Here is the link for the spreadsheet. I will populate it with tactics that I feel are worth uploading, instead of the tests I do for smaller changes (tested around 100 variations so far).
ZaZ said: Here is the link for the spreadsheet. I will populate it with tactics that I feel are worth uploading, instead of the tests I do for smaller changes (tested around 100 variations so far).
P.S.: Updated the link to give view permissions. Expand
this is the real way testing to be carried out, by testing with the lower league squads, highly appreciated @ZaZ so the spring tactic is the best of all I can confirm?
max 737 said: this is the real way testing to be carried out, by testing with the lower league squads, highly appreciated @ZaZ so the spring tactic is the best of all I can confirm? Expand
It has the best win rate and goal difference, but if you want to score more goals, Summer is better, and if you want to concede less, then Winter is better.
Skillz92 said: okay brilliant thanks for letting me know, how long does it normally take? Expand
It should be ready later today or tomorrow, since I am only running the four main tactics, and the reference (Highway from @Gerrard). I will also post the results of two runs per tactic (each with 8 different teams), and I did some adjustments to the league to reduce variance (like setting injury proneness to 1).
ZaZ said: It should be ready later today or tomorrow, since I am only running the four main tactics, and the reference (Highway from @Gerrard). I will also post the results of two runs per tactic (each with 8 different teams), and I did some adjustments to the league to reduce variance (like setting injury proneness to 1). Expand
Skillz92 said: okay i will keep and eye on here. thanks again Expand
Results are in the sheet now. Keep in mind that is just one test methodology, and results could be very different using another methodology like FM-Arena. Also, the tactics were tested in real teams, which means their squad could have some positive or negative impact for different formations, roles, or instructions.
hi @ZaZ : thanks for adding the excel file. looking at the sheet, is it safe to assume you ran the same tactics twice for the listed team. Lets take spring as an example, the 1st time you ran it, it ranked 4th but the 2nd time you ran it, it was ranked 12th among all the tactics you have tested till date.
sponsorkindest said: hi @ZaZ : thanks for adding the excel file. looking at the sheet, is it safe to assume you ran the same tactics twice for the listed team. Lets take spring as an example, the 1st time you ran it, it ranked 4th but the 2nd time you ran it, it was ranked 12th among all the tactics you have tested till date. Expand
Yes, that's exactly how you should read it. If you want a combined score of both runs, you can get the average of the scores. For example, the combined score of Spring 0.1 for points would be 7.5, or (6.1 + 8.9) / 2. Keep in mind that my scores are an average of ranks (discarding best and worst values), so smaller is better. In other words, it means Spring 0.1 was on average 7.5th place when you rank points of all runs for each team.
@ZaZ : thanks again! final question, so the reason why there is a high variability in ranking 4th and 12th (for spring) has to do more with the randomness within the game and you are testing for 38 games per team rather than for 10000 games like they do in fm-arena and to a lesser extend the volatility of the tactic. what i mean by volatility is a tactic like summer tend to give weird scores because of high scoring games and that might produce some weird season's during testing.
ZaZ said: Results are in the sheet now. Keep in mind that is just one test methodology, and results could be very different using another methodology like FM-Arena. Also, the tactics were tested in real teams, which means their squad could have some positive or negative impact for different formations, roles, or instructions. Expand
sponsorkindest said: @ZaZ : thanks again! final question, so the reason why there is a high variability in ranking 4th and 12th (for spring) has to do more with the randomness within the game and you are testing for 38 games per team rather than for 10000 games like they do in fm-arena and to a lesser extend the volatility of the tactic. what i mean by volatility is a tactic like summer tend to give weird scores because of high scoring games and that might produce some weird season's during testing. Expand
Some tactics naturally have more variance than others, and I am also not controlling several factors from the game like morale or injuries, like FM-Arena tests do. It is not like having that variance is bad, it is just how statistics work. That's why I ran twice each tactic, so you can combine the results and have a more precise insight of performance. Instead of running the test and showing the best results, I show them all without a bias, and let my scoring system interpret if they did well or not compared to the other tactics. It is not as precise as FM-Arena's test, but it has value as an approximation. Someone who checks the table will know Spring 0.1 can perform not as well in a bad day, and Spring 0.11 will very often produce better results.
Skillz92 said: when is spring 0.11 releasing on here? Expand
the versions 0.XY are tentative, so I am still testing to see what will become Spring 0.2. I will leave it below if you want to try, since it is the current best. The difference is that the two CMs move to AM during possession. I am only updating to that table the versions that are worth showing, because there is no point showing versions with small changes that didn't perform well enough.
ZaZ said: Some tactics naturally have more variance than others, and I am also not controlling several factors from the game like morale or injuries, like FM-Arena tests do. It is not like having that variance is bad, it is just how statistics work. That's why I ran twice each tactic, so you can combine the results and have a more precise insight of performance. Instead of running the test and showing the best results, I show them all without a bias, and let my scoring system interpret if they did well or not compared to the other tactics. It is not as precise as FM-Arena's test, but it has value as an approximation. Someone who checks the table will know Spring 0.1 can perform not as well in a bad day, and Spring 0.11 will very often produce better results.
the versions 0.XY are tentative, so I am still testing to see what will become Spring 0.2. I will leave it below if you want to try, since it is the current best. The difference is that the two CMs move to AM during possession. I am only updating to that table the versions that are worth showing, because there is no point showing versions with small changes that didn't perform well enough. Expand
I also think they have ramped up the RNG this year, or at least it feels that way. I reckon last year Spring 0.1 would have been more consistent in your own testing even without FM-Arena's 1000 games
To make it easier for people to find the best results, I added a field for "ZaZ Score", which basically gives a grade to each tactic (10 is the best).
Steelwood said: I also think they have ramped up the RNG this year, or at least it feels that way. I reckon last year Spring 0.1 would have been more consistent in your own testing even without FM-Arena's 1000 games Expand
I will run two more tests for each tactic, for a total of 4 runs of 8 teams. In each run, the results of best and worst teams are discarded, and my final score also discards the best and worst runs during calculation. That should give a better grasp of what tactic is expected to perform better, despite the variance and not running as many tests as FM-Arena.
ZaZ said: To make it easier for people to find the best results, I added a field for "ZaZ Score", which basically gives a grade to each tactic (10 is the best).
I will run two more tests for each tactic, for a total of 4 runs of 8 teams. In each run, the results of best and worst teams are discarded, and my final score also discards the best and worst runs during calculation. That should give a better grasp of what tactic is expected to perform better, despite the variance and not running as many tests as FM-Arena. Expand
Have you tried testing the 'extreme' offensive (most GS) and defensive (least GA) tactics from FM24 as a IP/OOP pair in FM26? I'm curious to see if the IP/OOP system can handle these extremes within the same 'tactic'.
johnconnerson said: Have you tried testing the 'extreme' offensive (most GS) and defensive (least GA) tactics from FM24 as a IP/OOP pair in FM26? I'm curious to see if the IP/OOP system can handle these extremes within the same 'tactic'. Expand
I will investigate in the future. Thank you for the feedback.
Here is the link for the spreadsheet. I will populate it with tactics that I feel are worth uploading, instead of the tests I do for smaller changes (tested around 100 variations so far).
P.S.: Updated the link to give view permissions.
ZaZ said: Here is the link for the spreadsheet. I will populate it with tactics that I feel are worth uploading, instead of the tests I do for smaller changes (tested around 100 variations so far).
P.S.: Updated the link to give view permissions.
this is the real way testing to be carried out, by testing with the lower league squads, highly appreciated @ZaZ
so the spring tactic is the best of all I can confirm?
max 737 said: this is the real way testing to be carried out, by testing with the lower league squads, highly appreciated @ZaZ
so the spring tactic is the best of all I can confirm?
It has the best win rate and goal difference, but if you want to score more goals, Summer is better, and if you want to concede less, then Winter is better.
Thank you @ZaZ I will test it out and let you know. Many thanks!!
@ZaZ what me was the spring tactic created for?
Skillz92 said: @ZaZ what me was the spring tactic created for?
I don't remember when it was tested, but I am running my tactics again in the current engine.
ZaZ said: I don't remember when it was tested, but I am running my tactics again in the current engine.
okay brilliant thanks for letting me know, how long does it normally take?
Skillz92 said: okay brilliant thanks for letting me know, how long does it normally take?
It should be ready later today or tomorrow, since I am only running the four main tactics, and the reference (Highway from @Gerrard). I will also post the results of two runs per tactic (each with 8 different teams), and I did some adjustments to the league to reduce variance (like setting injury proneness to 1).
ZaZ said: It should be ready later today or tomorrow, since I am only running the four main tactics, and the reference (Highway from @Gerrard). I will also post the results of two runs per tactic (each with 8 different teams), and I did some adjustments to the league to reduce variance (like setting injury proneness to 1).
okay i will keep and eye on here. thanks again
Skillz92 said: okay i will keep and eye on here. thanks again
Results are in the sheet now. Keep in mind that is just one test methodology, and results could be very different using another methodology like FM-Arena. Also, the tactics were tested in real teams, which means their squad could have some positive or negative impact for different formations, roles, or instructions.
hi @ZaZ : thanks for adding the excel file. looking at the sheet, is it safe to assume you ran the same tactics twice for the listed team. Lets take spring as an example, the 1st time you ran it, it ranked 4th but the 2nd time you ran it, it was ranked 12th among all the tactics you have tested till date.
sponsorkindest said: hi @ZaZ : thanks for adding the excel file. looking at the sheet, is it safe to assume you ran the same tactics twice for the listed team. Lets take spring as an example, the 1st time you ran it, it ranked 4th but the 2nd time you ran it, it was ranked 12th among all the tactics you have tested till date.
Yes, that's exactly how you should read it. If you want a combined score of both runs, you can get the average of the scores. For example, the combined score of Spring 0.1 for points would be 7.5, or (6.1 + 8.9) / 2. Keep in mind that my scores are an average of ranks (discarding best and worst values), so smaller is better. In other words, it means Spring 0.1 was on average 7.5th place when you rank points of all runs for each team.
@ZaZ : thanks again! final question, so the reason why there is a high variability in ranking 4th and 12th (for spring) has to do more with the randomness within the game and you are testing for 38 games per team rather than for 10000 games like they do in fm-arena and to a lesser extend the volatility of the tactic. what i mean by volatility is a tactic like summer tend to give weird scores because of high scoring games and that might produce some weird season's during testing.
ZaZ said: Results are in the sheet now. Keep in mind that is just one test methodology, and results could be very different using another methodology like FM-Arena. Also, the tactics were tested in real teams, which means their squad could have some positive or negative impact for different formations, roles, or instructions.
when is spring 0.11 releasing on here?
sponsorkindest said: @ZaZ : thanks again! final question, so the reason why there is a high variability in ranking 4th and 12th (for spring) has to do more with the randomness within the game and you are testing for 38 games per team rather than for 10000 games like they do in fm-arena and to a lesser extend the volatility of the tactic. what i mean by volatility is a tactic like summer tend to give weird scores because of high scoring games and that might produce some weird season's during testing.
Some tactics naturally have more variance than others, and I am also not controlling several factors from the game like morale or injuries, like FM-Arena tests do. It is not like having that variance is bad, it is just how statistics work. That's why I ran twice each tactic, so you can combine the results and have a more precise insight of performance. Instead of running the test and showing the best results, I show them all without a bias, and let my scoring system interpret if they did well or not compared to the other tactics. It is not as precise as FM-Arena's test, but it has value as an approximation. Someone who checks the table will know Spring 0.1 can perform not as well in a bad day, and Spring 0.11 will very often produce better results.
Skillz92 said: when is spring 0.11 releasing on here?
the versions 0.XY are tentative, so I am still testing to see what will become Spring 0.2. I will leave it below if you want to try, since it is the current best. The difference is that the two CMs move to AM during possession. I am only updating to that table the versions that are worth showing, because there is no point showing versions with small changes that didn't perform well enough.
@ZaZ which is the current best?
ZaZ said: Some tactics naturally have more variance than others, and I am also not controlling several factors from the game like morale or injuries, like FM-Arena tests do. It is not like having that variance is bad, it is just how statistics work. That's why I ran twice each tactic, so you can combine the results and have a more precise insight of performance. Instead of running the test and showing the best results, I show them all without a bias, and let my scoring system interpret if they did well or not compared to the other tactics. It is not as precise as FM-Arena's test, but it has value as an approximation. Someone who checks the table will know Spring 0.1 can perform not as well in a bad day, and Spring 0.11 will very often produce better results.
the versions 0.XY are tentative, so I am still testing to see what will become Spring 0.2. I will leave it below if you want to try, since it is the current best. The difference is that the two CMs move to AM during possession. I am only updating to that table the versions that are worth showing, because there is no point showing versions with small changes that didn't perform well enough.
I also think they have ramped up the RNG this year, or at least it feels that way. I reckon last year Spring 0.1 would have been more consistent in your own testing even without FM-Arena's 1000 games
max 737 said: @ZaZ which is the current best?
To make it easier for people to find the best results, I added a field for "ZaZ Score", which basically gives a grade to each tactic (10 is the best).
Steelwood said: I also think they have ramped up the RNG this year, or at least it feels that way. I reckon last year Spring 0.1 would have been more consistent in your own testing even without FM-Arena's 1000 games
I will run two more tests for each tactic, for a total of 4 runs of 8 teams. In each run, the results of best and worst teams are discarded, and my final score also discards the best and worst runs during calculation. That should give a better grasp of what tactic is expected to perform better, despite the variance and not running as many tests as FM-Arena.
ZaZ said: To make it easier for people to find the best results, I added a field for "ZaZ Score", which basically gives a grade to each tactic (10 is the best).
I will run two more tests for each tactic, for a total of 4 runs of 8 teams. In each run, the results of best and worst teams are discarded, and my final score also discards the best and worst runs during calculation. That should give a better grasp of what tactic is expected to perform better, despite the variance and not running as many tests as FM-Arena.
so autumn 0.1 is the best?
max 737 said: so autumn 0.1 is the best?
It got the best results in the experiments, for that configuration.
Just leaving Summer 0.11 here for anyone interested. It is better than 0.1, but still tentative, so it might change until the 0.2 version.
P.S.: My next batch of experiments is about different roles for OoP, like pressing or tracking.
Last update of the seasons before testing OoP stuff is below, Winter 0.11. It is tentative, but better than 0.1.
Have you tried testing the 'extreme' offensive (most GS) and defensive (least GA) tactics from FM24 as a IP/OOP pair in FM26? I'm curious to see if the IP/OOP system can handle these extremes within the same 'tactic'.
johnconnerson said: Have you tried testing the 'extreme' offensive (most GS) and defensive (least GA) tactics from FM24 as a IP/OOP pair in FM26? I'm curious to see if the IP/OOP system can handle these extremes within the same 'tactic'.
I will investigate in the future. Thank you for the feedback.
Hello, thank you for all these tactics. I would like to know when we will have access to the table section of FM 26.
Cyclingstrategist said: Hello, thank you for all these tactics. I would like to know when we will have access to the table section of FM 26.
we dont yet know unfortunately
Cyclingstrategist said: Hello, thank you for all these tactics. I would like to know when we will have access to the table section of FM 26.
Soon.
I predict at the end of this week we'll start testing tactics for FM26.
Zippo said: Soon.
I predict at the end of this week we'll start testing tactics for FM26.
Great News, can't wait