Alright. v4 didn't have an amazing result overall, but it is now the 2nd highest scoring tactic on the table. I'll take it, especially since the play is still visually very appealing to me. As mentioned, basically all of these tactics can probably be improved if you swap in a DLP at the DM spot, but I just prefer the look of the RPM. I haven't played my actual save in like 2 weeks because of all the testing, so I think I'm truly done with this thread for a while. Thanks for coming on this journey with me, and I'm glad I got some good results with v3 at least.
Ok hopefully my last version. I think the best performing version would be to take v3 and just swap in a DLP with no other changes, but then it becomes a lot like Void and I'm trying to keep this a bit more unique. Also, yhe most attractive football has come from this new version 4 I just uploaded.
V4 has dribble less, has less dribbling PIs for players, and lost the "play through middle" TI. Otherwise is similar to v3, but it did perform better in my tests. It definitely feels different to watch, with very pretty short passing plays and a lot more pass & move goals. V3 was very successful but so many goals came from just booting it to a forward on the counter and watching them run through the defense on the way to a goal, while my ideal goal is a lot of teamwork and quick passes to penetrating runners.
Tactic:
Results (I managed Man City and Valencia on holiday mode):
Ok, I'm really pleased with a 6.8! It's also one of the top scoring tactics, so I finally achieved my initial goals in this thread.
I'm still working on a few more tweaks. I've played with switching the DM to a DLP, and adding in less dribbling (I added the "Dribble Less" TI , and got rid of all the "Dribble More" PIs, which was previously on almost every role).
Comparisons with other versions:
The v3 uploaded in this thread still looks like the best tactic for an elite team, since nothing has approached that top Man City result - which is why I haven't uploaded a new version yet. But the other options look very strong in their own way, as good or better than v3. I'll say that the Dribble Less versions I've tested are extremely attractive to watch, with way more short passing and teamwork-based goals. The RPM version was the most fun to watch but the DLP performed slightly better.
I'll keep noodling and see if there's anything I want to upload, but I'm glad to see that you don't need to ONLY use SS and DWs to score a lot. My next trick will be to get similar results without using CM-As in the midfield - seems impossible, though, since they're just so good this patch.
I have a new testing league that I made for myself. It's probably a bit less accurate but sims significantly faster than the FM Base version. I use Man City and Valencia (the best and worst teams in the testing league). Results:
And here's apples to apples comparisons in this testing league for the 3 cyclops, Zaz Blue, and Void 2.0:
This was a very solid result for v3 - managed to keep the Raumdeuter and add in a Trequarista, for more unique roles. I also tested a version that has Attacking mentality and is otherwise identical - it did better for the bad team but was less impressive for Man City, overall had slightly more points but a slightly worse GD. So if you aren't a top team, Attacking mentality might be a safer option. The VA was more pleasing to watch in my opinion, though, so I chose to submit that version.
Instead of looking at the average score for each instruction or player role, maybe it makes more sense to count how many times everything pops up in the best tactics? As in, I don't think it's very useful to say "Dribble Less has an average score of 6.5 and Run At Defense had a 6.8" - I think you can get more out of the analysis by saying something like "Out of the top 50 tactics, 1 had Dribble Less as an instruction, 9 had Run at Defense, and the other 40 had no instruction" (or whatever the actual numbers are).
malte722 said: Well, i could do that but i feel its not my tactic, it's a blend between different tactics.. i may ask for permission to do so? Expand Not sure if you're talking about me, but what you posted looks totally different from my tactic here. Most of the roles are changed, the formation is very different, and the instructions are changed as well.. I bet if you watch a team play with your tactic, it wouldn't look anything like what I posted here - the wingers vs Raumdeuter is a very different experience. It definitely looks like a new tactic to me, and those results deserve a new thread!
Fair enough! I did another test though and I think it was a fluke, v3 ended up being very similar to v2. Doesn't seem like there's a huge difference between DLP and RPM after all. I'll keep tinkering to see if I can get a meaningful and repeatable performance increase.
Looks like I nailed my guess for v2, with a 6.0 score. I also lied, I'm still tinkering. I'm not going to keep uploading because I know there's protocols against just trying to brute force things. However, I changed the DM to a RPM and added in the instructions "slightly shorter passing," "play through middle," and "be more expressive" and got an insane result, a 2.15 p/g and 1.43 gd/g. For reference when I did the same test with zaz blue I got 2.13 and 1.54. Most likely just a lucky small sample but I think it's indicative of the top end potential here.
I'm adding in a v2 to this thread. Probably not as good as whatever tweak malte made, but at least with my personal tests it was an upgrade. The big changes were moving to Attacking mentality and changing one of the Raumdeuters to an AP-A. I also made smaller changes to both player and team instructions, and added in some corner set pieces.
Improved Results:
It's possible this was just a big high roll, but it's markedly better than my original. I ran another test of V1, as well as a test of Zaz Blue and Void 2.0, using all the exact same testing database and settings for an apples to apples comparison:
So this version was slightly less high scoring than v1 but really improved the defense. Definitely not in the same ballpark as Zaz and the GD suggests that it's much worse than Void still. And obviously this is not the same environment and a much smaller sample size than the full FM-Arena test. Still seems to be better than v1.
I don't think I'm going to iterate any further since at this point with the new instructions I'm dangerously close to just copying the Valhalla tactic linked earlier in the thread. And I'm also fine if this doesn't warrant a 2nd test, considering I would at best expect this to be like a 6.0, not exactly groundbreaking, and there's always the potential it tests the same or even worse than v1. If you're in love with this formation I still would suggest Valhalla, but as someone who likes the way Raumdeuters work, I'm hopeful this is an acceptable and slightly unique option.
Hmm, the test ended up being pretty unexpected. 5.5 score is actually as good as I'd hoped for from a points perspective, but the GF total was pretty uninspiring, so I think it's an overall failure.
Like the rest of the thread, I'm curious to know what malte did to tweak this. Either he got insanely lucky with a small tweak, or he did something major that deserves its own thread - since my tactic as presented does not seem capable of consistently getting those results, even with Liverpool.
oh wow. My apologies for not doing the research, I didn't realize there was already such a similar tactic. Yes, Valhalla is basically the same thing. My guess is that mine might score a little more but would probably be much less successful overall, so it's fine if this doesn't deserve a test - and I'll look at other similar tactics more before I submit again!
Hi! This tactic is kind of a mix between the Penetration Psycho and Zaz Blue tactics on this site. Penetration Psycho is the highest scoring option on the table right now, and Zaz Blue is the most successful overall. I don't expect this to be the best performing test, but I wanted to see if I could make a tactic that could even approach the amount of goals from PP can while being slightly less porous defensively.
I tested it using the FM Base testing database (without FMRTE options). It basically achieved what I was looking for - a pretty successful tactic that poured in the goals:
For anyone that doesn't know, the FM Base database has a 165CA team and a 140CA team, to simulate a "good team" and "bad team" setup. There are 40 teams in the test league. The 165CA team finished 3rd with the 2nd most goals and 2nd best differential. The 140CA team somehow scored the most goals in the league, but gave up the 3rd most. Still, 11th place and 9th best differential is pretty successful in that database. I also didn't do much work on set pieces, so things could probably be improved even more there.
I'm curious to see how it does in this testing environment.
V4 has dribble less, has less dribbling PIs for players, and lost the "play through middle" TI. Otherwise is similar to v3, but it did perform better in my tests. It definitely feels different to watch, with very pretty short passing plays and a lot more pass & move goals. V3 was very successful but so many goals came from just booting it to a forward on the counter and watching them run through the defense on the way to a goal, while my ideal goal is a lot of teamwork and quick passes to penetrating runners.
Tactic:
Results (I managed Man City and Valencia on holiday mode):
I'm still working on a few more tweaks. I've played with switching the DM to a DLP, and adding in less dribbling (I added the "Dribble Less" TI , and got rid of all the "Dribble More" PIs, which was previously on almost every role).
Comparisons with other versions:
The v3 uploaded in this thread still looks like the best tactic for an elite team, since nothing has approached that top Man City result - which is why I haven't uploaded a new version yet. But the other options look very strong in their own way, as good or better than v3. I'll say that the Dribble Less versions I've tested are extremely attractive to watch, with way more short passing and teamwork-based goals. The RPM version was the most fun to watch but the DLP performed slightly better.
I'll keep noodling and see if there's anything I want to upload, but I'm glad to see that you don't need to ONLY use SS and DWs to score a lot. My next trick will be to get similar results without using CM-As in the midfield - seems impossible, though, since they're just so good this patch.
New formation:
I have a new testing league that I made for myself. It's probably a bit less accurate but sims significantly faster than the FM Base version. I use Man City and Valencia (the best and worst teams in the testing league). Results:
And here's apples to apples comparisons in this testing league for the 3 cyclops, Zaz Blue, and Void 2.0:
This was a very solid result for v3 - managed to keep the Raumdeuter and add in a Trequarista, for more unique roles. I also tested a version that has Attacking mentality and is otherwise identical - it did better for the bad team but was less impressive for Man City, overall had slightly more points but a slightly worse GD. So if you aren't a top team, Attacking mentality might be a safer option. The VA was more pleasing to watch in my opinion, though, so I chose to submit that version.
Not sure if you're talking about me, but what you posted looks totally different from my tactic here. Most of the roles are changed, the formation is very different, and the instructions are changed as well.. I bet if you watch a team play with your tactic, it wouldn't look anything like what I posted here - the wingers vs Raumdeuter is a very different experience. It definitely looks like a new tactic to me, and those results deserve a new thread!
Improved Results:
It's possible this was just a big high roll, but it's markedly better than my original. I ran another test of V1, as well as a test of Zaz Blue and Void 2.0, using all the exact same testing database and settings for an apples to apples comparison:
Cyclops v1: 1.74 P/G, 2.54 GF/G, 1.91 GA/G, 0.63 GD/G
Zaz blue: 2.13 P/G, 2.46 GF/G, 0.93 GA/G, 1.54 GD/G
Void 2.0: 1.97 P/G, 2.72 GF/G, 1.34 GA/G, 1.38 GD/G
Cyclops v2: 2.06 P/G, 2.49 GF/G, 1.39 GA/G, 1.11 GD/G
So this version was slightly less high scoring than v1 but really improved the defense. Definitely not in the same ballpark as Zaz and the GD suggests that it's much worse than Void still. And obviously this is not the same environment and a much smaller sample size than the full FM-Arena test. Still seems to be better than v1.
I don't think I'm going to iterate any further since at this point with the new instructions I'm dangerously close to just copying the Valhalla tactic linked earlier in the thread. And I'm also fine if this doesn't warrant a 2nd test, considering I would at best expect this to be like a 6.0, not exactly groundbreaking, and there's always the potential it tests the same or even worse than v1. If you're in love with this formation I still would suggest Valhalla, but as someone who likes the way Raumdeuters work, I'm hopeful this is an acceptable and slightly unique option.
Like the rest of the thread, I'm curious to know what malte did to tweak this. Either he got insanely lucky with a small tweak, or he did something major that deserves its own thread - since my tactic as presented does not seem capable of consistently getting those results, even with Liverpool.
I tested it using the FM Base testing database (without FMRTE options). It basically achieved what I was looking for - a pretty successful tactic that poured in the goals:
For anyone that doesn't know, the FM Base database has a 165CA team and a 140CA team, to simulate a "good team" and "bad team" setup. There are 40 teams in the test league. The 165CA team finished 3rd with the 2nd most goals and 2nd best differential. The 140CA team somehow scored the most goals in the league, but gave up the 3rd most. Still, 11th place and 9th best differential is pretty successful in that database. I also didn't do much work on set pieces, so things could probably be improved even more there.
I'm curious to see how it does in this testing environment.
Thanks!