Improvement to Strength of Schedule

Discuss teams, ride/hotel sharing, trash talk, and event results here

Moderators: lunchmoney, TFF Mods

User avatar
rolo
Super Star
Super Star
Posts: 1188
Joined: Wed May 27, 2015 9:38 am
Location: Paradise Stadium, where the pitch is green and the cheerleaders are pretty.

Re: Improvement to Strength of Schedule

Post by rolo »

Re-railing this thread, we're talking about Strength of Schedule as a tiebreaker, not to seed round 1, right?

I think it's a general principle that the only games which matter for placing in a tournament, are those played at that tournament. Games from other tournaments shouldn't matter, so I'd strongly disagree with using NAF ranking as any kind of tiebreaker.

Reason: ''
"It's 2+ and I have a reroll. Chill out. I've got this!"
Image
User avatar
mubo
Star Player
Star Player
Posts: 749
Joined: Mon Dec 22, 2008 7:12 pm
Location: Oxford, UK

Re: Improvement to Strength of Schedule

Post by mubo »

Pairing "top" coaches early is an accepted way of getting a reliable winner if the number of coaches is too large for a successful Swiss. Note though, that it is necessarily harder to win for someone not in the top group.
https://en.wikipedia.org/wiki/McMahon_system_tournament

I don't really see the value in doing it in a pure Swiss system though. It effectively moves the most random round from R1 to R2.

Reason: ''
Glicko guy.
Team England committee member
User avatar
RoterSternHochdahl
Emerging Star
Emerging Star
Posts: 325
Joined: Mon Oct 15, 2012 1:04 pm
Location: Düsseldorf
Contact:

Re: Improvement to Strength of Schedule

Post by RoterSternHochdahl »

one could call the first round draw the premature coin-toss-tiebreaker

Reason: ''
"Chess is two stoic soviet sleeper agents silently conducting 300 possibility calculations per second. Blood bowl is a game where a halfling makes a shepherds pie so you lose all your re rolls." (Thanks to nonumber)
Nippy Longskar
Star Player
Star Player
Posts: 526
Joined: Mon Jan 15, 2007 11:20 pm
Location: Manchester

Re: Improvement to Strength of Schedule

Post by Nippy Longskar »

rolo wrote:
I think it's a general principle that the only games which matter for placing in a tournament, are those played at that tournament. Games from other tournaments shouldn't matter, so I'd strongly disagree with using NAF ranking as any kind of tiebreaker.
I guess my question would be why is that a general principle? I get the concept that tournaments should be self-enclosed bubbles, but if this throws up issues with how to divide those with the same record in the tournament then why not make use of the available data to help?

Also just to clarify my suggestion was SoS based on oppo's NAF W/L ratio, rather than ranking.

Reason: ''
That's huge
User avatar
Purplegoo
Legend
Legend
Posts: 2256
Joined: Mon Jan 14, 2008 1:13 pm
Location: Cambridge

Re: Improvement to Strength of Schedule

Post by Purplegoo »

I guess the idea of SoS is to say ‘you two have finished on the same number of tournament points. The fairest way to decide ‘who has done best’ is to look at who you have played and award it to the guy that’s had the harder run’. I think it ever so slightly fairer to decide how hard the run was contained within the single tournament bubble. A victory over Joe’s Undead on table one in one tournament should probably count more than a victory over Joe’s Ogres randomly drawn in game one, or on table 20 in another. It might be that winning the lottery of killing Joe’s Snotlings is the key to the best SoS in that, theoretical win % tournament? I don’t think it’s a bad idea, I just think it ever so slightly worse than the current SoS standard.

I do think SoS is the best system we have, although I appreciate the argument of feeling out of control. It could be argued that you are not in control of TDs or (more certainly) CAS either, but I digress. An idea for an ‘in control’ TB that we’ve started to use in Online NAF tournaments is racial tiers. I cannot give (for instance) Humans more skills than Undead with the current FUMBBL infrastructure, but I can have Humans ‘win’ the tournament if they finish on the same record as Undead. In tournaments with no tiered skill packages, this is simple to set up (we all know what T1, 2, 3 are, I appreciate Score! does not). Tournaments with tiered skill packages are theoretically designed to give those Humans a leg up compared to Undead, but not make them better. In practice, we know that tiered rulesets don’t do this, they simply identify the same number of competitive teams, just from elsewhere in the tier structure. I don’t see that as an issue with using tiers as TBs, though, if a coach has identified High Elves as the best race in the rules and selects them, well done him, he deserves the TB?

So what about racial tier / SoS / net TD + CAS / guess the weight of the cake? You are absolutely in control of the race you select, after all?

Reason: ''
Wulfyn
Emerging Star
Emerging Star
Posts: 323
Joined: Mon May 19, 2014 9:33 pm

Re: Improvement to Strength of Schedule

Post by Wulfyn »

mubo wrote:Pairing "top" coaches early is an accepted way of getting a reliable winner if the number of coaches is too large for a successful Swiss. Note though, that it is necessarily harder to win for someone not in the top group.
https://en.wikipedia.org/wiki/McMahon_system_tournament

I don't really see the value in doing it in a pure Swiss system though. It effectively moves the most random round from R1 to R2.
I'm a big fan of the McMahon system, and I think that it is the accepted default in Go is a good testament to it's reliability. One of the big problems I think in SoS and TD dif is the randomness of the first draw in Swiss. Whilst Joe is correct that he might have a harder game than Geggster if the two were to finish on equal points, if they both played people much weaker than them then how much is that a separator? The big weakness of the Swiss system is the randomness of the first draw.

The McMahon system divides the competitors into 2 halves and has a false first round where no games are played but the top half is given a bonus point for winning (or 2 in a 2/1/0 system). Effectively it is decided that the top half of the players all get a free and equal win. So why is that fair?

Well in the first round proper it now means that the top half play each other. Nobody is getting a super easy first round (or a super hard one). That means that nobody gets a big advantage or disadvantage. It splits the field very quickly. So the players in the top half who lose will be caught by the players in the bottom half who win. That gets the midfield in place much faster as well. Overall the ranking tends to find people of similar levels really really quickly, as opposed to Swiss where in a large tournament 2 equally skilled players could end up 3 games in having played a series of weaker opponents (who have won through playing weaker opponents).

Top 25% = 1 bonus win and easier first game (but harder than it could be), but face each other quickly.
Q2 25% = 1 bonus win but harder first game as they play other top half players.
Q3 25% = 0 bonus win but easier first game as they play other bottom half players.
Bottom = 0 bonus win and still a hard first game (but easier than it could be), but face each other quickly.

As players improve they will work their way into the top half of the draw for future tournaments.


edit: It not only helps produce a more reliable winner, but also helps rank coaches closer to their ability throughout the entire field.

Reason: ''
User avatar
sann0638
Kommissar Enthusiasmoff
Posts: 6609
Joined: Mon May 08, 2006 10:24 am
Location: Swindon, England

Re: Improvement to Strength of Schedule

Post by sann0638 »

Quick McMahon question - do you remove the "extra win" at the end? Clicked on the link, but wasn't clear.

Reason: ''
NAF Ex-President
Founder of SAWBBL, Swindon and Wiltshire's BB League - find us on Facebook and Discord
NAF Data wrangler
Greshvakk
Experienced
Experienced
Posts: 82
Joined: Wed Mar 04, 2015 12:12 am

Re: Improvement to Strength of Schedule

Post by Greshvakk »

sann0638 wrote:Quick McMahon question - do you remove the "extra win" at the end? Clicked on the link, but wasn't clear.
I don't believe so you are effectively only giving half the field (the top seeds) the chance to win the tournament - that's their reward for having guaranteed harder games if they keep winning. I believe it is possible for someone in the bottom half of the draw to win if the top draw with each other a lot and therefore someone winning from the bottom half 'catches up' with the top top player. I think that's how it works but Wulfyn or Mubo may correct me.

Reason: ''
Image
Wulfyn
Emerging Star
Emerging Star
Posts: 323
Joined: Mon May 19, 2014 9:33 pm

Re: Improvement to Strength of Schedule

Post by Wulfyn »

No, you keep the extra win. That is the benefit to having a forced harder game at the start (else someone low ranked could end up winning whilst avoiding all the best players).

It makes Q4 players less likely to win overall, but then they are far less likely to win anyway. Q2 and Q3 is worth between 0 and 1 wins because of the paired draws (Q2 get a win bonus but are less likely to win their opening game, and Q3 don't get a win bonus but are more likely to win their opening game).

Reason: ''
User avatar
sann0638
Kommissar Enthusiasmoff
Posts: 6609
Joined: Mon May 08, 2006 10:24 am
Location: Swindon, England

Re: Improvement to Strength of Schedule

Post by sann0638 »

Sounds interesting, certainly. Would take someone brave to implement it at their tournament I think, as it could well impact attendance (imo).

Reason: ''
NAF Ex-President
Founder of SAWBBL, Swindon and Wiltshire's BB League - find us on Facebook and Discord
NAF Data wrangler
Wulfyn
Emerging Star
Emerging Star
Posts: 323
Joined: Mon May 19, 2014 9:33 pm

Re: Improvement to Strength of Schedule

Post by Wulfyn »

I'm not sure why it would. You often hear about a lot of people who only seek to attend an event because of the great atmosphere and the ability to nerd out for a weekend. Many people go with few illusions that they will win it. Some people even say that those who go to win are doing it wrong. All of those people should not be fussed by this system at all.

Many reports of the McMahon say that it increases people's enjoyment of the tournament, regardless of their skill level. There is far less chance for a casual player to get boned in the first day by meeting 3 heavyweights. They are more likely to play like-minded players of a similar ability level for more of the tournament, producing more interesting and closer games where the skill gap does not mean that only big dice can change the result.

Reason: ''
User avatar
Dionysian
Emerging Star
Emerging Star
Posts: 360
Joined: Thu Jan 31, 2013 4:01 pm

Re: Improvement to Strength of Schedule

Post by Dionysian »

@Wulfyn

While McMahon is great, not only in terms of enhancing the competitive nature of the tournament circuit and accuracy of NAF rankings, but also increasing the enjoyment of every attendee - the latter benefits won't be immediately obvious to many.

I agree that were it to ever be trialled in any significant manner it would likely catch on pretty quickly and become popular, but I'm fairly sure that, at least in the UK, there isn't currently the cultural acceptance (nor probably the competitive infrastructure) to conduct that trial.

Perhaps in a country where BB is more widely considered a game of skill and tournament players more generally strive to improve, rather than simply attribute their failures to poor luck.

Reason: ''
Oventa
Veteran
Veteran
Posts: 246
Joined: Tue Apr 01, 2008 7:58 am
Location: Munich, Germany

Re: Improvement to Strength of Schedule

Post by Oventa »

Does anybody know how many rounds are usually played in a go tournament?
I mean in bloodbowl due to weekend and time it is usually 6 rounds, independent of attendance (12 as well as 150)
In a go tournament is there also a fixed number or is this tuned to attendance?

Reason: ''
Image
User avatar
Purplegoo
Legend
Legend
Posts: 2256
Joined: Mon Jan 14, 2008 1:13 pm
Location: Cambridge

Re: Improvement to Strength of Schedule

Post by Purplegoo »

I do like that system, but I think I'd only try applying it to BB in some cases. I'm not sure I'd like to see it every week at 20-30 man tournaments (the pool of players you'd meet in the year might reduce even further, and is it really necessary when the field is small?). However, I do think it would help improve the big ones. It's often said that the largest tournaments are crap shoots, it would not seem unreasonable to try and reduce that, perhaps with this method?

I think you'd run into a number of cultural issues as Dionysian points out and others infer. I don't think it would be popular.

The more Go comes up in the news, here, there and everywhere, the more interested I am in trying it. Is there a free, simple tool to learn with someone could recommend (sorry for the aside)?

Reason: ''
User avatar
sann0638
Kommissar Enthusiasmoff
Posts: 6609
Joined: Mon May 08, 2006 10:24 am
Location: Swindon, England

Re: Improvement to Strength of Schedule

Post by sann0638 »

Just went through the tutorial here, it's super easy: https://online-go.com/

British Go Open seems to be 6 rounds over a weekend.

Reason: ''
NAF Ex-President
Founder of SAWBBL, Swindon and Wiltshire's BB League - find us on Facebook and Discord
NAF Data wrangler
Post Reply