Team database - bug reports

More
3 years 2 months ago #144409 by Playaveli
Bayern and Barcelona were screwed by default in 96/97... poor clubs!!! :D

Please Log in or Create an account to join the conversation.

More
3 years 2 months ago #144410 by Gorzo
Replied by Gorzo on topic Team database - bug reports

Not as far as I know (after all, they are actually playing at CB/D, so they're not in a wrong position, and the RB/LB guys are generally much happier out wide than in a back 3). This can be tested quite easily by running a few leagues with such a team, then changing the position labels only on that team and running again. I can do this tonight (provided that in SWOS 2020 one can simply edit/replace the data files)

As I recall the 96/97 Bayern players were in very wrong positions, yes? Mario Basler at centre-back?


yes Basler as CB etc
Barcelona, Lyon and many other clubs. I have a corrected version for positions.
you can replace files yourself as far as I know :)

Please Log in or Create an account to join the conversation.

More
3 years 2 months ago - 3 years 2 months ago #144414 by Synchronated
Teams involved in this test:

[some control teams]
7s in a 442 - 442 standard, all player skills are 7
6s in a 442 - 442 standard, all player skills are 6...
5s in a 442
4s in a 442
3s in a 442
[then all 5s in various types of various formations]
433/451 group:
5s 433 STD - standard SWOS 4-3-3 with RW M LW A A A
5s 433 MMM AAA - 4-3-3 with M M M A A A instead
5s 433 MMM RWLWA - 4-3-3 with M M M RW LW A instead
5s 451 STD - standard SWOS 4-5-1
352 group:
5s 352 STD - standard SWOS 3-5-2 with RB D LB RW M M M LW
5s 352 DDD RWLW - 3-5-2 with D D D RW M M M LW
5s 352 DDD RBLB - 3-5-2 with D D D RB M M M LB
343 group:
5s 343 STD - standard SWOS 3-4-3 with RB D LB RW M M LW A A A
5s 343 DDD - with D D D at the back instead
5s 343 DDD SYNC - with wing backs changed to RBLB and wingers changed to RWLW
424/4231 group:
5s 424 STD - with M M RW LW A A
5s 424 4231 - with M M RW LW M A
[all GKs have £1.5m value]
Man Utd, Arsenal, Liverpool left alone as default teams.

League plays x10 (190 matches), results and source team.008 test data file attached

Observations (on default tactics):
- 442 control teams with different skills perform as expected
- 433/451 group: 433 STD and 451 STD both outperform variants of the 433 which have had their position labels changed [at least on default tactics files]
- 424/4231 group: the game really doesn't like the attacking mid in this formation being M instead of standard A.
- 352 group: 352 with def changed to DDD performs similarly the RB-D-LB standard, and beats out standard 442 and 433. but change the wingbacks to RB/LB too and it goes wrong
- 343 group: changing the def to DDD does impact performance in 343. changing wingbacks to RB/LB and wide attackers to RW/LW does no further damage

Observations (on Mozg tactics):
- 433/451 group: 433 STD and 451 STD both still outperform variants of the 433 which have had their position labels changed
- 424/4231 group: the game still really doesn't like the attacking mid in this formation being M instead of standard A. which is a shame. as that guy (in IRL systems like this) is usually an attacking midfielder (if he's a striker it might as well be 442...) and not correctly labeled with A.
- 352 group: 352 with def changed to DDD outperforms slightly (so, still comparable) the RB-D-LB standard on Mozg tactics. but change the wingbacks to RB/LB too and it goes wrong still
- 343 group: changing the def to DDD still does impact performance in 343. changing wingbacks to RB/LB and wide attackers to RW/LW does no further damage
- I think from this test that Mozg tactics make a difference to the result generator, at least in terms of the 433 variants. the default league table really penalised the midfield 3 for putting Ms at LW and RW. in Mozg tactics those guys are actually in midfield and the variants have done better
Attachments:
Last edit: 3 years 2 months ago by Synchronated.

Please Log in or Create an account to join the conversation.

More
3 years 2 months ago - 3 years 2 months ago #144415 by Synchronated
That's those teams against each other. Tests vs real teams (with default tactics as I think this is both what most people are using and where the biggest impact of any negative change would be felt):

3 352 variants vs 20-21 Italian league: DDD RWLW is above STD by 21pts over 190 games (points per real seaon). improvement again but probably negligible/could go either way. DDD RBLB performs poorly. can't help but wonder whether this would really affect results much compared to the abilities of the players/teams if they were real, varied teams.

results attached. more tests to come
Attachments:
Last edit: 3 years 2 months ago by Synchronated.

Please Log in or Create an account to join the conversation.

More
3 years 2 months ago - 3 years 2 months ago #144416 by Playaveli
Where is this test with all teams being super teams (complete 7s in all categories) and each one using a different standard formation, having the player roles as standard SWOS tactics would like them.
Only then we'd know how well each one is holding up.

I might give this a try tomorrow on my stream as I did here:
www.sensiblesoccer.de/forum/swos-related...-experiment-analysis
Last edit: 3 years 2 months ago by Playaveli.

Please Log in or Create an account to join the conversation.

More
3 years 2 months ago - 3 years 2 months ago #144418 by Synchronated
You can compare the 5s with each other to see how they hold up. The standard versions of each tested tactic are all here with all 5s specifically for said comparison.

We can't further test 7s against real teams because they would destroy them. So you only have 'lab data' on 7s not 'real world data' (and I suspect the 7v7 lab data would be similar to the 5v5 lab data, but even then less representative of a test of real team ratings)

At least we know from the above that, if any of these are a preferred course then
- it is better to put an IRL M M M RW LW A into 451 than shoehorn it into a 433 (bcs apart from anything else SWOS has 2 of the central midfielders playing on the wing and clearly doesn't like it when they are labeled Ms). the 451 standard version easily outperforms the altered 433 versions and outperformed 433 standard in default tactics
- using DDD for a back 3 in 352 is fine at least in terms of not negatively affecting result performance (and why should it, the players in question are playing at centre back) and may even positively affect it
- using RB/LB for the wing backs in 352 is not fine, they still need to be RW/LW
- using M for the attacking mid in a 424/4231 has a negative league result effect even if it's correct data
Last edit: 3 years 2 months ago by Synchronated.

Please Log in or Create an account to join the conversation.

More
3 years 2 months ago - 3 years 2 months ago #144428 by adrianozzo

We choose the third kit, to avoid similarities. Orange is better contrast to white-black home kit.

Yeah, sure, but playing in orange two thirds of the season (against all teams with a black or white kit, even home matches) is kind of lame
Last edit: 3 years 2 months ago by adrianozzo.

Please Log in or Create an account to join the conversation.

More
3 years 2 months ago #144441 by Playaveli
I guess you have no choice. Playing in similar looking kits is even lamer... :P

Please Log in or Create an account to join the conversation.

More
3 years 2 months ago - 3 years 1 month ago #144445 by captaincofresi

We choose the third kit, to avoid similarities. Orange is better contrast to white-black home kit.

This was a reply to a comment on the Juventus third kit versus the second kit.

(1) I play with Liverpool and mostly I do play with the all red home kit. The next kit that shows up most in SWOS2020 is the all black away kit, in the Premier League this works fine. Only recently I played a match with a all white away kit versus Barcelona. Although in 2020 / 2021 there is no all white away kit for Liverpool. So is that just chosen by the SWOS2020 team for best contrast to other shirts? Or is that chosen by the CPU and not by the SWOS2020 team at all? And thus 'something' in SWOS2020 now overrules the all black second kit.

Officially for Liverpool the first kit is red, the second kit is turquiose (greenish / blueish; although in real life I have not seen this much), the third kit is black (seen this one mostly used in real life and also in SWOS2020 in the English Premier League): www.nike.com/nl/liverpool-fc?cp=63928816...AvD_BwE&gclsrc=aw.ds . So I do kinda like that all black is the second kit in the SWOS2020 game, and just if a third kit is possible in the future I would opt for all green (as they recently used versus Manchester United in the FA Cup).

(2) And I am also wondering if the lack of the center red banner on the blue home kit of PSG is left out on purpose, is thas also due to better contrast to other kits? Or is that because of the strange way SWOS chooses the home / away kits?

EDIT: What I don't get is why, when playing a red/blue team like Barcelona, Liverpool is not just playing with the all black second kit. Since there is NO color clash there. Or is the CPU algorithm still set to the 'orginal' first / second kits of those teams, that it has to give Liverpool the all white kit (because the second kit in the past was blue maybe?). I am just curious how SWOS / SWOS2020 decides when not to use the first or second kits available in the game.
Last edit: 3 years 1 month ago by captaincofresi. Reason: Different insight and just curious how this 'third kit' works in SWOS / SWOS2020

Please Log in or Create an account to join the conversation.

More
3 years 1 month ago #144504 by Kanchelskis
Hi guys I've fixed Africa using your players from the database (for the new ones using Skill Generator):
Congo in now Republic of Congo with proper players
Added DR Congo and Gambia as new teams
Kits updated for Nigeria.

Interested in having it?

Please Log in or Create an account to join the conversation.

Time to create page: 0.288 seconds
Powered by Kunena Forum