The World Cup Discussion Thread (OOC, Version IV)

[95X]

Just wanted to say I had an idea as the last two days I've faced opponents who have no roster nor have RP'd.

After posting RPs about the match I've sent them a nice TG letting them know my team played theirs in the World Cup Qualifiers, looked up what date they signed up for the tournament, and included a link to both the post where the score was posted as well as the link to my match report RP. (Granted, I haven't written lengthy RPs this tournament.)

My idea is to let other users know that yes, what they signed up for in some cases months ago really is taking place and that they still have time to post a roster or RP, and to wish them luck in the qualifiers as any good sportsperson would do.

[Zwangzug]

Host announcement

Yesterday, when inputting updated ranks into the scorinator file, I made an error transferring from our spreadsheet to xkoranate. This caused Cassadaigua's rank for MD3 to be much lower than it should have been. Fortuitously, they defeated Olastor anyway, and there will be no further action to remedy this--their rank was correctly updated for tonight. Just posting to be accountable and let you all know.

[Farfadillis]

In a similar vein, I missed South Covello's first RP, which thankfully happened before a bye so there were no consequences, and Turori's roster yesterday, before a match Turori thankfully won.

[Zwangzug]

Host announcement

Yesterday, when inputting updated ranks into the scorinator file, I made an error transferring from our spreadsheet to xkoranate. This caused Cassadaigua's rank for MD3 to be much lower than it should have been. Fortuitously, they defeated Olastor anyway, and there will be no further action to remedy this--their rank was correctly updated for tonight. Just posting to be accountable and let you all know.

To elaborate briefly on this. My judgment in this particular case isn't intended to set a precedent or justify what I might have done in similar but different scenarios. The primary reason I, with my cohosts' blessing, went forward on this is pragmatic; I believe that if I had to scorinate group 9 separately and manually make tables for them, there would be a significant chance of me making further errors, which would do more damage than the errant scorination already did. I also think that, even if the ranks entered had been correct, the high seeds defeating Olastor by two goals at home is a not-unlikely result. (To be very very clear, this is a subjective and non-mathematical judgment, but it gives further weight to allowing the result to stand.) In a perfect world, hosts would be flawless and make no copy-paste errors, and so we would not need to have these discussions, but this is not a perfect world, and I thank you for your patience now and as we go forward.

[Audioslavia]

I think you made the right decision.

Furthermore (and this is a note to everyone, not just to the current hosts): If you're co-hosting the World Cup, you are in charge. Go with what you think is the best option. The community is entitled to their say, but if it's in the middle of a tournament, make your decision (as Zwang and the boys have done) and stick to it.

Good job.

[Cassadaigua]

I've already told this to Zwangzug, but I will mention it here as well, as the impacted party. I appreciate their transparency and support their decision.

[Squornshelan Remnant States]

A timely reminder that all hosts are (most likely) human, all hosts make errors, and that they do their best to make the right decisions about how to proceed once a mistake has been made. We can also remember that they are also all volunteers.

[Sarzonia]

I agree with Squorn, Cass, and Audio. More than the decision itself, which I support, I like the transparency from Zwangzug and Farfadillis.

[Independent Athletes from Quebec]

Good decision made by both of you. Thank you!

[Sarzonia]

Just out of curiosity, do the hosts have my style modifier set to zero?

[Farfadillis]

Just out of curiosity, do the hosts have my style modifier set to zero?

Just checked: yes.

[Sylestone]

how many teams qualify for the WC?
Thanks

[Delaclava]

how many teams qualify for the WC?
Thanks

30 from the qualifiers, plus the two hosts automatically. In this case, it's the top two teams from each of the 15 groups.

[Sylestone]

how many teams qualify for the WC?
Thanks

30 from the qualifiers, plus the two hosts automatically. In this case, it's the top two teams from each of the 15 groups.

Thank you!

[NSWC Signups]

We have one bid so far to host Cup of Harmony 77- Zwangzug & Kelssek.

The deadline for anyone to bid to host Cup of Harmony 77 will be 11:59 PM EDT on June 4th, 2020.

[Farfadillis]

After an argument that arose on Discord over what to do when the host has made a mistake and input the wrong ranking for one of the teams, who then happened to win, I promised I'd write an at length explanation of the math behind the decision to rescorinate and why, leaving aside practical problems that could arise from rescorinating such as having to create tables manually, it is in a sense actually the correct approach, despite how counterintuitive it is.

I will use the Muradil-Turori match I screwed up as an example. Turori's chances of winning that match without their roster bonus were roughly ~84%, and their chances with a roster bonus were about 87% (ignoring home advantage, which is besides the point of this post).

Let's assume I've made a mistake and caught it quickly. We will now figure out Turori's chances of winning the match, and we'll do so by analyzing every possible case:

In 84% of cases, Turori wins; in 16%, they do not. Let's say I see Turori won and decide not to rescorinate. That means Turori actually had a chance of at least ~84% of winning. Let's further assume that I feel bad that Turori lost, and I feel a rescorination is the fairest course of action.

We are now in the ~16% of cases in which Turori did not win. Of these 16% of cases, Turori will win a further ~87% (now with roster bonus), which comes out to 0.87*0.16=0.1392, so Turori actually wins in a further ~14% of cases.

Turori's chances of winning have therefore become ~84%+~14%=~98%. This is a considerably higher chance of winning than if I had remembered to award them their roster bonus.

The crux of the matter here is that we are essentially giving Turori two chances to win, not just one.

All of this said, this is actually just one way to look at the problem. This is, if I'm not mistaken, what is called a Bayesian approach. In probability, there's a philosophical divide between Bayesians and frequentists, and I don't want to go down that rabbithole. You can go read about it if it interests you, but I feel going into detail about this would detract from the point of the post. What's important is that there are two, equally valid approaches to this kind of thing, and I've only talked about one. I'll now explain rescorination from a frequentist perspective.

First off, I suspect most of you are frequentists, even if you don't know it. I know I am, which I've only come to realize after thinking this post through. As a result, I expect most of you to find this second part of the post both more intuitive and more agreeable. Just know that this is actually a philosophical matter, and arriving at "right" answers is basically impossible by design. I'll just limit myself to the math.

From a frequentist perspective, we could argue that if Turori won, then the roster bonus wouldn't have changed that. They already had enough luck to win it. If anything, the roster bonus would've increased the margin by which they won. Similarly, we could argue the game has to be rescorinated if they lost, since the roster bonus could've swayed that result the other way.

I agree with this sentiment - as I said, I'm a frequentist at heart. However, it's easy to go from this to "rescorinate only if Turori did not win", and doing that would be 100% wrong.

To illustrate this, let's imagine Turori lost 2-0 to Muradil. Now, sure, the roster bonus would've helped Turori get a better result, but, as I mentioned, it only really gave him a 3% bigger chance of winning. Turori only lose by more than one goal in a small number of matches; they definitely had bad enough luck that they would've lost even with a roster bonus, assuming luck stays the same (which, remember, is a key assumption for not rescorinating when Turori wins).

The conclusion we arrive at is that not rescorinating is correct, but only if the result has been one-sided enough for the roster bonus to not matter (ignoring that GD can actually be important, at least). Taking this approach we arrive at a problem, however: with how NSFS and SQIS work, we have no way of knowing just how lucky or unlucky Turori were in the first result. It is therefore impossible to know exactly how much the roster bonus would've swayed the result in Turori's favor. We could definitely think it could've, and then unfairly rescorinate a fair and square Muradil win.

I am, as a result, of the opinion that the Bayesian approach is the best at the moment. It guarantees fairness, whereas the frequentist approach falls short because of external limitations. Therefore, rescorinating, even if the RPing team won, is the best approach.

A couple other things I'd like to mention:

1) I have plans to, at some point in the future, create a scorinator that, besides outputting a result, also uses one random seed as input (and outputs it). In other words, a scorinator that tells you exactly how lucky each team was, and one that can generate a result non-randomly if you give it a random seed.

As an example:

Turori has a result that is in the top 5% of lucky results. Let's call this seed 0.95. The scorinator outputs "Muradil 1-5 Turori". You notice the mistake you made. You add the roster bonus. I input the same seed into the scorinator, but with the updated ranks. The scorinator now outputs "Muradil 1-6 Turori". Result fairly corrected.

2) All of this post has one enormous caveat, which I briefly mentioned in my first sentence. There needs to be room left for pragmatism. You might notice I did not rescorinate the very same Turori-Muradil match I am talking about. This is because I caught the result late and rescorinating it could've resulted in a shitstorm at that point, though I don't think Vil would've caused much of a fuss. Another scenario in which rescorination could be imprudent, as Zwang pointed out, is that rescorinating just the one result leads to having to create tables manually which, as we know, only complicates things further and makes future mistakes more likely. Hosts aren't perfect, they aren't supposed to be, and sometimes shit happens. Needless to say, I would've backed Zwang in her decision not to rescorinate even if she'd caught the mistake early.

This caveat does not quite apply to, say, WC knock-out matches, though, where every result matters immensely and keeping things as fair as possible is vital and, furthermore, technical difficulties are unlikely to arise from rescorination.

3) I do not intend to engage in a philosophical debate over Bayesianism vs frequentism. If anyone has any concerns to raise regarding this post, please make sure you've understood all of the math behind it before making your arguments, and make sure they're grounded in mathematics. I would be more than happy to answer any questions any of you may have and/or engage in a mathematical argument (I'm not infallible, maybe my math is wrong here, as unlikely as I think that is). However, I would not be happy to engage in philosophical bickering, which is what I fear would happen if somebody tries to argue against this post without first understanding the math behind it.

I also am aware of the fact that many/most of the people in this community do not have a good grasp of probability theory. I've made my best effort to make this post as informative and approachable as possible, but sometimes some concepts in math just take some time and effort to fully be understood. Many of my uni professors have heavily criticized my inability to explain my mathematical ideas clearly, so apologies in advance.

[Kandorith]

Farfadillis that explanation has finally made me understand scorinators somewhat...

I do not know much about hosting etcc, but I do think it was the most fair decision and the right one.

Which also made me think, holy crap I have rotten luck by constantly drawing and just barely winning from roster and RP-less teams. xD (And too be fair it's starting to get me a little bit demotivated but eh...)

[Jeckland]

Let's be honest, we already know who's qualifying from Group 1 so who cares.

Jokes aside, great explanation Farf. Would never have thought of it that way.

[Kandorith]

Let's be honest, we already know who's qualifying from Group 1 so who cares.

Jokes aside, great explanation Farf. Would never have thought of it that way.

Of course we do, it's going to be TJUN-ia and Astograth. Duh.

[Liventia]

If anyone has any concerns to raise regarding this post, please make sure you've understood all of the math behind it before making your arguments, and make sure they're grounded in mathematics. I would be more than happy to answer any questions any of you may have and/or engage in a mathematical argument (I'm not infallible, maybe my math is wrong here, as unlikely as I think that is). However, I would not be happy to engage in philosophical bickering, which is what I fear would happen if somebody tries to argue against this post without first understanding the math behind it.

I also am aware of the fact that many/most of the people in this community do not have a good grasp of probability theory. I've made my best effort to make this post as informative and approachable as possible, but sometimes some concepts in math just take some time and effort to fully be understood. Many of my uni professors have heavily criticized my inability to explain my mathematical ideas clearly, so apologies in advance.

Whatever the merits of your explanation, "you can only argue with me in terms of a mathematical theory you may not know about" is a disingenuous way of shutting down debate.

[Torisakia]

If anyone has any concerns to raise regarding this post, please make sure you've understood all of the math behind it before making your arguments, and make sure they're grounded in mathematics. I would be more than happy to answer any questions any of you may have and/or engage in a mathematical argument (I'm not infallible, maybe my math is wrong here, as unlikely as I think that is). However, I would not be happy to engage in philosophical bickering, which is what I fear would happen if somebody tries to argue against this post without first understanding the math behind it.

I also am aware of the fact that many/most of the people in this community do not have a good grasp of probability theory. I've made my best effort to make this post as informative and approachable as possible, but sometimes some concepts in math just take some time and effort to fully be understood. Many of my uni professors have heavily criticized my inability to explain my mathematical ideas clearly, so apologies in advance.

Whatever the merits of your explanation, "you can only argue with me in terms of a mathematical theory you may not know about" is a disingenuous way of shutting down debate.

Honestly you can't bitch about scorination if you don't know how it works in the first place. That's like complaining that you can't put your new desk together when you threw away the instructions.

[Liventia]

Honestly you can't bitch about scorination if you don't know how it works in the first place. That's like complaining that you can't put your new desk together when you threw away the instructions.

… Are you seriously suggesting a multiple World Cup and Olympic Games host doesn't know how scorination works?

I was merely pointing out that, as valid as Farf's point may be, telling people they have no right to rebut him unless they are able to speak in extremely technical mathematical terminology is not a great look.

[Torisakia]

Honestly you can't bitch about scorination if you don't know how it works in the first place. That's like complaining that you can't put your new desk together when you threw away the instructions.

… Are you seriously suggesting a multiple World Cup and Olympic Games host doesn't know how scorination works?

I was merely pointing out that, as valid as Farf's point may be, telling people they have no right to rebut him unless they are able to speak in extremely technical mathematical terminology is not a great look.

I was talking more to the people who bitch at the hosts for scorination.

I honestly don't even know what this situation is about. I just wanted to give my unwanted commentary.

[Delaclava]

If anyone has any concerns to raise regarding this post, please make sure you've understood all of the math behind it before making your arguments, and make sure they're grounded in mathematics. I would be more than happy to answer any questions any of you may have and/or engage in a mathematical argument (I'm not infallible, maybe my math is wrong here, as unlikely as I think that is). However, I would not be happy to engage in philosophical bickering, which is what I fear would happen if somebody tries to argue against this post without first understanding the math behind it.

I also am aware of the fact that many/most of the people in this community do not have a good grasp of probability theory. I've made my best effort to make this post as informative and approachable as possible, but sometimes some concepts in math just take some time and effort to fully be understood. Many of my uni professors have heavily criticized my inability to explain my mathematical ideas clearly, so apologies in advance.

Whatever the merits of your explanation, "you can only argue with me in terms of a mathematical theory you may not know about" is a disingenuous way of shutting down debate.

Which is why Farf went ahead with this clear, well-written post in an attempt to explain said theory. This actually opens up constructive discussion, rather than suppressing it. It's not unreasonable to ask that people are at least somewhat thoughtful in their engagement.

[Northwest Kalactin]

… Are you seriously suggesting a multiple World Cup and Olympic Games host doesn't know how scorination works?

I was merely pointing out that, as valid as Farf's point may be, telling people they have no right to rebut him unless they are able to speak in extremely technical mathematical terminology is not a great look.

I was talking more to the people who bitch at the hosts for scorination.

I honestly don't even know what this situation is about. I just wanted to give my unwanted commentary.

I wouldn’t talk in this if you don’t know what the situation is.