Home » Forum Home » Tournaments » Tournaments - General

Topic: Pente Masters 2020: Solving the seed problem
Replies: 48   Views: 77,950   Pages: 4   Last Post: Jul 30, 2020, 8:53 PM by: watsu

Search Forum

Back to Topic List Topics: [ Previous | Next ]
Replies: 48   Views: 77,950   Pages: 4   [ Previous | 1 2 3 4 | Next ]
karlw

Posts: 968
Registered: Mar 7, 2006
From: Eugene, Oregon
Age: 32
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 18, 2020, 4:58 AM

@zoeyk: Acknowledging that, this whole time I've been trying to offer solutions that require minimal effort to implement and are also sure to be the least offensive to the people here who care, since these rules (avoiding repeated matchups and seeding according to multiple factors) are almost universally applied for major tournaments in other areas. I'm putting a lot of time and effort into this, and so far the response has been "I don't have the time/inclination to deal with this minor but consequential issue."

karlw

Posts: 968
Registered: Mar 7, 2006
From: Eugene, Oregon
Age: 32
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 18, 2020, 9:03 AM

And watsu, that seeding you proposed is a joke. Not that this has any real bearing on the tournament right now, but based on the fact that seeding is supposed to be a reflection primarily of players' most recent tournament-level performance, an honest-to-goodness accurate seeding would look something like this:

1 pente_gon
2 karlw
3 spavacz
4 be_water
5 watsu
6 madhippo
7 zoeyk
8 richardiii

Unfortunately this is more of a reflection of who has been active recently rather than who the real top players are, but seeding should be empirical rather than hypothetical.

Notice this is much closer to the current seeding, and would indeed be the seeding or close to it if you, zoey and richardiii had played at least a dozen rated sets against top players in the months leading up to this tournament. If only...

watsu

Posts: 1,442
Registered: Dec 16, 2001
Home page
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 18, 2020, 1:16 PM

Sure, it's a joke, but it's more accurate than what we got and it's the only way we have an historical established tb rating from all of us from a single site. Do you honestly believe that if Zk and rich had been active over the past decade they'd be at 7 and 8 and you'd be at 2? I certainly don't believe I'd be 5 if all of us had been active in Pente tournaments on this site in the past few years. I lost 3/4 TB to nosovs when I was perhaps at my best; I believe Rich drew about 4 straight sets against him in a tournament.


Message was edited by: watsu at Jul 18, 2020, 1:23 PM


Message was edited by: watsu at Jul 18, 2020, 1:28 PM


Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
watsu

Posts: 1,442
Registered: Dec 16, 2001
Home page
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 18, 2020, 2:45 PM

Brainking has always (to my knowledge, at least) given points to the lower rated player for splitting a set with the higher rated player in TB. There'd be no rating inflation correction needed to account for a decade of that not being the case here. Arguably at least, our ratings there are closer to accurate reflections of our abilities than they are here, despite the fact that Rich and Zk haven't played there recently.
ETA: I just looked up stats on your TB results against Nosovs at BK in comparison to Rich's
Rich - 9/18 with one P1 loss and 1 P2 wn.
You - 2/6 with one P1 loss


Message was edited by: watsu at Jul 18, 2020, 3:11 PM


Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
karlw

Posts: 968
Registered: Mar 7, 2006
From: Eugene, Oregon
Age: 32
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 18, 2020, 7:33 PM

You're missing the point entirely. You must have skipped or ignored the part where I said seeding is supposed to primarily reflect recent performance (as opposed to yearly rankings, which take into consideration the entire year as well as legacy). This is used by tournament organizers as an incentive to force top players to continually play to ensure a high seed (something which was needed here, hint hint). Your inactivity is the only reason that you would be 5, zoey 7 and rich 8 in my seeding. A true ranking would have all three of you higher, but that is more subjective and I'm not going to try to do that here because that is even more off-topic.

I want to repeat, however, that I do NOT think we should reseed the entire tournament; that would be too dramatic. I've already stated above my recommend measures.

watsu

Posts: 1,442
Registered: Dec 16, 2001
Home page
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 18, 2020, 7:43 PM

No, I actually neither skipped nor ignored the recent performance part. How did you get TB seedings for pgon, spavacz and be_water by that method? Probable answer - from legacy TB ratings (over a year old) which were primarily gained before draws penalized the higher rated player on this site.


Message was edited by: watsu at Jul 18, 2020, 7:46 PM


Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
karlw

Posts: 968
Registered: Mar 7, 2006
From: Eugene, Oregon
Age: 32
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 18, 2020, 7:50 PM

100% incorrect. You don't give me enough credit.

Because top level pente has been nearly non-existent for the last 5 years, you have to be a little more generous when defining "tournament-level" pente and "recent." In brief:

pente_gon: Won 12/20 games in May '19 10-set duel with karlw, won pente masters 2018, had highest active rating on pente.org in 2018.

karlw: Won 8/20 in above duel, won pente masters 2019, highest active rating in 2019.

spavacz: Won 16th anniversary world championship, placed 1st or 2nd in all three(!) of Brf's most recent tournaments.

be_water: Won most recent Brf tournament

watsu, zoeyk, richardiii, madhippo: no major tournament wins in the past 5 years to my knowledge

On second thought, I might swap you and be_water, but really other than pente_gon, madhippo and me, no one on this list has been active enough in the last 4 years to be seeded accurately, which is a shame.

Edit: madhippo did win the Winter's End 2020 tournament.


Message was edited by: karlw at Jul 18, 2020, 7:59 PM


watsu

Posts: 1,442
Registered: Dec 16, 2001
Home page
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 18, 2020, 8:06 PM

Ten set duel was sort of experimental due to the structure imposed, but okay. 16th anniversary tourney was a live tourney. Madhippo went undefeated in the most recent TB tourney. Okay, so I didn't give you enough credit. I doubt I should be 5th, in any case, since the last TB tourney I competed in was a decade ago.
ETA: I was pretty active in 2017, but not in vanilla TB Pente tourneys, so I wouldn't credit my TB rating as far as you apparently do to rate me at 5th.


Message was edited by: watsu at Jul 18, 2020, 8:08 PM


Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
karlw

Posts: 968
Registered: Mar 7, 2006
From: Eugene, Oregon
Age: 32
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 18, 2020, 8:10 PM

I don't think there's much to argue at this point since we mostly agree, and this is a moot point anyway, but in the absolute dearth of major TB tourneys since 2015, a Tournament organizer attempting to seed would have to go to the next closest recent events or the next most recent events, which would be, in rough order,

-minor TB pente tourneys
-head-to-head grandmaster TB matches
-major live pente tourneys
-non-tournament rated TB sets
-major TB tourneys before 2018

Anyway, this whole discussion only serves to reinforce my claim that creating accurate seeds for this tournament would be a nightmare and only serve to make everyone angry. This is why my two proposed solutions do not attempt to do that (other than reseed one player who had a very incorrect provisional rating from 1 set).

watsu

Posts: 1,442
Registered: Dec 16, 2001
Home page
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 18, 2020, 9:01 PM

Anyway, if I were in charge of seeding this nightmare I'd put myself below madhippo due to his recent tourney win and active play. Which the legacy BK ratings reflect accurately. IDK when the last tourney Rich or ZK played in was, but they are at least as current on the TB tournament score as I am, if not more current and their (more accurate than pente.org) legacy ratings from BK are higher than mine, so I'd probably seed them higher than me as well.
But, it's a seeding or reseeding nightmare, no question.
ETA: both Rich and Zk participated in St. Patty's day tourneys within the last 10 years, with Rich being co champ in one and Zk finishing second in another. I'd seed myself 8th here for sure.


Message was edited by: watsu at Jul 18, 2020, 9:16 PM


Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
zoeyk

Posts: 2,220
Registered: Mar 4, 2007
From: San Francisco
Age: 45
Home page
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 18, 2020, 11:22 PM

So, watsu seeds me 3rd, karl seeds me 7th. I'm fine with being seeded 5th as a compromise. Lol

Scire hostis animum - Intelligere ludum - Nosce te ipsum - Prima moventur conciliat - Nolite errare
zoeyk

Posts: 2,220
Registered: Mar 4, 2007
From: San Francisco
Age: 45
Home page
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 18, 2020, 11:52 PM

all right, since we are throwing lists around here, here's mine. Normally R3 would be in the 1 spot. but hes been MIA, and we also don't know how much he is even willing to put into deep study at this point. spavacz is a badass, but he hasn't been studying i don't think. be_water maybe not studying prior to recently too? Karl and Pgon have been here studying the most, so we expect more out of their P2 lines. My history in tourneys is hit n miss. I won one, came in 2nd in another, and got slaughtered by nosovs in some others. and I have been MIA. so, im kind of middle of the road currently possibly. Watsu, I don't know in turn based. I know in forums hes great at analyzing. even finds stuff i miss, but on a live table, i generally beat watsu. so hard for me to quantify his placement. Hippo is a new up and coming star. sharp and quick study, but perhaps some green on the horns, with room to grow, and yet to claim him street cred which would elevate him into the upper top ranks. now if we were doing a live tourney, the seeding here might look much different.

p.s., i was very close to flip flopping me n spavacz here

1 pente_gon
2 karlw
3 be_water
4 richardiii
5 spavacz
6 zoeyk
7 watsu
8 madhippo

Scire hostis animum - Intelligere ludum - Nosce te ipsum - Prima moventur conciliat - Nolite errare
watsu

Posts: 1,442
Registered: Dec 16, 2001
Home page
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 18, 2020, 11:58 PM

Okay, karl can have #2, zk can take his #4 seed. Then Rich to Zk's #10 spot? Everyone happy now? No toes stepped on?

Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
zoeyk

Posts: 2,220
Registered: Mar 4, 2007
From: San Francisco
Age: 45
Home page
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 19, 2020, 12:04 AM

ok let me put this another way guys. observe:
1 pgon
2 karl


the four in middle, shuffle them at random because it don't matter. the middle four are all top notch players who are rusty with no study time put in recently.

then...

7 watsu
8 hippo

Scire hostis animum - Intelligere ludum - Nosce te ipsum - Prima moventur conciliat - Nolite errare
watsu

Posts: 1,442
Registered: Dec 16, 2001
Home page
Re: Pente Masters 2020: Solving the seed problem
Posted: Jul 19, 2020, 12:18 AM

If we were in charge of seeding, I'd agree. But, the 7th and 8th seeds as the tourney stands might not. Hence, my suggestion of reliquishing #2 so that Karl, you and Rich all move up into the empty seats. No one loses their current seed.

Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
Replies: 48   Views: 77,950   Pages: 4   [ Previous | 1 2 3 4 | Next ]
Back to Topic List
Topics: [ Previous | Next ]


Powered by Jive Software