I don't know about you, but I'm tired of having a ton of Dombrowskis and Twellmans come through the ranks just because there happen to be two Twellmans and three Dombrowskis already in the database. So, I think it would be good to create a large .edt file with a ton of American names (which means names from all over the world) to give a larger scope of names.

However, to make an .edt for a future newgen, we have to use the following:
"FUTURE_REGEN" "First Name" "Common Name" "Surname" "DD.MM.YYYY" "Nation" "Favourite Team"
which means we are creating a specific player, which doesn't seem right to me.

Thus, I think the easiest way to do this is to create a bunch of players in the database, and just have them retired before the game loads. Voila, a bunch of players that never appear in the game, but contribute their names to the overall American pool of players.

So, where do we get these names? I was thinking just using the names of current college players, as that would give a good cross section (ethnically) of who plays soccer with a prospect to possibly play professionally. Using a random name generator such as this one would be effective too:
http://www.kleimo.com/random/name.cfm

Of course, if you wish to put your own name, or to contribute with other names, then that would be cool too.

We can of course combine databases now, and newly created players shouldn't overwrite each other (unless they go by ID, which possibly could overwrite players), so if we were to pool our time and talent, we could get a very comprehensive list going. What do you say?

I noticed - I admit too late to make any real use of it - the option to add extra domestic names to the master DB for a country. It was a feature I had missed and I was pissed. It was new for this year for HR's.

I did/do have a standing request that American names just pull from the entire world for regens, given that's about right. They've done okay with hispanics, but fall flar elsewhere. Canadians are the same, do okay on the francophone names, but not enough ethnic diversity in there for thr rest.

I think it would be good to populate the PDL teams with real youngsters in the hope that it will work. We could also (though not technically correct) add the U17 squad to Bradenton in the PDL.

Also, in practice you might only need this:

"FUTURE_REGEN" "First Name" "" "Surname" "01.01.1993" "United States" ""

Much of which could be copy and pasted with just the name changed (and even then you could probably enter the same surname on a couple and just change the year). So what they'd all have the same birthday each year?!

While its true we could do this, I don't know how birthdays will affect things. I don't want a ton of newgens all having New Years as their birthday. From what I understand, the year doesn't matter too much, but then again, I've never used and .edt to create a newgen, so I'll have to play around with it.

If we do have to use the "FUTURE_REGEN" method, then how does the game know the skin tone and hair color for the newgen? Is there some way to set this as well?

This is a good resource for US names: Most Common Surnames

I'm sure someone who's taken a programming class more recently than I could easily come up with a program that takes random first and last names, pairs them with a random DOB and displays them in the "FUTURE_REGEN" ... format.

I don't think we can control the visual ethnicity of each name, but to me its not so big of a deal.

That's a good resource, I'll have to take a look at that. Approximately 8000 people in America have my last name. Pretty cool.

I could write a program for that, but I'd need a database to pull from, and to be honest, I don't know if that's even worth it.

I would seem to be something that could be coded easily, but what do I know?

It would be a nice feature to have and something that is on the list for next version now I know the functionailty exists. Unfortunately, from what I can tell, they need to be manually entered, 1 at a time in the DB, so a good name edt generator might actually be much more effective.

I'd be happy to make an excel database, i'll even try to separate the names based on what they selected as their ethnicity on the US census. If I do that, can you make the program to randomly select names and pair them with a random DOB?

It would be great if we had 5k or even 10k realistic names to add so future US games stay realistic.

If you're going to go as far as creating a whole database by hand, then what is the need to create a program to take those names and add a birthday?

Two items:

1. I can easily create excel databases which run simple functions to work out the names and percentages based on ethnicity. I'd guess a few hours of work. I certainly wouldnt have to type out every name on those lists. It would take far longer to then create another database with names, DOB's, etc..

2. I couldnt create 'random' player names if I had to choose each element. As you know from our other discussions, I'd prefer to have as little human influence as possible. Sure I could try to pick appropriate names, but it would be 'my' selection of names. Its not a horrible situation if I have to generate them by hand, but a program would remove a great deal of human influence.

I wonder if we even need a D.O.B. for this.

What if were just put names in the following form:

"FUTURE_REGEN" "First Name" "" "Surname" "" "United States" ""

That would make writing a randomizing program unnecessary. I wonder if it would work.

I was thinking about this and as soon as Riz can complete my 2nd nationality request I think this will dramatically increase the number of names in the mix by default, as it will draw on traditional names for those countries as well.

Here's what I requested. It's not 100%, I'm sure, but seemed to be a reasonable cross-section of both the current and up-coming soccer playing community. I've done Canada as well.

This is for the newgen pool. (IE currently into the PDL) but i've asked the name changes/nationalites across the board.

US Only 31.00%
US/Mex 6.00% (this is currently in as Mex/US and should be US/Mex - that was my original error. Not declared)
US/Italian 3.00%
US/Polish 0.50%
US/Portugese 2.00%
US/Irish 2.00%
US/English 1.00%
US/Canadian 2.00%
US/Puerto Rico 8.00%
US/German 2.00%
US/Honduran 0.50%
US/Nigerian 0.50%
US/Ghanaian 0.50%
US/Greece 0.50%
US/Scotland 1.00%
US/Wales 0.50%
US/El Salvador 1.50%
US/Haitian 1.00%
US/Japanese 0.25%
US/Chinese 0.25%
US/Vietnamese 0.50%
US/Ukraine 0.25%
US/Norway 0.25%
US/Sweden 0.25%
US/Denmark 0.25%
US/Serbia 0.25%
US/Croatia 0.25%
US/France 0.25%
US/Spain 0.25%
US/Guatemala 0.25%
US/Russia 0.25%
US/India 0.25%
US/Iraq 0.25%
US/Venezuela 0.25%
US/Peru 0.25%
US/Argentina 0.25%
US/Colombia 0.25%
US/Carribean 0.25%
US/Jamaica 0.25%
US/African nations (random) 2.25%
US/T&T 0.25%
US/Bolivia 0.25%
0.25% - whoops, missed one.
72.00%

Canada    9.00%
Please use these same proportions for all Canadian regens (2nd colum indicates % of all canadians)

Canada Only British sounding 25.00%
Canda Only but french name 15.00%
Canada/African 6.00%
Canada/Argentina 0.25%
Canda/Brazil 0.25%
Canada/Chinese 0.25%
Canada/Croatian 2.50%
Canada/English 4.50%
Canada/French 7.00%
Canada/German 1.50%
Canada/Greece 0.50%
Canda/Honduras 0.50%
Canada/Honk Kong 0.25%
Canada/Indian 0.25%
Canada/Italian 6.00%
Canada/Jamaica 1.00%
Canada/Mexico 0.50%
Canada/Netherlands 4.00%
Canada/Nigeria 1.50%
Canada/Polish 7.00%
Canada/Portuguese 5.00%
Canda/Romania 0.25%
Canada/Russian 0.25%
Canada/Scotish 5.00%
Canada/Slovenia 0.25%
Canada/Switzerland 0.25%
Canada/T&T 1.00%
Canada/Ukraine 0.25%
Canada/US 1.00%
Canda/Serbian 3.00%
100.00%

African 5% 5.00%
Brazil 2%   2.00%
Random S. American 4%   4.00%
Other (Euro) 3% 3.00%
Mexico 3% 3.00%
N. Zealand 1% 1.00%
Other/World 1% 1.00%




100.00%

I've got the databases made (long airport layovers), and using some Excel plug ins I can produce random samples based on ethnicity. It will include a random birthday.

So how many newgens do we want?

Al, the databases I made take the names from the 2000 US Census data. I've got them organized by ethnicity (that was submitted by the census taker). From there I've further developed the data.

I can now see, for example, the number of African Americans who have the last name "Williams" (716,704). The percentage of Williams within the African American community (.0311..) and the percentage of African American Williams' in the whole US name pool (.004419...).

I'd be happy to pass this information along, as it literally contains the real US names (as accurate as of 2000 at least). If this can be added to the names list, or used in future versions I'd be honored.

(also, is it possible to make these newgens with a second nationality?)

I'm not even sure how I would use that? But I'd certainly take it from you to find out.

As for the number of regens....well, 200-300 a year, but I'm not sure if it would add them to the PDL newgens or replace them? We'd have to test to find out.

Optimally I would think that adding a ton of names without birthdays would simply add those names to the collective pool of players, that way there is still some random assignments of things, so we don't have the same player names coming up every time.

Aljarov, well done on the nationality thing. Hope that gets sorted soon. However, only 31% of the newgens will have only American nationality? That doesn't seem right. Perhaps I'm just looking at it the wrong way, but basically only 1 in 3 American players will not have a second nationality? In my survey I found it to be about 2 in 3 American players will not have a second nationality.

crash2772, I don't know think these things work perfectly, as far as making however many newgens you put into the file. Also, having a bunch of Williams or Jones or Smiths in the file will cause the percentage of those names to increase, as these names don't directly affect just the specific newgen they are intended to create, but also on the percentage of first/last names in the database. Think of it as adding first and last names to a giant database. If you have three Dombrowskis, and only one Smith, Dombrowski is three times as likely to pop up rather than Smith. So, including a plurality of Smiths to balance that out is crucial. I personally would not like to include date of birth.

31% out of a possible 72%, so that's half. I think a lot of the US/Mex in the current DB are overlooked. In fact, with all the US players in Mexico I had to go back and even have them add in the US as they were in as just Mexican (Orozco, Padilla, Torres etc...)

I think half is about right, especially going forward as it seems an increasing number of minority players are finally starting to come through if you look at the ranks of e the US U17 and U20 sides.

Fair enough, I knew I had to be looking at it the wrong way. Still, 43% seems low to me, but it will be better than it currently is, for sure.

Ok I've put together a list of ethnically realistic Newgens, using my census database and excel randomizing the combination's. They also have totally random birth dates between 1989 and 2000. You'll be happy to know that there are no Dombrowski's in the list. Rapidshare Download

In this file there are 3 American Indians, 400 African Americans, 400 Hispanic Americans, 10 Arab Americans, and 1000 White Americans. I didnt do the Asian sub-groups yet, because on the census these ethnic group all check the same box. I would have to separate them out to create proper names. Some of the common Arab American surnames are also in the Asian category, but these surnames were also far more likely to select 2 or more races, which made it easier to extract.

If these names do indeed get added to the US pool of eligible names, then this file could cause problems in the future. Currently first names that are overwhelmingly one ethnicity, such as Jose, are only paired with Hispanic American last names in my file. However, if in later generations the game would simply randomize them, we would have Jose paired with any ethnicity, potentially producing strange results such as Jose Yazzie (overwhelmingly Native American - 97.3%) or Jose Zhang (98.16% Asian American). Of course this name can exist in real life, but it would be EXTREMELY rare according to the census numbers.

Al, can you confirm the game will add these Newgen names into the US pool of names? I assume the database's pre-existing names are some how categorized in the game, so very unlikely pairings arent created. Is that correct? If once the newgen was added to the game, we corrected the skin tone, could that help organize future name categories?

If it does add these newgens to the name pool and there is no way to categorize them, then I would suggest limiting the Newgens surnames to those common for African Americans, Asian Americans, Native Americans, and White Americans as a great number of their first names are universal according to census data. Therefore future newgens would have the highest probability of maintaining a statistically likely pairing.

FYI, if anyone wants me to put together a specific list for them let me know. Its all in excel spreadsheets and one click will randomize it all again.

When the game creates Newgens, does it take it from a pool of names or simply from existing players? I was working with the assumption there is a list of names from which these new players are created.

If that is the case, then the file I've created takes the approx 5000 most common surnames from the US census. I've then separated them based on the ethnic group the respondent selected, so if the game can separate the names then even better. From what I read, one of the editing changes this year is the ability to edit skin tones (presumably in the newgens as well).

I'm not sure if anyone has tried the file I made previously, but since then I went ahead and made a few more thorough files. One contains 5,000 unique and random White- american Newgens, and another that contains 3,000 unique and random African-americans.

I'm still awaiting word on if it is possible to further edit/specify various traits with regard to the regens. Otherwise, I imagine all 8,000 of these will turn out looking like white-americans in the bio and on the field.

Together this file accounts for almost 4,300 unique last names, most of which are not in the original database. They have been weighted, according to the census numbers, so there are more Smiths, Jones, and Williams than the less common names.

The first names are a similar situation, and this will add about 1,000 new first names to the US database.

If anyone wants the White-American file, or both, I'm happy to share just PM me. If it is possible to give more characteristics to the newgens in the edt file, then I'll put together a comprehensive file with close to 10k new newgens and over 5,000 new surnames.

I wish I had read this before...the ootp (out of the park) forums have done extensive work on names and ethnicities due to the options available in ootp. They even have things divided historically. I'm sure that we (you)can find invaluable info there to aid in your name authenticity settings..

Could you direct me to the resources you mention? I'd love to expand up on my project! Overall, its all about long term playability. The more unique and realistic names I can come up with, the more we all benefit in long term saves.

Just so everyone knows, these files are relatively small so I dont think even a massive name file would have any noticeable effect on your PC's. The 5,000 person list I made is only 332 kb.

*wonders what I'll do without a team of Canadian players all having Nakajima-Farran for a last name*

Over the weekend I found the time to research the impact of additional names on the database.

I created an .edt regen file that contained 10K names, loaded up the game, and searched for a regen in the first few years with one of the new names. Unfortunately there were NONE. Now I do think that long term some of these players will pop in, but it wont be enough to greatly impact the newgen names.

As a result, I looked into some alternatives. Using the .xml files, I created a list of players 'Tester, Tester' to see if people added this way would impact the database significantly. They did! I added 384 "Tester, Tester" and 125 new players began a game with this first name (so these are game generated coaches, players, etc...). Also, 252 players began the game with this last name. Thirty-two shared both the first and the last.

Of course, for realism I didnt want to add hundreds of imaginary players. I tried working with retired persons, but they dont designate a nationality, so they didnt have any impact. In the end I think I found a satisfactory solution. In my tests, I can add referee's that are too old to actually referee. As a result, these people dont appear in the game as a referee's but their names do come into effect in great numbers for newly created coaches, youth players, etc...

I tried a file containing 1025 referee's, and 938 newgens came into the game using the last names. As a result, using an .xml file seems to the most effective way of adding new names to the database. These files are larger, but still manageable. With a file 1000 names deep, most teams in the US had 2-5 newgens with the new names. If you have a newgen class of 6-8, about half should have these new names!

Unfortunately I dont have a faster way to make thousands of people, so its a slow process-- but I'll work on it and release it as soon as possible.

In case you were wondering, after adding 1000 new names to the database there were no additional Dombrowski's (nor Twellmans) added to the game at start up. However, after a season when the new youth players came out, there was a new Dombrowski. It seems that no matter what we do, since there are so many Dombrowski family members in the game at start up they will forever have an impact. If some of these names start to get too numerous, I would recommend using a program like FMRTE to edit the name to something else. In this program you have to use names already in the database, but with thousands added there should be plenty of choices.

Crash, instead of doing referees, you can create an active player, but set his retirement date to 2008, so he retires before the game loads. This is different from a retired person in the database. I don't know if this will impact the game, but I imagine it would. All it would take is the added:

to the rest of the player's XML file, which sets the player's retirement date to June 1, 2008.

As long as every third Canadian player is no longer named Hoilett, I'll be pleased.

This would work too, but at this point I've got word pad files with thousands of cut and pasted referee's ready to retire. I would imagine the impact is similar, but with the ref method the game automatically retires the player since he's too old.

Just for clarity's sake, none of the ref's have their ages adjusted by the game and show up as a real person. This can happen if you set a player too old/young, but not the ref's apparently (at least if you set their birth year as 1954).

I've got this set to be US players (390) but since I imagine you dont mind the names the US database would generate I could easily change the country code to Canada.

Once I get all 5k names organized, I'll offer a CAN version too.

FYI this wouldnt have many French-Canadian names.

Ok I've spent countless hours working on this, and I finally have some files ready to be tested by more people than just myself. I've created several .xml files.

In these files you'll find the most popular White and African American names, as well as some more obscure White names.

If anyone wants to test these on their machine, please PM me your email address and I'll send it to you.

FYI, although names like Smith and Jones occur frequently, when I tested the game, there were still a lot of Dombrowski's thanks to their family. The only way I see to minimize these would be to create files with literally hundred's of thousands of players. All together, the files I mention here creates about 1,500 players. While the diversity of the newgen names is up, it is still common to find Dombrowski's. I think the best option is to use FMRTE to edit the newgens name away if you dont like it. BTW, it is one of the 5,000 most common names in America, and its obviously a soccer loving family, so having a few more isnt ridiculous.

Hows this coming along? I'd love to test this out.

Lol. This or Nakajima-Farran, which really should never appear.

Yes, Canada's side is dominated by the Hoilett - Nakajima-Farran - Scicluna (I've never met a single Scicluna in my life) trifecta of evil.

How has the US/Canada name pack progressed?

This is a fantastic addition to FM which I always load when playing. I would suggest that crash upload it to here since I'm enjoying it so much. I would also suggest someone start on a Canadian version since that would be very much appreciated.

Hi, I'm planning to do an American and a Canadian version after the release of 10.3.0 patch. What I'm going to do is to create new staff with the surnames that appear only 1 time in the DB (the game only picks for newgens the surnames appearing 2 or more times), and then I'll add surnames that are not currently in the data base.

You might want to get together with Crash to avoid duplicating work.

Yep, I'll try to get in contact with him.

Scicluna is a Maltese last name, no idea how it gets used as a regen name so often in Canada

I'd be happy to help you with it Al. I know how it could be implemented into the db :)

Just make sure to add Focker!! :)