Evoweb Gameplay Lab, FIFA 20 (Frostbite FIFAs?)

[papinho81]

I was hoping this did something when I added it!
Does it slow down the turning animation? It annoys me a lot how fast they can change direction! I hope we can find something on that front.

I think it does. The CPU dribbling looks more grounded. It needs more testing though.
Also I didn't try the next variable yet, the stretch one, but I bet it controls the giant step they take sometime when dribbling.

@Topaz I am curious to see what you are silently cooking!

 

[Chris Davies]

Spent the last five hours messing around (this is why I stopped editing months ago, you can't stop or you forget what you've tried)...

I've actually rolled back lots of changes and tried to get the same effects in other ways. For example, the walk speed increase is a quick fix to make defenders alive, but it means they position themselves at a really silly pace IMO (the kind of speed where you're overshooting the guy you're marking by 3ft either side, over and over again).

What I've found is a better solution for me personally, is to lower the walk speed slightly, and then lower the dribble walk speed, the dribble jog speed, and the jog speed, so that the ratios are more even between each one, and so that the highest speed (jog) is 0.25 max (currently I'm testing 0.215).

I think "dribble" translates as "moving with the ball at your feet", which makes the default ratios look crazy (e.g. walking is 0.1, walking with a ball at your feet is twice as fast - which explains A LOT about how I feel the game plays)...

(I've also discovered an "actors_movement" set of variables that include a global "max speed" and "max acceleration" - with the speed being 53 and the acceleration being 28, IIRC. They make an immediate impact, but I've reset them for now. There are even stopping distances, so you can change how long it takes for your player to slow down or change direction at top speed. One for another night...)

On top of those adjustments, I've increased the field "roll friction" and "slide friction" by 0.1 each, which helps the ball speed, but importantly I've also gone into the "physics_physics" tree and increased ball density slightly, then increased "field static friction" and "field dynamic friction" by 0.2 each, with "ball static friction" and "ball dynamic friction" increased slightly too.

If you go overboard, you can't get a fast passing move going - but without doing all this, every pass is a rocket... Especially when you increase the ball weight (which I have done, to 0.461, as this is a good balance between occasional rocket shots and *always* rocket shots).

I also found a pretty useful value in the "goalkeeper" tree - there's a "one-armed save" value that's set to 1.2 and means keepers can make some really ridiculous saves on the default sliders. Knocking it down to 1.0 takes away a few of the silly wonder-saves.

Bedtime now...

 

[Redhat]

I look forward to seeing what you guys come up with. I've played a lot of FIFA 20 this year, but grew tired of it a month or so ago. I'd love to go back in with some gameplay modifications.

 

[manmachine]

What I've found is a better solution for me personally, is to lower the walk speed slightly, and then lower the dribble walk speed, the dribble jog speed, and the jog speed, so that the ratios are more even between each one, and so that the highest speed (jog) is 0.25 max (currently I'm testing 0.215).

I agree with most. Raising the walk speed seems to speed up the game a fair bit, however with a lowered dribble jog speed the actual animation (while slower) looked kind of funny to me. It seems you and @papinho81 like to experiment with the physics values while I enjoy digging into the positioning stuff. Maybe we can solve defensive issues with tweaking actual behavior values rather than physics. I'm testing on normal speed and sometimes I find the game ok, sometimes too fast. Overall I think I'd love a bit slower movement...

On top of those adjustments, I've increased the field "roll friction" and "slide friction" by 0.1 each, which helps the ball speed, but importantly I've also gone into the "physics_physics" tree and increased ball density slightly, then increased "field static friction" and "field dynamic friction" by 0.2 each, with "ball static friction" and "ball dynamic friction" increased slightly too.

...especially since I've now gone ahead and slowed down most common pass types by 10%.



These are various pass types and their speed, higher values on the y axis means slower since the axis is the "travel time", how long it takes for a pass to reach destination. Green lines are normal ground passes, blue line is lobbed passes, red lines are driven passes. Dotted lines will be my modded values for a first test.

Here's something that made me laugh, at how stupid this game is when it comes to positioning/formations & behavior. I think it's fair to say we're going to get the most benefits by fixing positioning logic. The red line is how much players press (and leave their position) with a high press tactic, blue line is low pressure teams. If it looks like 2 ridiculous extremes, that's because that's exactly what it is. Unfortunately I just can't skip testing some of this stuff (slowing passing was relatively easy), but I hope I can have something today that you guys can test too.

 

[Anth James]

I agree with most. Raising the walk speed seems to speed up the game a fair bit, however with a lowered dribble jog speed the actual animation (while slower) looked kind of funny to me. It seems you and @papinho81 like to experiment with the physics values while I enjoy digging into the positioning stuff. Maybe we can solve defensive issues with tweaking actual behavior values rather than physics. I'm testing on normal speed and sometimes I find the game ok, sometimes too fast. Overall I think I'd love a bit slower movement...



...especially since I've now gone ahead and slowed down most common pass types by 10%.

View attachment 44885

These are various pass types and their speed, higher values on the y axis means slower since the axis is the "travel time", how long it takes for a pass to reach destination. Green lines are normal ground passes, blue line is lobbed passes, red lines are driven passes. Dotted lines will be my modded values for a first test.

Here's something that made me laugh, at how stupid this game is when it comes to positioning/formations & behavior. I think it's fair to say we're going to get the most benefits by fixing positioning logic. The red line is how much players press (and leave their position) with a high press tactic, blue line is low pressure teams. If it looks like 2 ridiculous extremes, that's because that's exactly what it is. Unfortunately I just can't skip testing some of this stuff (slowing passing was relatively easy), but I hope I can have something today that you guys can test too.

View attachment 44886

Really interesting. Good find on the pressing tactic. Are you able to share your passing adjustments for us to test?

 

[papinho81]

Interesting read!

Agree the walkspeed thing was a quick and dirty fix found inadvertently to help with the defending behaviour. It is better if we can directly tackle the dedicated variables. It worth keeping an eye at it though because this difference in speed between the normal and the dribble walk is bizarre.

As for the pass speed, the difference in perception might come from the setting we use. As a manual player and with the default game, I am constantly under hitting my passes and they look like if they have no zip. The power bar is not an option as for everything else than ground passes it becomes super tricky.
I like the idea of slowing down the initial speed of the passes but in my opinion the ball has to roll more than it does on default. I also think the ball should roll more than it does, it is stopping to quickly on default.
I think @manmachine curves looks promising in order to get a finest control on this.

Anyway good job fellas!

 

[Chris Davies]

I agree with most. Raising the walk speed seems to speed up the game a fair bit, however with a lowered dribble jog speed the actual animation (while slower) looked kind of funny to me. It seems you and @papinho81 like to experiment with the physics values while I enjoy digging into the positioning stuff. Maybe we can solve defensive issues with tweaking actual behavior values rather than physics. I'm testing on normal speed and sometimes I find the game ok, sometimes too fast. Overall I think I'd love a bit slower movement...



...especially since I've now gone ahead and slowed down most common pass types by 10%.

View attachment 44885

These are various pass types and their speed, higher values on the y axis means slower since the axis is the "travel time", how long it takes for a pass to reach destination. Green lines are normal ground passes, blue line is lobbed passes, red lines are driven passes. Dotted lines will be my modded values for a first test.

Here's something that made me laugh, at how stupid this game is when it comes to positioning/formations & behavior. I think it's fair to say we're going to get the most benefits by fixing positioning logic. The red line is how much players press (and leave their position) with a high press tactic, blue line is low pressure teams. If it looks like 2 ridiculous extremes, that's because that's exactly what it is. Unfortunately I just can't skip testing some of this stuff (slowing passing was relatively easy), but I hope I can have something today that you guys can test too.

View attachment 44886

This whole post is another reason why I stopped modding months ago (until you pulled me back in with this thread, @papinho81) - because I don't have the knowledge to unlock stuff like this (or the time, given I'm still working and on a training course too).

This post is huge - it nails the biggest problems I have. I'm adjusting physics to achieve the effect of making passes take longer to leave the player's boot (they zip off like bullets) and take more than a second to arrive, but only because I don't know any other way to do it.

If you can roll out a fix to that... Man oh man. Delicious.

 

[papinho81]

I am wandering what difficulty and defending setting are you testing?

 

[Chris Davies]

I am wandering what difficulty and defending setting are you testing?

Good question. Legendary and tactical here. Normal speed, but I prefer slow purely because it makes it easier to keep up with the ridiculous pace of ground passes and dribbling.

 

[papinho81]

Stupid me I should have added my settings. WC and tactical for me, with normal speed and all sliders on default.

 

[manmachine]

I'm testing on legendary / normal speed / default (50/50) sliders.

Couple of questions for those who played this a lot more than me. I play against the same 2-3 teams for testing so I don't know really... Does the cpu try long shots? The cpu loves going wide for a cross or trying to dribble/pass deep into the box but very very rarely shoots from outside the box. Need to know if this is normal for the game in which case I want to find a way to increase the number of shots or if it's something that's actually caused by me modding random stuff.

Is the gameplay still inconsistent? (I guess I know the answer to that one). I can play a couple of games and test normally, then all of a sudden the speed changes and it's lightning fast. So I have to quit playing and not make any progress. Argh, very annoying!

Apart from that, it's not that bad. Maybe I can share some of my changes today. What's more convenient? An .fbproject file for the frosty editor or an exported .fbmod file for the mod manager?

 

[papinho81]

I didn't play it much more than you but I have played against more teams and I would say the CPU loves intricate play inside the box but I have seen some/few long range effort too.
Can't remember the file their is a variable that tells the cpu (should be in the cpuai) what distance should be perceived as far ect.

I didn't particularly notice a session variation speed. At least it didn't stood out to me.

What is your defending setting for the record?

 

[Chris Davies]

I didn't particularly notice a session variation speed. At least it didn't stood out to me.

I notice this MASSIVELY. On PC and on console.

So many times in career mode, I'll have a good-feeling game, and the next game will be in fast-forward (with lots of tiki-taka). I'd put it down to team styles, but I swear I can reload the same match and get a different speed.

However, I don't think many people see this - so perhaps it's just me...

 

[manmachine]

What is your defending setting for the record?

Legacy, always. Shouldn't make a difference. In this game you should keep a distance so that you have time and space to react to the cpu dribbling, so I just use manual jockey 99% of the time.

I am seeing shots, but the cpu's insistence on going wide or dribbling his way into the net is just ridiculous at times, lol. Just try a few more shots damn it! If there's a high shot count in a game, 99% of them must be from crosses/corners/rebounds etc.

 

[papinho81]

I notice this MASSIVELY. On PC and on console.

So many times in career mode, I'll have a good-feeling game, and the next game will be in fast-forward (with lots of tiki-taka). I'd put it down to team styles, but I swear I can reload the same match and get a different speed.

However, I don't think many people see this - so perhaps it's just me...

Oh we'll, I guess I just didn't play the game enough and it is a matter of time before it kicks in.
I usually have time for 1 or 2 game by session, it is may be down to that.

Also before every game there is a random draw that gives a stat boost/handicap to the home and the away team. Could it be that? It might be specific to career mode though.

 

[Topaz]

I think it does. The CPU dribbling looks more grounded. It needs more testing though.
Also I didn't try the next variable yet, the stretch one, but I bet it controls the giant step they take sometime when dribbling.

@Topaz I am curious to see what you are silently cooking!

Honestly bud, have been trying to atleast get the positioning right. It does seem like tactics is really respected in FIFA 20 which is good and bad. Bad because the CDM constraint I am trying to set on the grid shifts based on the tactics. Also something I noticed is that a player's role is also important too when it comes to positioning the CDM so the grid position is all relative based on a player's role and the overall team tactics. However even that doesn't fully solve the problem. The rate at which a CDM tracks back is proportional to the deviation between the defensive work rate and the attack work rate of an individual player. For example if a player has High defensive work rate but also High attacking work rate, thier positioning will be shifted upwards and the time to track back is lower than if a player had High. defensive work rate but Low attacking work rate. It's pretty tedious and frustrating to get a value that is good for all situation.

 

[manmachine]

For example if a player has High defensive work rate but also High attacking work rate, thier positioning will be shifted upwards and the time to track back is lower than if a player had High. defensive work rate but Low attacking work rate. It's pretty tedious and frustrating to get a value that is good for all situation.

Have you tried to tweak the gp_positioning_difficulty_runtime? Looks like you can adjust defensive/attacking workrates separately for each difficulty level. I have increased the defensive ones. There are 6 entries, but 7 difficulty levels I think, which is a bit strange.

 

[Topaz]

Have you tried to tweak the gp_positioning_difficulty_runtime? Looks like you can adjust defensive/attacking workrates separately for each difficulty level. I have increased the defensive ones. There are 6 entries, but 7 difficulty levels I think, which is a bit strange.

Oh i haven't yet. I thought it had to do just with difficulty. I want to make it so that attack workrate doesn't impact tracking back rate. Will play around with that now.

 

[manmachine]

Here's a first test modding passing only and nothing else. But i included the slight increase to the ball size and the pitch friction change.

Passes are 10% slower, driven passes are also slowed down but still relatively faster. It's easy to eliminate them completely if needed.
There is a huge variety of passes in the game. There are even different values for the goalkeeper passes and for the first pass on kick off. No idea why, but I didn't bother to change them. So the first kick of the game or the odd pass from the GK might seem faster. Same with free kicks & corners. No changes to shooting/free kicks etc. Just the basic pass types.

As a bonus (well you might hate it) I've increased the number of contextual errors a bit. You should notice it. Some passes are underhit, overhit, or miss their target a bit (even if they are successful) which can slow down the pace. Passing from the back under pressure with your defenders can be more risky. You will also notice some bad crosses, and the long diagonal balls to the winger should fail more often.

(Tested with 50/50 pass speed/error. No idea what happens if you use non default values!)

http://www.mediafire.com/file/lp60io6w9ezw8ao/passingtestmod.fbmod/file

Edit: Oh I had started experimenting with shot errors and you know what, I included some tweaks to shooting as well... So, you might also see some more bad shots after all.

 

[papinho81]

@manmachine I just gave it a very quick and super short go (can't play much now).
I already seen the increase in error. When the def clear the ball under pressure it doesn't reach the ST like without it, which I really liked. The pass speed looks better to me even though the ball lacks a bit of weigth to my taste (Also I can't stand the bubling when the players are dribbling with the default weight, but surely a matter of preference here).

Does the error you introduced is supposed to affects shooting? Because, whao!, I had wild attempts in front of goal

 

[manmachine]

@manmachine I just gave is a very quick and super short go (can't play much now).
I already seen the increase in error. When the def clear the ball under pressure it doesn't reach the ST like without it, which I really liked. The pass speed looks better to me even though the ball lacks a bit of weigth to my taste (Also I can't stand the bubling when the players are dribbling with the default weight, but surely a matter of preference here).

Does the error you introduced is supposed to affects shooting? Because, whao!, I had wild attempts in front of goal

I started looking into shooting, and you know what, some error values might be shared or I might forgot a file related to shooting in the modfile, lol. But yeah i'd like to see more shots, more errors in shots, and decrease their speed in the future.

That's the best part about passing now. It's harder to just ping a pass to your striker or winger, the team pressing you will likely win it back. I can increase these contextual errors now that I know how if people like an even more scrappy game.

 

[papinho81]

Don't forget to share your insights about the variables you edited and/or a project file at some point so anyone can benefit from your break through and tweak things to its liking (if you don't mind of course). No urgency on doing that though, take all the time you need. Fun should come first or you can loose your mind in the frosty space.

 

[manmachine]

Don't forget to share your insights about the variables you edited and/or a project file at some point so anyone can benefit from your break through and tweak things to its liking (if you don't mind of course). No urgency on doing that though, take all the time you need. Fun should come first or you can loose your mind in the frosty space.

It's the contexteffect files in the gp_kickerror folder. There are hundreds of different contextual errors that you can tweak if you have 1 or 2 years with nothing else to do, I just increased the simpler looking "missrate_scalar" variables which I guess act as modifiers to increase the amount of misskicks, based on player attributes, pressure, angle, ball physics etc. That's why it's now harder to pass with your cbs when they're under pressure (lower attributes + pressure).

See fifa already has believable contextual errors programmed in that could make it a football game after all, it's just that the default gameplay is ping pong passing where you mindlessly press the pass button with your eyes (and brain) shut and the ball just zips to its destination.

 

[papinho81]

Thanks!

 

[manmachine]

Positioning and marking is the toughest nut to crack. There are a lot of moving parts: formations, tactics, player instructions, line settings, marking values, workrates... I can find a balance and the gameplay will be amazingly good with Team A vs Team B, but like crap when you play with other teams. If we can crack this though... boy oh boy. There's even the chance that this can be transformed into an actual football game. (sorry but in its default state, it's anything but!)

 

[papinho81]

I think if we could make the defense shape more ball oriented and to have a bit less space between the lines, it would be a good start for me.

Sorry I didn't have much time to test things lately.

 

[papinho81]

@Chris Davies did you figured out what was difference between the ball density and the ball weight? Because in theory, they should be related.

 

[manmachine]

I think if we could make the defense shape more ball oriented and to have a bit less space between the lines, it would be a good start for me.

Sorry I didn't have much time to test things lately.

Only by a little bit I think and there's the problem that pressing is already ball oriented (nothing like real high press). Players shut down passing options when pressing, in fifa they just mindlessly chase the ball. So you have an easy pass option and the high press tactic is basically trash. If this is applied in the defensive third you would have defenders out of position and an easy pass for a 1v1 every single time. But they do need to be more alert/aggressive/smart if you consider how easy it is to dribble/turn past them or the midfielders who don't mark beyond an imaginary invisible line.

 

[papinho81]

Only by a little bit I think and there's the problem that pressing is already ball oriented (nothing like real high press). Players shut down passing options when pressing, in fifa they just mindlessly chase the ball. So you have an easy pass option and the high press tactic is basically trash. If this is applied in the defensive third you would have defenders out of position and an easy pass for a 1v1 every single time. But they do need to be more alert/aggressive/smart if you consider how easy it is to dribble/turn past them or the midfielders who don't mark beyond an imaginary invisible line.

Sorry I meant more ball sided. A bit like the tactic you can activate with the D pad.
I agree the pressing is badly implemented. It reminds me the gegenpress of pes. One wave of players running like mads at the ball while the rest of the team stays in position. So basically behind the wave you a huge gap and your players which are far too easy to reach. With two/three passes you can escape the pressing wave and reach the free players just behind. If the teams were acting more like team units that actually adjust their positioning according to the ball location it would be more difficult to escape the pressing. And you would be probably forced to play longer ball which are more prone to errors.

 

[papinho81]

I ve had a bit of time to continue testing things which are not necessarily relevant to the core issue of the game

gp_physics_soccerfield_runtime
Field_GrassDensityCoeff: default: 35 tested: 1;99;1000;20
Observations: Affects the ball rolling friction and the height and type of rebounds. With high values rebounds are lower and have more back spin grip.

Here are two video illustrating the changes, the second one is CPU vs CPU:

Spoiler: Videos

I found quite interesting that with a very high value that brake the passes, the cpu hits the ball with a lot of power to make sure it reaches a teammate. In the end it looks like the passes are faster.
I thinks this down to the pass TravelTimes @manmachine shown us. The ball, to be on time in the teammates leg needs to have a lot of initial pace considering the brake that will come after. In general the CPU does fine with those extreme change in physics because they are part of their equation BUT it does bring a bit more error in the execution. May be just not the type of error we want to see in the game.