problem with english.
User avatar
The TJT
MARPaholic
MARPaholic
Posts: 2479
Joined: Wed Mar 06, 2002 10:56 am
Location: 20 Grand Palace

careless calculus and whispers

Post by The TJT »

zhorik wrote:Even if you have the scenario where the player from the winners bracket only plays one game in the final, then if you think the other player is doing that and you are from losers bracket, pick only one of the games to play yourself and play the other one for a half day. If you guess right (50%) and can beat the player at the game you pick (50%), you have about a 25% chance of winning! If instead you use the rule that the losers bracket champion must beat the winners bracket champion twice, this is also 50% for the first match and 50% for the second match: again 25%.
The odds for "loser" winning the "right" finals-format is correct, 25%.

Odds(*) for "loser" winning the "wrong" finals-format are not correct though:
Firstly, if the guess would be right, then the chances of winning both two games would be 0,25 (not 0,5). This is similar to situation where you flip two coins trying to get heads on both; It doesn't matter whether you throw them separately("right" finals format) or at same time("wrong" finals format), the odds are still 0,25.

Second, when we count in that "winner" can choose which game he plays more, "winner" has more options to choose from than those two you mentioned (95%gameA, 5%gameB and vice versa) ...Winners bracket player("winner") can also play these games with a mix of 50-50, 60-40, 40-60 etc.

Let's say one day of play(10% of total playtime) makes a difference if a game is won or lost:
If there's less than days difference in play then chances of winning a game is equal. However, if there's more than one days play difference -"loser" will win one game while losing the other! ...Thus losing the final.
For example: if "loser" would play for half a day gameB, so, "loser" plays:
gameA = 95%, gameB = 5% (half a day= 5% of 10 days).
"Loser" could only win BOTH games if "winner" would play gameA =100-0% and gameB 0-15%...So loser has 15% chance of matching the guess...If hes guess does not match, he will lose one game for playing it one day less than "winner".
After the guess is correct, loser has 25% chance of actually winning both games...So total chance of "loser" winning would be in this example 0,15x0,25=0,0375 =3,75% !!!

Another example: let's say both players choose randomly between 100-0% how much he will play one game and if 2 days more playtime(20%) makes a difference, which certainly does! - Odds for "loser" to guess "winners" play distribution near enough, would be 36%(**), which would make total chances of winning the final 0,36x0,25 = 9%...wow!


(*)This obviously didn't count other variables that can skew the odds greatly, such as: If one just plays more, has more talent or having luck with game selection in the final.

(**)If winner chose gameA play% from 80-100%, 0-20%...loser has average 70% chance of guessing wrong
If winner's random choice would be from 80-20%...loser has 60% chance of guessing wrong.
Chance for "loser" to randomly guess right = 1-[(0,4x0,7) + (0,6x0,6)] =0,36
User avatar
Zhorik
MARP Seer
MARP Seer
Posts: 788
Joined: Mon Sep 27, 2004 8:16 pm

Re: careless calculus and whispers

Post by Zhorik »

The TJT wrote:
zhorik wrote:Even if you have the scenario where the player from the winners bracket only plays one game in the final, then if you think the other player is doing that and you are from losers bracket, pick only one of the games to play yourself and play the other one for a half day. If you guess right (50%) and can beat the player at the game you pick (50%), you have about a 25% chance of winning! If instead you use the rule that the losers bracket champion must beat the winners bracket champion twice, this is also 50% for the first match and 50% for the second match: again 25%.
The odds for "loser" winning the "right" finals-format is correct, 25%.

Odds(*) for "loser" winning the "wrong" finals-format are not correct though
It still looks correct to me. Notice I said that the player from the winners bracket only played one game in the final. That means the losers bracket player is 100% to win that game, even with just a half day of play, since the other player only played the other game. Per your logic, this is only a 5% loss of time for the player from the losers bracket and shouldn't have much effect on whether he wins the other game, so that game is 50/50. His chance of picking the correct game is also 50/50, so the total percent is 25%.

Of course as you correctly pointed out later in your post, the problem is guessing how the player from the winners bracket is going to allocate his time between the games. My point was that the argument "the winner's bracket player only has to play one game and will win" isn't true. Putting all of his eggs in one basket seems risky. If the winners bracket player goes for a more conservative time allocation of 50-50 or 60-40, there is a good chance that the losers bracket player will be doing something similar, in which case the chance for the losers bracket player isn't that far from 25% as well.
The TJT wrote: (*)This obviously didn't count other variables that can skew the odds greatly, such as: If one just plays more, has more talent or having luck with game selection in the final.
I agree that these variables will have a HUGE impact on the true probabilities and largely make this sort of mathematical analysis moot.
User avatar
TRB_MetroidTeam
MARPaholic
MARPaholic
Posts: 1055
Joined: Fri Oct 22, 2004 4:50 am
Location: PR - Brazil

Post by TRB_MetroidTeam »

Hmmm... if without changes in the method... I suggest the final round (with 2 simultaneous games) would be 2 weeks, instead of 10 days!
Valter "TRB" - The Real Battle
Metroid Team, from Brazil
Crazy Taxi and Crazy Taxi 2 world champions
Post Reply