"The Winter Patch" just dropped and as you may already know it doesn't have any M.E. changes.
Anyway, we have updated our tactic testing league and soon we'll start retesting tactics on the updated DB.
We implemented some tweaks to the tactic testing league that greatly reduce the RNG in the league, these tweaks reduce the RNG about 2 times. Also, we made slight tweaking to the players' attributes and AI managers' tactical settings.
Soon, after our testing algorithm finishes retesting the tactics from the "Hall of Fame" under the new DB, you can manually request retesting your tactic under the new DB.
Please note, that the "retesting requests" share the same cooldown with the regular tactic sharing so per 24 hours you can only make 2 retesting request or share 2 new tactics or a combination of both.
Will you still be testing the tactics that were submitted in the last 24 hours? will these be tested after the Hall of Fame tactics?
Thanks Expand
The priority will be given to the tactics from the HOF, probably, not all of them will be retested from the start but obviously, a bunch of top tactics from the top of HOF should be retested at first then next, probably, a mix of newly uploaded tactics and retesting requests.
Wow, some massive points boost for a few tactics there...I know there haven't been any ME changes but I wonder if balanced/positive mentality has received some sort of hidden boost
CBP87 said: Wow, some massive points boost for a few tactics there...I know there haven't been any ME changes but I wonder if balanced/positive mentality has received some sort of hidden boost
Expand
Probably these boosts are due to attributes changes in the database’s players, I think there’s anything about 24.3.0 patch
Gianaa9 said: Probably these boosts are due to attributes changes in the database’s players, I think there’s anything about 24.3.0 patch Expand
I would assume the attribute changes would be generic across all roles, so maybe 13 across the board for all attributes rather than certain increases in certain attributes which would benefit certain roles
We've updated the testing league and made changes that increase the difficulty in the league, now tactics should be more efficient to come on top and also, the changes reduce the RNG which improves the accuracy of the results.
We've updated the testing league and made changes that increase the difficulty in the league, now tactics should be more efficient to come on top and also, the changes reduce the RNG which improves the accuracy of the results. Expand
We've updated the testing league and made changes that increase the difficulty in the league, now tactics should be more efficient to come on top and also, the changes reduce the RNG which improves the accuracy of the results. Expand
First of all, thank you for your hard work. However, I don't think RNG is any different, this can be seen by the upper and lower limits of the score presented, still generally between 2 and 3, while the more difficult score and the number of tests also offset each other's bonus. There were also nearly identical tactical test results, such as the Problem Solvin V18 and the 424 Classic CBP87 v1.8. RNG is closely related to the number of test venues, and the effectiveness of these three tactical test reforms is only regressive in terms of RNG. I think the accuracy of the test results still needs more test matches to improve, and the tactics with higher scores need to unify the test matches, whether it is 4800 or 9600. Even that doesn't really tell the difference between two nearly identical tactics.
Results will be the same regardless of the difficulty and I think every time we (re)test them we will have a different "winner" because all the tactics are close.
Just wondering if FM-Arena introduces the 4222 and 5221 formations to the AI teams to see if there is any change in the results.
A Smile said: First of all, thank you for your hard work. However, I don't think RNG is any different, this can be seen by the upper and lower limits of the score presented, still generally between 2 and 3, while the more difficult score and the number of tests also offset each other's bonus. There were also nearly identical tactical test results, such as the Problem Solvin V18 and the 424 Classic CBP87 v1.8. RNG is closely related to the number of test venues, and the effectiveness of these three tactical test reforms is only regressive in terms of RNG. I think the accuracy of the test results still needs more test matches to improve, and the tactics with higher scores need to unify the test matches, whether it is 4800 or 9600. Even that doesn't really tell the difference between two nearly identical tactics. Expand
Perhaps they haven't updated the upper and lower limits yet. I doubt they would've restarted if the RNG wasn't actually reduced by the new testing league.
dzek said: Results will be the same regardless of the difficulty and I think every time we (re)test them we will have a different "winner" because all the tactics are close.
Just wondering if FM-Arena introduces the 4222 and 5221 formations to the AI teams to see if there is any change in the results. Expand
I'm not sure if results would be the same regardless of difficulty. Are some tactics not better against harder teams than they are against weaker teams, relative to other tactics?
Cherknam said: I'm not sure if results would be the same regardless of difficulty. Are some tactics not better against harder teams than they are against weaker teams, relative to other tactics? Expand When I say the same, I don't mean exactly the same. Of course the scores will drop but they will be more or less the same. That is my opinion and it just remains to be seen. Maybe I am wrong and there is no problem.
As for the other tactics I am suggesting I would like to see at least one test including these two tactics. I may be wrong again. There is no problem at all.
If you don't mind me asking @Zippo what is it that you've changed to make the test league more difficult, some of these point swings are enormous. It's like you've created a whole new test league
Guys, I just want to let you know that the switching to the new DB 3.0 didn't go smoothly and few small bugs managed to lurk in but don't worry, we already have found and fixed them.
It's just don't be surprised to see some tactics that already have been tested under the new DB 3.0 get retested once more, we retest them because we want to be sure the result aren't affected by the small bugs that I mentioned.
Can you explain what you've changed please? I'm shocked at how the previous top tactics are doing in the new testing DB. What did you change that resulted in such a drastic swing in the previous top tactics?
Nah, I wouldn't call it "a drastic swing". It's just about 6 - 9 points drop across the board, which translates into 8%-12% points drop in points. No way 8%-12% points drop can be considered as "a drastic swing", to define it as "a drastic drop" it should not be no less than 50% or more.
In general, if we take a regular season that consists of 38 matches then 6 - 9 points drop means that you get just 2 or 3 wins less than before.
So in the new DB the AI managers show a bit more resistance which translates into 9%-12% points drops for the human managers and that's all.
More importantly, in the new DB the balance between tactics has stayed 99% the same as it was before.
CBP87 said: What did you change that resulted in such a drastic swing in the previous top tactics? Expand
Some settings that are responsible for the way AI managers play. Sorry, I can't be more specific here.
Germaniac said: will the results still be representative of a normal gameplay? Expand No, of course not... we've been working hard on our testing DB to make it less and less representative of normal gameplay because at the end our goal is finding good tactics to play FIFA and PES and not FM.
Pfff... it would be ridiculous to assume that we've been updating our DB to make it better for finding good tactics to play FM! The less our DB represents normal gameplay, the better for us! Isn't it obvious?
Germaniac said: Why can't you be more specific? Expand I can only take a wild guess here... it could be that we want to keep the result of our hard work on learning the FM privately for ourselves? Yes, I know it's unbelievable selfishness but unfortunately, we're that evil. Once more, it's just my wild guess.
Zippo said: No, of course not... we've been working hard on our testing DB to make it less and less representative of normal gameplay because at the end our goal is finding good tactics to play FIFA and PES and not FM.
Pfff... it would be ridiculous to assume that we've been updating our DB to make it better for finding good tactics to play FM! The less our DB represents normal gameplay, the better for us! Isn't it obvious?
I can only take a wild guess here... it could be that we want to keep the result of our hard work on learning the FM privately for ourselves? Yes, I know it's unbelievable selfishness but unfortunately, we're that evil. Once more, it's just my wild guess. Expand
His question may have been worded rather rudely, but do you think it’s helpful to reply like that? Why not just give a straightforward answer and if you need to be vague to protect your work then say that.
Not sure if this should go here, but it seems that GD on the table is calculated from the GF-GA after each of them are rounded? If it's true then it can be less accurate than rounding the GD after you calculate GF-GA first. For example, 64.4-47.6 will be rounded (to nearest integer) to 64-48=16, but if you calculate 64.4-47.6=16.8 first it'll be rounded to 17.
Germaniac said: Not sure if this should go here, but it seems that GD on the table is calculated from the GF-GA after each of them are rounded? If it's true then it can be less accurate than rounding the GD after you calculate GF-GA first. For example, 64.9-48.1 will be rounded (to nearest integer) to 64-48=16, but if you calculate 64.9-48.1=16.8 first it'll be rounded to 17. Expand
Hi,
We are aware of such rounding might happen but still we think that adding decimal numbers to G.F., G.A. and G.D. would be "overkill" because when in the testing the minimal RNG is +/- 1-2 points when it comes to Pts and +/- 2-3 G.D. when it comes to G.D. then having decimal numbers for G.D. doesn't make sense. Yes, we added decimal numbers for Pts but still we aren't 100% sure they are needed even for Pts.
We are aware of such rounding might happen but still we think that adding decimal numbers to G.F., G.A. and G.D. would be "overkill" because when in the testing the minimal RNG is +/- 1-2 points when it comes to Pts and +/- 2-3 G.D. when it comes to G.D. then having decimal numbers for G.D. doesn't make sense. Yes, we added decimal numbers for Pts but still we aren't 100% sure they are needed even for Pts. Expand You don't need to add decimal numbers, just round GD after calculating GF-GA.
The biggest problem in my opinion, is sorting the table by GD will be inaccurate. For example:
Tactic A: 64.4-47.6=16.8 on the current table will be 64-48=16 Tactic B: 64.6-48.4=16.2 on the current table will be 65-48=17
As you can see Tactic A actually has higher GD than Tactic B, but on the current table when you sort the tactics by GD, Tactic B will be higher than Tactic A.
While I agree that a difference of 0.6 is insignificant due to the RNG, I think the order shouldn't be reversed. In extreme cases like the example above, people will think Tactic B has better GD by around 1, when it's actually worse by 0.6, a difference of 1.6.
Germaniac said: The biggest problem in my opinion, is sorting the table by GD will be inaccurate. For example: Expand First of all, we always can sort tactics in the tables on "raw" values(not rounded), probably, that happens right now but I'm not 100% sure about that. Anyway, it's really isn't a big deal taking into consideration the RNG.
Germaniac said: For example, 64.4-47.6 will be rounded (to nearest integer) to 64-48=16, but if you calculate 64.4-47.6=16.8 first it'll be rounded to 17. Expand
We had the tables the way you say before but it was quite confusing. For example, people were looking at a table and seeing that GF = 64 and GA = 48 so сertainly the GD should be equal to "16" in this case because 64-48=16 but somehow it was "17" in the table so they were reporting to us about a "wrong" calculation and so on. So after some consideration we decided that the current approach is better.
I can recall that in previous iterations of the test leagues, the AI teams had names like "Best 4-2-3-1 AI tactic" or something of that sorts. When testing this tactic:
It seems to struggle quite a lot versus AI Team 1, while being very good against the others. I'm now wondering which tactic this specific AI is using. I was not successful searching the site for this info, can someone (I guess @Zippo?) tell me which one that is? Also, is this test db available somewhere? I would love to get a three-at-the-back system in the 4-stars-range
Hi,
"The Winter Patch" just dropped and as you may already know it doesn't have any M.E. changes.
Anyway, we have updated our tactic testing league and soon we'll start retesting tactics on the updated DB.
We implemented some tweaks to the tactic testing league that greatly reduce the RNG in the league, these tweaks reduce the RNG about 2 times. Also, we made slight tweaking to the players' attributes and AI managers' tactical settings.
Soon, after our testing algorithm finishes retesting the tactics from the "Hall of Fame" under the new DB, you can manually request retesting your tactic under the new DB.
Please note, that the "retesting requests" share the same cooldown with the regular tactic sharing so per 24 hours you can only make 2 retesting request or share 2 new tactics or a combination of both.
Cheers.
Thanks @Zippo
Will you still be testing the tactics that were submitted in the last 24 hours? will these be tested after the Hall of Fame tactics?
Thanks
CBP87 said: Thanks @Zippo
Will you still be testing the tactics that were submitted in the last 24 hours? will these be tested after the Hall of Fame tactics?
Thanks
The priority will be given to the tactics from the HOF, probably, not all of them will be retested from the start but obviously, a bunch of top tactics from the top of HOF should be retested at first then next, probably, a mix of newly uploaded tactics and retesting requests.
Now, we just wait until FMRTE gets updated.
Wow, some massive points boost for a few tactics there...I know there haven't been any ME changes but I wonder if balanced/positive mentality has received some sort of hidden boost
CBP87 said: Wow, some massive points boost for a few tactics there...I know there haven't been any ME changes but I wonder if balanced/positive mentality has received some sort of hidden boost
Probably these boosts are due to attributes changes in the database’s players, I think there’s anything about 24.3.0 patch
Gianaa9 said: Probably these boosts are due to attributes changes in the database’s players, I think there’s anything about 24.3.0 patch
I would assume the attribute changes would be generic across all roles, so maybe 13 across the board for all attributes rather than certain increases in certain attributes which would benefit certain roles
Gianaa9 said: Probably these boosts are due to attributes changes in the database’s players, I think there’s anything about 24.3.0 patch
It's the power of RNG. 2,400 matches give enough room for such increases.
Meriten said: It's the power of RNG. 2,400 matches give enough room for such increases.
Barely, and unlikely. Some differences are huge.
what changed in new database?
smigler said: what changed in new database?
Hi,
We've updated the testing league and made changes that increase the difficulty in the league, now tactics should be more efficient to come on top and also, the changes reduce the RNG which improves the accuracy of the results.
Zippo said: Hi,
We've updated the testing league and made changes that increase the difficulty in the league, now tactics should be more efficient to come on top and also, the changes reduce the RNG which improves the accuracy of the results.
Thanks for your hard work on this.
Zippo said: Hi,
We've updated the testing league and made changes that increase the difficulty in the league, now tactics should be more efficient to come on top and also, the changes reduce the RNG which improves the accuracy of the results.
First of all, thank you for your hard work. However, I don't think RNG is any different, this can be seen by the upper and lower limits of the score presented, still generally between 2 and 3, while the more difficult score and the number of tests also offset each other's bonus. There were also nearly identical tactical test results, such as the Problem Solvin V18 and the 424 Classic CBP87 v1.8.
RNG is closely related to the number of test venues, and the effectiveness of these three tactical test reforms is only regressive in terms of RNG. I think the accuracy of the test results still needs more test matches to improve, and the tactics with higher scores need to unify the test matches, whether it is 4800 or 9600. Even that doesn't really tell the difference between two nearly identical tactics.
Results will be the same regardless of the difficulty and I think every time we (re)test them we will have a different "winner" because all the tactics are close.
Just wondering if FM-Arena introduces the 4222 and 5221 formations to the AI teams to see if there is any change in the results.
A Smile said: First of all, thank you for your hard work. However, I don't think RNG is any different, this can be seen by the upper and lower limits of the score presented, still generally between 2 and 3, while the more difficult score and the number of tests also offset each other's bonus. There were also nearly identical tactical test results, such as the Problem Solvin V18 and the 424 Classic CBP87 v1.8.
RNG is closely related to the number of test venues, and the effectiveness of these three tactical test reforms is only regressive in terms of RNG. I think the accuracy of the test results still needs more test matches to improve, and the tactics with higher scores need to unify the test matches, whether it is 4800 or 9600. Even that doesn't really tell the difference between two nearly identical tactics.
Perhaps they haven't updated the upper and lower limits yet. I doubt they would've restarted if the RNG wasn't actually reduced by the new testing league.
dzek said: Results will be the same regardless of the difficulty and I think every time we (re)test them we will have a different "winner" because all the tactics are close.
Just wondering if FM-Arena introduces the 4222 and 5221 formations to the AI teams to see if there is any change in the results.
I'm not sure if results would be the same regardless of difficulty. Are some tactics not better against harder teams than they are against weaker teams, relative to other tactics?
Cherknam said: I'm not sure if results would be the same regardless of difficulty. Are some tactics not better against harder teams than they are against weaker teams, relative to other tactics?
When I say the same, I don't mean exactly the same. Of course the scores will drop but they will be more or less the same. That is my opinion and it just remains to be seen. Maybe I am wrong and there is no problem.
As for the other tactics I am suggesting I would like to see at least one test including these two tactics. I may be wrong again. There is no problem at all.
If you don't mind me asking @Zippo what is it that you've changed to make the test league more difficult, some of these point swings are enormous. It's like you've created a whole new test league
Guys, I just want to let you know that the switching to the new DB 3.0 didn't go smoothly and few small bugs managed to lurk in but don't worry, we already have found and fixed them.
It's just don't be surprised to see some tactics that already have been tested under the new DB 3.0 get retested once more, we retest them because we want to be sure the result aren't affected by the small bugs that I mentioned.
Can you explain what you've changed please? I'm shocked at how the previous top tactics are doing in the new testing DB. What did you change that resulted in such a drastic swing in the previous top tactics?
CBP87 said: What did you change that resulted in such a drastic swing
Hi, @CBP87.
Nah, I wouldn't call it "a drastic swing". It's just about 6 - 9 points drop across the board, which translates into 8%-12% points drop in points. No way 8%-12% points drop can be considered as "a drastic swing", to define it as "a drastic drop" it should not be no less than 50% or more.
In general, if we take a regular season that consists of 38 matches then 6 - 9 points drop means that you get just 2 or 3 wins less than before.
So in the new DB the AI managers show a bit more resistance which translates into 9%-12% points drops for the human managers and that's all.
More importantly, in the new DB the balance between tactics has stayed 99% the same as it was before.
CBP87 said: What did you change that resulted in such a drastic swing in the previous top tactics?
Some settings that are responsible for the way AI managers play. Sorry, I can't be more specific here.
I hope that helps.
Cheers.
Zippo said: Some settings that are responsible for the way AI managers play. Sorry, I can't be more specific here.
Why can't you be more specific? Sorry, but it sounds rather suspicious.
Also, if you're tinkering with the way AI managers play, will the results still be representative of a normal gameplay?
Germaniac said: will the results still be representative of a normal gameplay?
No, of course not... we've been working hard on our testing DB to make it less and less representative of normal gameplay because at the end our goal is finding good tactics to play FIFA and PES and not FM.
Pfff... it would be ridiculous to assume that we've been updating our DB to make it better for finding good tactics to play FM! The less our DB represents normal gameplay, the better for us! Isn't it obvious?
Germaniac said: Why can't you be more specific?
I can only take a wild guess here... it could be that we want to keep the result of our hard work on learning the FM privately for ourselves? Yes, I know it's unbelievable selfishness but unfortunately, we're that evil. Once more, it's just my wild guess.
Zippo said: No, of course not... we've been working hard on our testing DB to make it less and less representative of normal gameplay because at the end our goal is finding good tactics to play FIFA and PES and not FM.
Pfff... it would be ridiculous to assume that we've been updating our DB to make it better for finding good tactics to play FM! The less our DB represents normal gameplay, the better for us! Isn't it obvious?
I can only take a wild guess here... it could be that we want to keep the result of our hard work on learning the FM privately for ourselves? Yes, I know it's unbelievable selfishness but unfortunately, we're that evil. Once more, it's just my wild guess.
His question may have been worded rather rudely, but do you think it’s helpful to reply like that? Why not just give a straightforward answer and if you need to be vague to protect your work then say that.
Not sure if this should go here, but it seems that GD on the table is calculated from the GF-GA after each of them are rounded? If it's true then it can be less accurate than rounding the GD after you calculate GF-GA first. For example, 64.4-47.6 will be rounded (to nearest integer) to 64-48=16, but if you calculate 64.4-47.6=16.8 first it'll be rounded to 17.
Germaniac said: Not sure if this should go here, but it seems that GD on the table is calculated from the GF-GA after each of them are rounded? If it's true then it can be less accurate than rounding the GD after you calculate GF-GA first. For example, 64.9-48.1 will be rounded (to nearest integer) to 64-48=16, but if you calculate 64.9-48.1=16.8 first it'll be rounded to 17.
Hi,
We are aware of such rounding might happen but still we think that adding decimal numbers to G.F., G.A. and G.D. would be "overkill" because when in the testing the minimal RNG is +/- 1-2 points when it comes to Pts and +/- 2-3 G.D. when it comes to G.D. then having decimal numbers for G.D. doesn't make sense. Yes, we added decimal numbers for Pts but still we aren't 100% sure they are needed even for Pts.
Droid said: Hi,
We are aware of such rounding might happen but still we think that adding decimal numbers to G.F., G.A. and G.D. would be "overkill" because when in the testing the minimal RNG is +/- 1-2 points when it comes to Pts and +/- 2-3 G.D. when it comes to G.D. then having decimal numbers for G.D. doesn't make sense. Yes, we added decimal numbers for Pts but still we aren't 100% sure they are needed even for Pts.
You don't need to add decimal numbers, just round GD after calculating GF-GA.
The biggest problem in my opinion, is sorting the table by GD will be inaccurate. For example:
Tactic A: 64.4-47.6=16.8 on the current table will be 64-48=16
Tactic B: 64.6-48.4=16.2 on the current table will be 65-48=17
As you can see Tactic A actually has higher GD than Tactic B, but on the current table when you sort the tactics by GD, Tactic B will be higher than Tactic A.
While I agree that a difference of 0.6 is insignificant due to the RNG, I think the order shouldn't be reversed. In extreme cases like the example above, people will think Tactic B has better GD by around 1, when it's actually worse by 0.6, a difference of 1.6.
Germaniac said: The biggest problem in my opinion, is sorting the table by GD will be inaccurate. For example:
First of all, we always can sort tactics in the tables on "raw" values(not rounded), probably, that happens right now but I'm not 100% sure about that. Anyway, it's really isn't a big deal taking into consideration the RNG.
Germaniac said: just round GD after calculating GF-GA.
Germaniac said: For example, 64.4-47.6 will be rounded (to nearest integer) to 64-48=16, but if you calculate 64.4-47.6=16.8 first it'll be rounded to 17.
We had the tables the way you say before but it was quite confusing. For example, people were looking at a table and seeing that GF = 64 and GA = 48 so сertainly the GD should be equal to "16" in this case because 64-48=16 but somehow it was "17" in the table so they were reporting to us about a "wrong" calculation and so on. So after some consideration we decided that the current approach is better.
I can recall that in previous iterations of the test leagues, the AI teams had names like "Best 4-2-3-1 AI tactic" or something of that sorts. When testing this tactic:
https://fm-arena.com/thread/12012-svonnstem-3-4-3-libero-mid-focus/
It seems to struggle quite a lot versus AI Team 1, while being very good against the others. I'm now wondering which tactic this specific AI is using. I was not successful searching the site for this info, can someone (I guess @Zippo?) tell me which one that is? Also, is this test db available somewhere? I would love to get a three-at-the-back system in the 4-stars-range