[FM23] 4-team groups at World Cup 2026

[themodelcitizen]

Apparently FIFA are second-guessing 3-team groups in 2026 (can you blame them?), which we see in FM23.

Here's a format with 4-team groups:

Screenshots:

Spoiler

DOWNLOAD (backup)

The hard-coded seeding (for 16 groups) is off but seems to fix itself by shifting seeds into the next pot.

I can't get teams into specific groups (i.e. USA in group A, Canada in group B, Mexico in group C) without breaking stuff, but the default 3-team group version has this same problem anyway.

As far as I can see everything else is working fine, I haven't touched qualifying.

Edited September 7, 2023 by themodelcitizen

[themodelcitizen]

I don't think there's a way to put Mexico and Canada in specific groups without breaking qualifying (same issue in default game).

In the groups there's usually 2 games at the same time due to the crowded schedule. Qatar was doing games at 1, 4, 7, and 10 PM local, but I really don't see 10 PM local happening in North America, lol. Mind you, 7 pm kickoffs in 2026 aren't necessarily simultaneous, not that the game recognizes timezones.

I tried to use the Spain '82 stage flag in the second round (matchday 1 winners rest on matchday 2) but it breaks offset dates/specific stadia

[itiadeadrian]

Hi, 

That looks super awesome.
I'm presently also trying to do something like this and I was wondering if I could ask you some questions?

https://community.sigames.com/forums/topic/570908-changing-48-team-world-cup-format/
Are you using the 48 team version rules as a base and editing from there?

If so how are you managing to get your knockout rounds to work
At present mine generates the quarters and then goes straight to the final no semis. Progessing from Stage to stage works but the rounds in the cup stage are a real issue. I've got team mapping and cup tie mapping and the dates all set up and on the test pages it does generate rows for the games but they have no entries

Also did you setup up the 2nd groups stage to form its groups in any particular order i.e. presetting what group and position in the previous stage goes into which group in the next stage? would you be ok telling me how you managed that 

Any help that you could give would be really appreciated

Thank you for your time

[themodelcitizen]

Yeah, using the rules as a base. I *think* this breaks the hard-coded hosting so I was working on that, adding hosts until 2098. Did you get it to keep Qatar in 2022 but also go back to the random-ish hosts from 2030?

For the knockouts, I think I had to go in and check the boxes for "next round for winner" and "next round for loser" - set the appropriate round for the winner (semis) and set the loser to "-1". Cup tie mapping might be working against you if there aren't actually teams corresponding to that team index.

For the second group stage, i Just had teams "qualify" from the first stage (I think I scrapped the "add room for teams" entry that was in there), and did a "treat seedings as pots" with teams seeded by their position in the previous stage. Obviously there are too many first seeds but it seems to adjust (same as the first round). Also the entry to keep teams from same previous group apart... aside from that I wanted it random, barring the ability to put Canada or Mexico in their home group (should they qualify) and keep the US out of it (likely something we can figure out)

[itiadeadrian]

So because I only took the 48 team rules and I've only been working in the editor, haven't created an actual fmf file my tests don't create a 2022 event, don't know how I going to merge it back

And thanks my mind was reading round as stage and since I was going of the original the round values were higher than they should have been and i didn't see an issue because i thought it was using sub stage to extrapolate

But yeah I've run 5 sims with 2020,2030 and 2034 and Spain's hosted all the 2030s but 2034 was Japan and Korea twice and China and Australia once, I didn't personally setup any hosting order so I guess it's random-ish

*Did a run to 2050 and got Spain, China, South Africa, France, Japan, Brazil

**Ran another to 2090 and got Spain, Japan, South Africa, Germany, Korea, Brazil, England, USA, China, Germany, South Africa, Japan, France, Brazil, Korea, Germany

I don't know if I'd call this random and I find it upsetting but seeing as I don't want to personally pick the host I'll leave it like this unless you have a suggestion of how I can increase the range of nations

 

[themodelcitizen]

Yeah that sounds right, same as default. It's a pretty short list it can choose from. I had it doing that but as soon as I started cleaning the up the 32 team rules (stadiums got messed up if I left as is) then the whole hosting got thrown off. Which might not be that bad, I'd rather pick myself and allow countries like Nigeria and Indonesia to one day host. It's just then we might lose all the cool news items about host bids and such 

 

[themodelcitizen]

  Finally sorted out the 4-team groups at the World Cup without messing up any hard-coded stuff like hosting:

Spoiler

Just working on getting Canada and Mexico into  their home groups now. Final is in Dallas until we hear otherwise. Download link in OP is updated

[giantswjw]

Does this work after 2026 too?

[themodelcitizen]

Yes, same 4-team group format. The hosts are selected in the same semi-random method as with the default game, Spain seem to host 2030 in most tests

[anitamui]

在 2023/8/17 在 AM7点39分, themodelcitizen说:

Yes, same 4-team group format. The hosts are selected in the same semi-random method as with the default game, Spain seem to host 2030 in most tests

Can Confirm, i used Pick random host from 40 nation(Including Spain) and Spain still hosted 2030

[V2023]

On 04/12/2022 at 18:04, themodelcitizen said:

The hard-coded seeding (for 16 groups) is off but seems to fix itself by shifting seeds into the next pot.

I can't get teams into specific groups (i.e. USA in group A, Canada in group B, Mexico in group C) without breaking stuff, but the default 3-team group version has this same problem anyway.

Maybe a bit late, but I think I found a solution to the hard-coded seeding. It appears that the hard-coding only depends on the stage number, so when I added two hidden stages as stage 0 & 1, I was able to change the seeding without any problems.

[themodelcitizen]

Do go on? Or post it to take a look at? Can you still leave "teams" blank or have you filled that out too? I'm always hesitant to make changes like that in case there's stuff going on in the background tied to stage numbers, but it sounds like it's worth it... at the moment i"m just praying each country gets into one of "their" groups

[V2023]

On 02/09/2023 at 03:00, themodelcitizen said:

Do go on? Or post it to take a look at? Can you still leave "teams" blank or have you filled that out too? I'm always hesitant to make changes like that in case there's stuff going on in the background tied to stage numbers, but it sounds like it's worth it... at the moment i"m just praying each country gets into one of "their" groups

Here you go. I filled out teams, since I couldn't think of any other way to create the new pots and to solve the problems of host countries groups being messed up. I found this method by rusty to be the most useful for that. I haven't run a test over several World Cups so far, but I didn't notice any problems with the 2026 WC when I tested it.

V1_FIFA_WC.fmf

[themodelcitizen]

Interesting, I'm gonna think about that. I can't imagine why there's any inherent advantage to using the in-built stages 0 and 1, and the background team selection has to be similar to how you've done it (but yours works better for seedings).

If I were you I might just change it so the top ranked qualified teams (seeding 1) are maxed at 11 teams and not 12, because the host technically takes one of those slots.

Then under "group settings" move the instruction for "max. number of seeds per group" for seeding 1 to the bottom of the list - if it's at the top, the game might put one of them in the same group as seed 0 (the host), which we don't want. But if the game goes first in assigning the max. number of seeds from seeding 2, 3 and 4 first, by the time it gets to seeding 1 it will have no choice but to put them in the remaining groups (that 0 isn't in).

Make sense? Lol, I also noticed in testing that Morocco can (very rarely) host, nice addition to the usual pool of Australia, Korea, Japan, China and S. Africa from the Asian/African choices

[V2023]

10 minutes ago, themodelcitizen said:

If I were you I might just change it so the top ranked qualified teams (seeding 1) are maxed at 11 teams and not 12, because the host technically takes one of those slots.

Then under "group settings" move the instruction for "max. number of seeds per group" for seeding 1 to the bottom of the list - if it's at the top, the game might put one of them in the same group as seed 0 (the host), which we don't want. But if the game goes first in assigning the max. number of seeds from seeding 2, 3 and 4 first, by the time it gets to seeding 1 it will have no choice but to put them in the remaining groups (that 0 isn't in).

Oh, that's an interesting method as well, but I actually used the max number of teams in overall list flag instead of the basic max number, so they should be maxed out at 11 (or in case of 2026 at 9).

[themodelcitizen]

Ahh I didn't see that, my mistake. I think the second part should still stand in that case, the game goes by those rules in a list so when it gets to the rule for only 1 team from seeding 1 - it doesn't have any reason at this point NOT to put one of them in the host group. But if that's the last instruction, there won't be any room in the host group left

[V2023]

25 minutes ago, themodelcitizen said:

I think the second part should still stand in that case, the game goes by those rules in a list so when it gets to the rule for only 1 team from seeding 1 - it doesn't have any reason at this point NOT to put one of them in the host group. But if that's the last instruction, there won't be any room in the host group left

While you're right, there is no reason for it to not do that, in all my testing this never happened. Still, you raise a good point and I'll add the invalid groups for seeding rule (for seeding 1 and group 0), this way there shouldn't be any way for this to happen.

[themodelcitizen]

Ahh yeah that's an even better way to do it. You're right, it may not be an issue in practice most of the time, although I swear I noticed it causing an issue in another tournament so have been keeping it in mind