World Baseball Classic Discussion Thread

[Tarrentum]

Sorry, more questions: if three or more teams are tied, are head-to-head records cumulative (i.e., amongst tied teams)?

Also, if through the tiebreaking process one or more teams are eliminated, does the process revert back to step one or continue from where it left off? Ex: there are three teams ties with the same record and all the teams are 2-2 against each other by, say, Team A beating Team B twice, Team B beating Team C twice, and Team C beating Team A twice. In the process of determining the top ranked team among them, the tiebreaker goes to total run differential, and Team A and Team B both have +7 while Team C has +2. Team C is eliminated, but what is the next step? Does it go back to head-to-head (where Team A would outrank Team B) or does it continue to total runs scored?

I tried to make sure all bases were covered. Remember the criteria for tiebreakers (assuming Overall Record is the same).
First is Head to Head record. If two teams are tied with no H2H competition - then it goes to Total Run Differential.
If TR Diff. is the same then it goes to Average Runs Scored Per Game (ARS/G). If ARS/G is tied then it goes to a coin flip. Total Runs were included so that people could see where that one stat is coming from (to make ARS/G and TR Diff).

The example you put up leaves out criteria 4 and 5 (ARS/G and Coin Flip), which would happen first (and last becuase a coin flip is pretty final).

[Sakurakoku]

My question has not been answered - the example I put up was for three teams. I was asking whether tiebreaking steps continue until only one team remains or if they revert back to step one when at least one of the teams involved are eliminated (but not necessarily n-1 teams are eliminated, where n is the number of teams involved in the tiebreak).

Also, there is no difference between total runs scored and average runs per game because every team will play the same number of games at every point in time. Unless forfeits are a possibility, both stats are operationally identical - if you win over another team in one stat, you will win with the other stat.

[Secristan]

My question has not been answered - the example I put up was for three teams. I was asking whether tiebreaking steps continue until only one team remains or if they revert back to step one when at least one of the teams involved are eliminated (but not necessarily n-1 teams are eliminated, where n is the number of teams involved in the tiebreak).

He's referring to cumulative H2H amongst tied teams. This is the traditional way we have done it. If three teams end up 8-4, and in total one was 3-1 against the other 2, one 2-2, and the other 1-3, the 3-1 gets the tiebreaker, then the 2-2, then the 1-3. If it was a straight 8-4 with 3 teams, and they were all 2-2 against one another, then it goes to the next criteria.

[Sakurakoku]

So if three 8-4 teams were all 2-2 against each other, we would go to the next tiebreaker - run differential. That I understand.

Suppose it breaks down like this:

Team A: +12 run differential
Team B: +12 run differential
Team C: +6 run differential

What happens here?

1: Team C is eliminated from the tiebreaker, Team A and Team B goes to the next tiebreaker (average runs per game/total runs scored).
2: Team C is eliminated from the tiebreaker, Team A and Team B revert back to step 1 of the tiebreaker (head to head).
3: All three teams move on to the next tiebreaker (average runs per game/total runs scored) because run differential was insufficient to determine one winner.
4: Something else.

[Tarrentum]

So if three 8-4 teams were all 2-2 against each other, we would go to the next tiebreaker - run differential. That I understand.

Suppose it breaks down like this:

Team A: +12 run differential
Team B: +12 run differential
Team C: +6 run differential

What happens here?

1: Team C is eliminated from the tiebreaker, Team A and Team B goes to the next tiebreaker (average runs per game/total runs scored).
2: Team C is eliminated from the tiebreaker, Team A and Team B revert back to step 1 of the tiebreaker (head to head).
3: All three teams move on to the next tiebreaker (average runs per game/total runs scored) because run differential was insufficient to determine one winner.
4: Something else.

While I misunderstood originally the first option would be correct. Team C is no longer part of the tie breaker scenario becuase it is not tied with A and B with a +12 run differential. Therefore, the higher ranked team between A and B would then be determined using ARS/G. If A and B have the exact same ARS/G then it is a coin flip.

I'd also like to point out that the host bid was accepted, albeit with a low number of respondents, with the five criteria listed. These are the five determining criteria for the WBC 16 and will be final. No other criteria will be implemented for tiebreakers.

[Sakurakoku]

Oh, yes, I'm not trying to argue for a different set of tiebreakers (though admittedly the set-up isn't what I'm used to seeing). I just wanted an explanation.

Thanks for putting up with my questions.

[Tarrentum]

To continue the Tiebreaker Discussion:

I felt the need to post this becuase of the standings.
First: We are only four games in to the 14 game Group Stage schedule
Second: For Yuppers FoodBrands Group and MercantileBancshares Group I had to throw out H2H comparison for tiebreaking in the standings for Match Day 4. The issue is that it is impossible to truly determine standings based solely on H2H (using that criteria first as dictated in the Host Bid) before using TR Diff.

Therefore, I am going to leave the standings as sorted by Winning Percentage, then TR Diff., then AR/G for the Standings up to MD4. This will clear itself out as we go deeper into the schedule.

[Secristan]

This has nothing to do with the hosts, who are doing a fine job, or to TBI who created the scorinator.

But there is way too much randomness with this Miscellanator. I didn't RP much in the last Classic, so I didn't speak up. But here, for both us and Sarzonia to be sitting at 2-2, behind 4-0 Freelasio is what is called a red flag. There is something wrong with this scorinator, it does not properly balance the important factors that go into a tournament as established as the WBC. I will not support any future bid that proposes its use unless it is significantly corrected.

If I wasn't enjoying my storyline, I would see no further reason for myself, or anyone, to continue roleplaying, quite frankly.

[Sarzonia]

One match I could easily chalk up to Margaret. Through four matches and having an unranked team that hasn't RP'd once or even posted a roster sit at 4-0 while the two highest-ranked teams whose players have RP'd AND posted rosters are 2-2? Something's not right.

I don't like being in the habit of questioning results because in the end, they don't mean much in the grand scheme of life. But I, too, would be hard pressed to support a future bid for the World Baseball Classic that uses Miscellanator.

[Liventia]

To continue the Tiebreaker Discussion:

I felt the need to post this becuase of the standings.
First: We are only four games in to the 14 game Group Stage schedule
Second: For Yuppers FoodBrands Group and MercantileBancshares Group I had to throw out H2H comparison for tiebreaking in the standings for Match Day 4. The issue is that it is impossible to truly determine standings based solely on H2H (using that criteria first as dictated in the Host Bid) before using TR Diff.

Therefore, I am going to leave the standings as sorted by Winning Percentage, then TR Diff., then AR/G for the Standings up to MD4. This will clear itself out as we go deeper into the schedule.

While I withdrew from signing up, I do think you should not and never
a) deviate from what you promised in your bid without good reason; and
b) take such a decision seemingly unilaterally without agreement from your co-host.

Not looking good, I must say.

[Sakurakoku]

I don't think he means to actually change the tiebreaking procedures - rather, he's ignoring them for now because with only 4 out of 14 games played, odds are the head-to-head records aren't going to be meaningful. As the group stages progress they should fill out.

[Liventia]

If there's been a head-to-head meeting between tied teams, why shouldn't it count regardless of whether four or 14 matches have been played? I fail to see the point here.

[Sarzonia]

In a couple of competitions he hosted, Dancougar would not apply tiebreakers until about the two-thirds or three-quarters point. However, when it came down to the stretch run, he would apply them accurately. I think that's what Tarrentum's doing here. If it's not, that's a whole separate issue that I would consider far more serious than which scorinator is spitting out which random numbers.

Granted, I try to apply head-to-head in the tournaments I've hosted even from Matchday 2 when I first start publishing tables, but I haven't exactly run a very large tournament, at least not for several years.

[Quintessence of Dust]

I think the biggest factor is the home advantage. The Miscellanator help file states it averages out to about 0.5 runs per game.

Here's a list of upsets (a higher ranked team losing, taking no account of RP bonus), with those with the away team winning in italics:

Delaclava @ Grantanata
Sibirsky @ New High Charity 24
Lycabron @ Gracelona
Qazox @ Churchma
Tocrowkia @ Manticore Reborn
Yafor 2 @ Pacific_a
Mantwenic @ Castigonia
Sarzonia @ Michael VII
Allamunic States @ Elemanu
Silver Beach @ Freelasio

Bodegraven @ Sibirsky
Cassadaigua @ Milchama
Sakurakoku @ Virabia
Pacific_a @ Mantwenic
Elemanu @ Silver Beach
Voltronica @ Allamunic States

Delaclava @ Iron Coast
Sibirsky @ Grantanata
Milchama @ Conversants
Western cuba @ Zwangzug
Silver Beach @ Voltronica

Delaclava @ Tarrentum
Milchama @ Lycrabon
Yafor 2 @ Mantwenic
Sarzonia @ Silver Beach
Secristan @ Allamunic States

As you can see, the vast majority of upsets have involved a home team winning. Many teams have, after four days, played a 3-1 schedule (including, for example, Freelassio - who have played 3 home games, and are 4-0 - and Delaclava - who have played 3 road games, and are 1-3). The minority involving a road team winning have almost all been teams with little RP bonus or relatively low ranks anyway.

So I would hope that some of these issues would have resolved by the time we have a full round robin complete. And, of course, the home advantage will no longer be a factor in the playoffs.

I really hope this does not discourage anyone from RPing, however.

[Qazox]

Perhaps we need to go back to the -1/0/+1 home advantage instead of a flat-rate home advantage.

[Sarzonia]

The fact that a match is being played in a particular home stadium should not have as much impact as it seems to be having.

If you told me after four matchdays that Sarzonia would be 2-2 and would be 1-2 away, I wouldn't be surprised. However, if you told me that my away win would be against the team ranked next highest in my group and my losses would be to an unranked team and a team that's still significantly lower ranked than mine, I'd wonder what's up.

One odd match here and there I can understand. But a pattern of oddness at about a quarter of the way in is not trivial.

[Newmanistan]

I think the biggest factor is the home advantage. The Miscellanator help file states it averages out to about 0.5 runs per game.

Here's a list of upsets (a higher ranked team losing, taking no account of RP bonus), with those with the away team winning in italics:



I really hope this does not discourage anyone from RPing, however.

Using your list, that's upsets in 26 of 96 games, or over 25%. I have been fortunate not to be bit by this yet, but I can see what is going on. If in the the World Cup, newbies were beating top 20 teams on a 25% clip, there would be about 20 pages of "the scorinator us broke." I find that percentage way too high. In my opinion when Newmanistan plays a newbie nation like Castigonia, the odds of Newmanistan winning should not the equivalent of the "Yankees beating Baltimore when Baltimore couldn't win to save their life", but as the odds of the United States defeating a country that you wouldn't even think of as having a baseball team such as say, Iraq. If Newmanistan has a less the 75% chance of going into Castigonia and winning, then this system is indeed flawed. We should have well over a 90% chance. This is not a Major League Baseball simulation, it is a World Baseball Classic simulation.

These facts and concerns were developing during WBC15, and now with it happening again, demonstrate a scorinator not ready to be used for the WBC. This is not through a fault of Tarrentum and Quintessence of Dust, who I feel are doing a great job hosting.

[Quintessence of Dust]

If you told me after four matchdays that Sarzonia would be 2-2 and would be 1-2 away, I wouldn't be surprised. However, if you told me that my away win would be against the team ranked next highest in my group and my losses would be to an unranked team and a team that's still significantly lower ranked than mine, I'd wonder what's up.

In my opinion when Newmanistan plays a newbie nation like Castigonia, the odds of Newmanistan winning should not the equivalent of the "Yankees beating Baltimore when Baltimore couldn't win to save their life", but as the odds of the United States defeating a country that you wouldn't even think of as having a baseball team such as say, Iraq.

Here's the thing: both of these are impossible to quantify. No scorinator, no matter how complex or well tuned, can possibly take into account "what Sarzonia thinks his results should be", nor can Newmanistan's ranking be translated into "what USA's would be IRL". Changing the formula, or the way it's used, or devising a different formula, depends on objectively stating what form we think the results should take.

Just a couple of other things:

Using your list, that's upsets in 26 of 96 games, or over 25%.

Well - some of those aren't upsets in the sense of being less probable scorinator results, because I wasn't taking into account RP bonus differences.

Using your list, that's upsets in 26 of 96 games, or over 25%. If in the the World Cup, newbies were beating top 20 teams on a 25% clip, there would be about 20 pages of "the scorinator us broke."

The only reason there won't be 20 pages is there's fewer participants; the basic dynamic is no different.

[The Babbage Islands]

We should have well over a 90% chance. This is not a Major League Baseball simulation, it is a World Baseball Classic simulation.

These facts and concerns were developing during WBC15, and now with it happening again, demonstrate a scorinator not ready to be used for the WBC.

As one might expect, I disagree.

First, I don't design scorinators "for the WBC" or for any other specific tournament. I design to try to get numbers that accurately reflect top-flight play--and I think it's been established that the Miscellanator baseball module does that.

Second, any scorinator is affected by the inputs given to it--you, of all people, should remember the maxpoints discussion in World Cup 49. I have already noted in this discussion a likely modification to inputs in my future usage, one that I think will better differentiate teams given the small spread of ranking numbers used in baseball rankings (a smaller spread than in any other major sport, FWIW).

Perhaps we need to go back to the -1/0/+1 home advantage instead of a flat-rate home advantage.

Miscellanator does not use a flat-rate home advantage. The home advantage varies, but averages about 0.5 runs per game.

[Tarrentum]

In a couple of competitions he hosted, Dancougar would not apply tiebreakers until about the two-thirds or three-quarters point. However, when it came down to the stretch run, he would apply them accurately. I think that's what Tarrentum's doing here. If it's not, that's a whole separate issue that I would consider far more serious than which scorinator is spitting out which random numbers.

Granted, I try to apply head-to-head in the tournaments I've hosted even from Matchday 2 when I first start publishing tables, but I haven't exactly run a very large tournament, at least not for several years.

That's exactly what I'm doing. There are two many "A beat B but lost to C who beat B, but C hasn't played A" then throw in "E has beaten C and A has beaten D" and you can see where we are going with this. The point of my post was that there are too many tied teams right now for a correct H2H comparison becuase not everyone that has the same record in a group has played each other.

Liventia, while I respect your opinion, if you want to go through all of the matchups and give us a clear and concise standings then feel free...otherwise, you are not a part of this WBC on your own choosing. You had reservations and that's why you didn't participate, no need to come back and say, "I told you so."

[Liventia]

Not being a part of the WBC doesn't mean I'm not an observer. I've hosted two past WBCs myself. Let's make it clear. I withdrew because of time concerns; originally my post about my signup being provisional on host was in reference to a n00b who ultimately chose not to bid. You've also just pretty much ensured I will not be voting for any future bid from you with that tone.

[Newmanistan]

First, I don't design scorinators "for the WBC" or for any other specific tournament. I design to try to get numbers that accurately reflect top-flight play--and I think it's been established that the Miscellanator baseball module does that.

Second, any scorinator is affected by the inputs given to it--you, of all people, should remember the maxpoints discussion in World Cup 49. I have already noted in this discussion a likely modification to inputs in my future usage, one that I think will better differentiate teams given the small spread of ranking numbers used in baseball rankings (a smaller spread than in any other major sport, FWIW).

Miscellanator does not use a flat-rate home advantage. The home advantage varies, but averages about 0.5 runs per game.

First point: I get your point, and Miscellanator does do that. The difference in my frame of thought about WBC qualifying and your opinion of it, and also this will kind of reply to QoD's post; is that WBC qualifying should not be a "top flight" competition. Once we're into the playoffs, then it becomes that. Qualifying, when a strong nation plays a weak nation, the weak nation should only very rarely have a chance at winning. The World Cup qualifying does exactly this. As I scroll through the scores thread, after 17 games, most group leaders have zero or one loss. If basic forumla used in Miscellanator was used, then with a near 25% upset frequency, most of these top teams would have 3 or 4 losses. That would not go over well in the WCDT.

As for the World Cup 49 comparison, I have very much thought about that as being the culprit. With as many games having been "upsets", it looks like that type of situation. WC 49 is also what I was thinking about when I mentioned that if this was a World Cup, we'd have 20 pages on this by now. As for modifications to inputs as you mentioned before, yes, please, let's try that.

[Secristan]

I really hope this does not discourage anyone from RPing, however.

I'm sorry but at this time I just don't see the point. I'll be back for WBC 17.
No offense or anything, I probably still would if I wasn't about to host IBC 6.

[Elemanu]

sigh

Let's have a good look at these results. I'm going to concentrate on the BergBucks group, because a) it's my group, and b) this relates to Secristan and Sarzonia.

The current table looks like this:

N. Team            W  L
1. Freelassio      4  1
2. Elemanu         4  1
3. Sarzonia        3  2
4. Secristan       2  3
5. Voltronica      2  3
6. Silver Beach    2  3
7. Michael VII     2  3
8. Allamunnic Sta 1  4

Now, let's have a look at the remaining three games of these cycle. We can safely assume that the top two teams won't lose to unranked teams with this scorinator (Sarzonia has only lost to the ranked Silver Beach and a fluke to Michael VII thanks to a 7 run innings - Secristan has only lost to Sarzonia and the ranked Silver Beach and Allamunnic States). Therefore the table should look like this after every team has played each other once:

N. Team            W  L
1. Sarzonia        6  2
2. Secristan       5  3
3. Elemanu         5  3
4. Silver Beach    4  4
5. Voltronica      4  4
6. Freelassio      4  4
7. Allamunnic Sta 3  5
8. Michael VII     3  5

I don't think that's so bad...

[Secristan]

Now, let's have a look at the remaining three games of these cycle. We can safely assume that the top two teams won't lose to unranked teams with this scorinator (Sarzonia has only lost to the ranked Silver Beach and a fluke to Michael VII thanks to a 7 run innings - Secristan has only lost to Sarzonia and the ranked Silver Beach and Allamunnic States).
[/pre]

I don't think that's so bad...

My weekend games were losses to Allamunnic States and Silver Beach. I can't stay motivated after that. They have minimal ranking points, and have done nothing to increase that. I don't know how you say we can safely assume people won't lose to unrankeds. It's not the rank that matters, its the ranking points. There's no way we can safely assume anything. I'm still gonna follow it, and I'll likely give Sarzonia something to work off of for our next meeting, but other then that; I'm gonna do what's working for Freelasio, Silver Beach, Grantanata, Bodegraven, Mossulia, Peace Republic, and others.