Oh my god! The new patch ruined the tactics of previous tests. Hopefully the tactical masters (@ZaZ , @Magician , @Zealot , @Cyborg , @Egraam , and several others) can present us with good tactics. Good job on the tests, @Zippo.
Mark said: @Zippo are you testing all tactics over the 6 tests or just the leading tactic? Expand
For now all tactics rated 5.1 or higher should get 6 runs ( 912 matches ) but at some point in the future all tactics rated 4.8 or higher should get 6 runs ( 912 matches ) too.
Once again I think you are doing awesome work. Do you have any idea what’s causing quite some fluctuation between the same tests ? Take Tika taka for example , in a test their b team is almost relegated while in the test after it , it’s 2nd after team A. Given that all circumstances are equal , I don’t expect quite such a large difference in results ( 50 points ? ) I’m not a data analyst person so I don’t really have an answer to this.
Belmorn said: Once again I think you are doing awesome work. Do you have any idea what’s causing quite some fluctuation between the same tests ? Take Tika taka for example , in a test their b team is almost relegated while in the test after it , it’s 2nd after team A. Given that all circumstances are equal , I don’t expect quite such a large difference in results ( 50 points ? ) I’m not a data analyst person so I don’t really have an answer to this. Expand
Belmorn said: I didnt look it up right, its more like 30 points difference which still is quite large ?
Positive Tiki Taka - Team B test 3 -> 97 points Other tests its mostly close to 125 ish points Expand
So the highest difference is about 30 points on 912 matches distance and I find it's a quite small difference because when I test with Norwich in the real game I can have 40 points difference on 38 matches distance.
If you think that 30 points is quite high on 912 matches distance then make your own test in the real game and post your results. For instance, test it in EPL but don't pick the strongest teams in the league like Liverpool, Man Utd and so on because with that teams almost any tactics will give you 95-100 points, pick an average team like Everton or Leicester and play 912 matches ( 24 seasons ) and post your results and we'll see what will be the highest difference on 38 matches distance after 24 seasons.
Last year, the top tactic went from 178 to 145 points, or a 33 points difference. We are talking about the most stable tactic from last version, with 72% win rate. Now, the top tactic has only 54% win rate, also because of the new testing league that uses better tactics. Lower win rate means higher variation.
He tested only for 228 matches(6 seasons) and the highest difference was 32 points on 38 matches distance!!!
Test 1 ( 31 pts ) vs Test 4 ( 63 points ) so on 38 matches distance it was 32 points then if we translate it into 76 matches then it'll be 64 points difference!! And you say that 30 points is quite large?
Note that he tested only for 6 season instead of 24 seasons and I wonder what the biggest difference was if he tested for 24 seasons.
ZaZ said: also because of the new testing league that uses better tactics. Expand
I was just to ask this question when you posted this, if any admin could answer me would be great, since the test got harder (I think) what do you guess will be the highest score possible ? because I don't think we will be getting close to the 7´s from last year
Just to be clear, that variance between tests is expected. That's why they do six tests, to give higher confidence to results. With just one or two tests, you could claim one was a lucky shot or another was unlucky. If they could, they would test thirty times each, but that is not viable because of time constraints.
@Rince i somehow get the feeling everything I said is taken as an assault on the testing league which it clearly isn’t. If I don’t think it’s good / the best I wouldn’t even bother interacting. It clearly is the best on the net. It’s current iteration with the ai being further buffed AND the current ME just shows to me that if you want to win consistently then you have to have everything in order like morale / cohesion / tactical familiarity ( which you also posted elsewhere on this forum ). The attribute testing league clearly shows which attributes make a far bigger impact than any decent tactic will make. My current squad fe is around 110 Ca but challenging for German title because it’s loaded with pace / accel / stam / natural fitness / consistency / important matches and less on most mental / technical attribites.
@Sandro thank you for that link , team cohesion not being editable / freezeable can definitely contribute to slightly off expected results , as well as the obvious rng is rng or better said on a forum like this : fm is fm.
Belmorn said: @Sandro thank you for that link , team cohesion not being editable / freezeable can definitely contribute to slightly off expected results , as well as the obvious rng is rng or better said on a forum like this : fm is fm. Expand
Obviously, Team Cohesion has some influence but I don't think it's responsible for the results' variation we see in the test league. If you remove all the RNG factors from the game then it still will be that random because the game is that random and you can't do anything about it, you can just slight reduce the RNG by eliminating as many as possible random factors but still it'll quite random because the developers from SI want the game to be like that.
BiTeL33T said: I was just to ask this question when you posted this, if any admin could answer me would be great, since the test got harder (I think) what do you guess will be the highest score possible ? because I don't think we will be getting close to the 7´s from last year
at least in this match engine anyway Expand
Not an admin, but I think people are just taken aback by the difficulty of the testing league. It has gotten so much tougher with the improved opponents, so I think the 5.X from this year is equal to 6.X from last year or so. Currently my 5.3* Positive Tiki Taka tactic and @ZaZ Blue DM are winning leagues with mid table teams consistently, so I think that's clearly proof that the tactics are still good, just that the FM-Arena team has been constantly improving the Testing League and it has gotten a lot more exigent.
ta2199 said: Hi, just wonder can you test the affect of jadeness to performance ? @Zippo Expand
Hi,
As far as I know "Jadedness" is a quite simple thing. A high "Jadedness" decreases the rate at which players restore their conditions, it also might be that a high "Jadedness" increases the rate at which players lose their conditions during matches but I'm not sure about that. Unfortunately, it isn't that simple to test it.
As far as I know "Jadedness" is a quite simple thing. A high "Jadedness" decreases the rate at which players restore their conditions, it also might be that a high "Jadedness" increases the rate at which players lose their conditions during matches but I'm not sure about that. Unfortunately, it isn't that simple to test it. Expand
I remember somewhere said that high jadeness make player perform worse. So I think It would worth for a try.
Oh my god! The new patch ruined the tactics of previous tests. Hopefully the tactical masters (@ZaZ , @Magician , @Zealot , @Cyborg , @Egraam , and several others) can present us with good tactics. Good job on the tests, @Zippo.
Would it be possible to share with us the database used for the tactics tests? I would like to run some tests as well, if possible.
I just want to thank you for all the work you do (is this the right place for it???)
@Zippo is the new testing table for patch 2.0 as per the new database *
Tejash said: @Zippo is the new testing table for patch 2.0 as per the new database *
The patch 22.2.0 results on the new database.
@Zippo are you testing all tactics over the 6 tests or just the leading tactic?
Mark said: @Zippo are you testing all tactics over the 6 tests or just the leading tactic?
For now all tactics rated 5.1 or higher should get 6 runs ( 912 matches ) but at some point in the future all tactics rated 4.8 or higher should get 6 runs ( 912 matches ) too.
Once again I think you are doing awesome work.
Do you have any idea what’s causing quite some fluctuation between the same tests ?
Take Tika taka for example , in a test their b team is almost relegated while in the test after it , it’s 2nd after team A.
Given that all circumstances are equal , I don’t expect quite such a large difference in results ( 50 points ? )
I’m not a data analyst person so I don’t really have an answer to this.
Belmorn said: I don’t expect quite such a large difference in results ( 50 points ? )
Where is that 50 points difference ?
Belmorn said: Once again I think you are doing awesome work.
Do you have any idea what’s causing quite some fluctuation between the same tests ?
Take Tika taka for example , in a test their b team is almost relegated while in the test after it , it’s 2nd after team A.
Given that all circumstances are equal , I don’t expect quite such a large difference in results ( 50 points ? )
I’m not a data analyst person so I don’t really have an answer to this.
Look at this - https://fm-arena.com/thread/442-my-attempt-to-measure-the-game-rng-random-number-generator/
Rince said: Where is that 50 points difference ?
I didnt look it up right, its more like 30 points difference which still is quite large ?
Positive Tiki Taka - Team B test 3 -> 97 points
Other tests its mostly close to 125 ish points
Belmorn said: I didnt look it up right, its more like 30 points difference which still is quite large ?
Positive Tiki Taka - Team B test 3 -> 97 points
Other tests its mostly close to 125 ish points
So the highest difference is about 30 points on 912 matches distance and I find it's a quite small difference because when I test with Norwich in the real game I can have 40 points difference on 38 matches distance.
If you think that 30 points is quite high on 912 matches distance then make your own test in the real game and post your results. For instance, test it in EPL but don't pick the strongest teams in the league like Liverpool, Man Utd and so on because with that teams almost any tactics will give you 95-100 points, pick an average team like Everton or Leicester and play 912 matches ( 24 seasons ) and post your results and we'll see what will be the highest difference on 38 matches distance after 24 seasons.
Last year, the top tactic went from 178 to 145 points, or a 33 points difference. We are talking about the most stable tactic from last version, with 72% win rate. Now, the top tactic has only 54% win rate, also because of the new testing league that uses better tactics. Lower win rate means higher variation.
Belmorn said: I didnt look it up right, its more like 30 points difference which still is quite large ?
Some guy tested Revelation in EPL with Watford and posted his results in the thread - https://fm-arena.com/find-comment/9280/
He tested only for 228 matches(6 seasons) and the highest difference was 32 points on 38 matches distance!!!
Test 1 ( 31 pts ) vs Test 4 ( 63 points ) so on 38 matches distance it was 32 points then if we translate it into 76 matches then it'll be 64 points difference!! And you say that 30 points is quite large?
Note that he tested only for 6 season instead of 24 seasons and I wonder what the biggest difference was if he tested for 24 seasons.
ZaZ said: also because of the new testing league that uses better tactics.
I was just to ask this question when you posted this, if any admin could answer me would be great, since the test got harder (I think) what do you guess will be the highest score possible ? because I don't think we will be getting close to the 7´s from last year
at least in this match engine anyway
Just to be clear, that variance between tests is expected. That's why they do six tests, to give higher confidence to results. With just one or two tests, you could claim one was a lucky shot or another was unlucky. If they could, they would test thirty times each, but that is not viable because of time constraints.
@Rince i somehow get the feeling everything I said is taken as an assault on the testing league which it clearly isn’t. If I don’t think it’s good / the best I wouldn’t even bother interacting. It clearly is the best on the net. It’s current iteration with the ai being further buffed AND the current ME just shows to me that if you want to win consistently then you have to have everything in order like morale / cohesion / tactical familiarity ( which you also posted elsewhere on this forum ). The attribute testing league clearly shows which attributes make a far bigger impact than any decent tactic will make. My current squad fe is around 110 Ca but challenging for German title because it’s loaded with pace / accel / stam / natural fitness / consistency / important matches and less on most mental / technical attribites.
@Sandro thank you for that link , team cohesion not being editable / freezeable can definitely contribute to slightly off expected results , as well as the obvious rng is rng or better said on a forum like this : fm is fm.
Belmorn said: @Sandro thank you for that link , team cohesion not being editable / freezeable can definitely contribute to slightly off expected results , as well as the obvious rng is rng or better said on a forum like this : fm is fm.
Obviously, Team Cohesion has some influence but I don't think it's responsible for the results' variation we see in the test league. If you remove all the RNG factors from the game then it still will be that random because the game is that random and you can't do anything about it, you can just slight reduce the RNG by eliminating as many as possible random factors but still it'll quite random because the developers from SI want the game to be like that.
BiTeL33T said: I was just to ask this question when you posted this, if any admin could answer me would be great, since the test got harder (I think) what do you guess will be the highest score possible ? because I don't think we will be getting close to the 7´s from last year
at least in this match engine anyway
Not an admin, but I think people are just taken aback by the difficulty of the testing league. It has gotten so much tougher with the improved opponents, so I think the 5.X from this year is equal to 6.X from last year or so. Currently my 5.3* Positive Tiki Taka tactic and @ZaZ Blue DM are winning leagues with mid table teams consistently, so I think that's clearly proof that the tactics are still good, just that the FM-Arena team has been constantly improving the Testing League and it has gotten a lot more exigent.
Zippo said: The patch 22.2.0 results on the new database.
Thats explain why the score is so low
@cptdoggo yeah i use Positive Tiki Taka myself and results are crazy, very good work, congrats
Hi, just wonder can you test the affect of jadeness to performance ? @Zippo
ta2199 said: Hi, just wonder can you test the affect of jadeness to performance ? @Zippo
Hi,
As far as I know "Jadedness" is a quite simple thing. A high "Jadedness" decreases the rate at which players restore their conditions, it also might be that a high "Jadedness" increases the rate at which players lose their conditions during matches but I'm not sure about that. Unfortunately, it isn't that simple to test it.
Zippo said: Hi,
As far as I know "Jadedness" is a quite simple thing. A high "Jadedness" decreases the rate at which players restore their conditions, it also might be that a high "Jadedness" increases the rate at which players lose their conditions during matches but I'm not sure about that. Unfortunately, it isn't that simple to test it.
I remember somewhere said that high jadeness make player perform worse. So I think It would worth for a try.
ta2199 said: I remember somewhere said that high jadeness make player perform worse. So I think It would worth for a try.
I'll look into it. If we can freeze the "Jadedness" then we'll test it but I'm not sure it can be "frozen".
I always thought jadedness was fatigue. =)
ZaZ said: I always thought jadedness was fatigue. =)
SI member said Jadeness is related to natural fitness. So I might got something to do with the player condition