A quick update. First round of testing against the top 2 tactics, ZaZ Fire 0.3 moved to the top and Perfect moved 2nd, just. It was the only time I had the 4 teams finish top 4 in all 4 tests. In fact the previous record was just 2 out of the 4. These are very competitive tactics, with @ZaZ looking the best. I will do another test tomorrow and see if this early form holds up.
Hi All. I thought I would have a bit of a play with some of the tactics that others are doing well with on this site as well as a few that have high downloads off FMBase. Ended up with 9 tactics and then added 3 of my old ones to see how they stack up. I divided the tactics into 3 Groups of 4 and then cycled them through 4 teams in Vanarama National South. So that meant 4 tests for each group so they all used each of the 4 teams. The 4 teams selected were ranked as the bottom 4 by Genie Scout in that Division and the season preview in FM23 ranked them 16 to 19th. So all pretty close and not rated much chance - St. Albans, Braintree, Oxford City and Hungerford.
Following the first round of testing I then seeded them and reran the tests. So I now have 368 games for each tactic. Results are below. The next step is running the top 2 tactics against the new ZaZ Fire 0.3 and Perfect by super korean to see how those two stack up. To make it fair the second run will be against mid table tactics.
They were all pretty good considering my expectations for these teams. The top 4 tactics were outstanding with not much between them through the whole process.
OpticFawn said: @ZaZ whenever I sign youngsters, they barely improve with your training session, one of my players increased who was aged 18 at the time, +1 in a few attributes like 4 years, what am I doing wrong?
P.s. I setup the rest and everything. Expand
FM Arena did some work on this and advised us earlier, in this thread at post #17 on page one, of some of their findings. See link below.
I dont quite understand how you can find value. Why wouldn’t you just use positional rating and buy the highest rated player you can afford? Expand
Sure. I looked at 2 tables, the Cost of Attribute Table by position and the ykykyky positional rating weightings. I only looked at attributes that were 50 or higher in the ykykyky positional rating weightings and then divided this score by the corresponding value on the Cost of Attribute Table. This then gave a value rating, higher the better.
My expectation of this score is that if a player has high values for these value scores their rating will be better value than those that are higher in the high cost attributes. I use these in the filters and it seems to be working, although I haven't done comparisons against those with that are higher in the high cost attributes. And I always include acceleration in my filters, other than for DM or GK.
Middleweight165 said: Hey @Mark Does the positional rating on GS not take into account the position value or is it just inaccurate? Why is your calculation necessary?
I'm looking at a player who has a position value of 1 and a positional rating of 66.9%, so using your calculation, his true positional value is 39.3%. Is that correct? Expand
Your calculation is correct. Of course you can improve your players position value and therefore their positional rating by training them and playing them in the position.
GS have a very minor adjustment for positional value, it is around 0.35% per point. My values are derived from the FM Arena testing over the last 2 years, and I believe they are much closer to the mark.
Middleweight165 said: @Mark am i right in saying that if I want a players true positional rating, in its simplest form, I need to reduce the GS positional rating by 2.17 for every point under 20? Expand
It is by 2.17%. For example you have a player rated 18/20 for a position and GS shows 56.84%. The calculation is 56.84% - (56.84% x (2.17% x 2)) = 54.37%
Jamolisim said: @Mark Does your chart on page 1 you show (not FM scout one) work for LLM as well or are ratings weighted differently? Thank you in advance for your response. Expand
I only play LLM. It works across the board but it is harder to find players at lower levels. You need to reduce your expectations
Evanscam97 said: Thanks anyway mark. I admittedly went down an excel rabithole yesterday evening and set up a quick "best xi" picker based off stats pulled from GS Arena, using the YK Ratings and Position scores and your formula to derive true positional rating...
Pulling the data from GS is very awkward, however. Cant find a simple way to copy and paste the data and exporting doesn't give the position data (x/20) for whatever reason. Having to screenshot and use a "picture to table" conversion. Expand
Yes, unfortunately that aspect is manual. I haven't been able to find the info in the game itself to be able to download it and it is one of many fields not downloadable thru third party applications due to FM licencing.
If it helps, you don't need to include any ratings below 10
Evanscam97 said: Hi there Mark, Been lurking for a while now, appreciate all that you (and the other great people on here) do for the community! Was just curious as I looked at the above post - Is there any way to change the positions that GS uses for these formations? I.e in the 4-1-3-2, switching TS (Default) to FS, and MC x3 to AM R/C/L etc in line with the @ZaZ formation for example.
Purely the lazy side of me would find it useful at a glance for seeing Best XI (Unless there another way in GS that I am unaware of) as well as the weakest members of the Best XI
Thanks again mate Expand
I cant see any easy way to do this. There is no functionality I can see in GS to facilitate it. You can maybe look for other structures that use the position you are after and see which players are picked.
johnconnerson said: Are the Potential positional ratings worth looking at when scouting young players? Or am I better off looking at something else?
Does the 'Probability Of Development' mean anything? Expand
I always look at current first. I do look at agility difference and potential ability when looking at younger players but cant say I have looked at Probability of Development.
dzek said: So for every position we use the same number? Expand
I am unsure what you mean by use the same number for every position, does that mean the same calculation?
Look at post #245 above. The player has a different Genie Rating for each position and a different position score out of 20 for each position. And resulting calculation is a different final rating for each position.
dzek said: Its working actually for positional rating also. I mean if i have a player with score 20 in winger and another player who has score 19 in winger position i can see which one is better.
For example player A has 80% positional rating in AMR with 20 position score and player B has 85% positional rating in AMR with 19 position score. With your formula i mention above we can determine who is the better right? Expand
dzek said: So if i use ykykyk balanced ratings in excel spreadsheet and do the calculations on my own, which is the exact number to use for different positions? Including weak foot attribute because in Genie Scout we cant set foot weights for any position and its something that i asked Eugene to add it and he told me that will add it later, maybe in the next game version.
I put 48 for every position but i saw your full formula here and you use different number each time. Please clarify this a little bit further. Thank you! Expand
That link is for the general rating not the positional rating.
dzek said: I have a question about your formula. Why you use 46 and not 48?
If I understand correctly, number 46 is the total attributes included within the formula. Is that correct? Expand
It is not quite that simple. I looked at the percentage impact in the FM Arena Attributes testing table for Playing Position Rating both in this year's and last year's games. It worked out to be very close to 2.5% per point or 1/40th. Then I needed to calculate what Genie Scout uses for a reduction, which is quite a minor adjustment. I took a download of around 1000 wingers and looked at the differences in ratings against their positional ratings out of 20. Moving from 1/40th to 1/46th was the nearest to the adjustment.
johnconnerson said: I have a question about this formula. Are you going off of the General Rating from a player's page in Genie Scout? Or the Positional Rating? When I plug in the General Rating into A1 and the position rating (the 1-20 one) I'm not getting close to GS's Positional Rating. Am I doing something wrong? Expand
Here is a screenshot of one of my players from Genie Scout. Top left you can see the score out of 20 for the 4 attacking positions. Further along you see a rating for each position.
Now here is the screenshot of my spreadsheet having calculated the ability for each position using the algorithm I provided above.
So the calculation for AML is the GS rating of 47.22% * (1-(20-16)/46). 20-16 = 4. 4/46 = 0.087. 1-0.913 = 0.913. 47.22% * 0.913 = 43.11%.
I am hoping this clarification helps you understand
If I have 2 players, A with a rating of 80% and a positional score of 20 and B with a rating of 85% and a positional rating of 18. Player B's adjusted scorer rating is around 80.5%, on paper he would perform better? I know the margins are very small so might not result in real world better performance Expand
My calculation would be that Player B would rate 81.3% so I would go with him. Training and playing him in that position may even lift the position rating up to 19 or 20.
Following the first round of testing I then seeded them and reran the tests. So I now have 368 games for each tactic. Results are below. The next step is running the top 2 tactics against the new ZaZ Fire 0.3 and Perfect by super korean to see how those two stack up. To make it fair the second run will be against mid table tactics.
They were all pretty good considering my expectations for these teams. The top 4 tactics were outstanding with not much between them through the whole process.
Table Update
Group 10 Update
Group 11 Update
Group 12 Update
FINAL TABLE
Fair enough
What does your leaderboard for tactics look like at the moment @Eric
P.s. I setup the rest and everything.
FM Arena did some work on this and advised us earlier, in this thread at post #17 on page one, of some of their findings. See link below.
Development Training Info
Can you expand a bit further on you trying to get value for your money that you described in this post https://fm-arena.com/find-comment/14024/
I dont quite understand how you can find value. Why wouldn’t you just use positional rating and buy the highest rated player you can afford?
Sure. I looked at 2 tables, the Cost of Attribute Table by position and the ykykyky positional rating weightings. I only looked at attributes that were 50 or higher in the ykykyky positional rating weightings and then divided this score by the corresponding value on the Cost of Attribute Table. This then gave a value rating, higher the better.
My expectation of this score is that if a player has high values for these value scores their rating will be better value than those that are higher in the high cost attributes. I use these in the filters and it seems to be working, although I haven't done comparisons against those with that are higher in the high cost attributes. And I always include acceleration in my filters, other than for DM or GK.
So not really tested but an assumption.
I'm looking at a player who has a position value of 1 and a positional rating of 66.9%, so using your calculation, his true positional value is 39.3%. Is that correct?
Your calculation is correct. Of course you can improve your players position value and therefore their positional rating by training them and playing them in the position.
GS have a very minor adjustment for positional value, it is around 0.35% per point. My values are derived from the FM Arena testing over the last 2 years, and I believe they are much closer to the mark.
It is by 2.17%. For example you have a player rated 18/20 for a position and GS shows 56.84%. The calculation is 56.84% - (56.84% x (2.17% x 2)) = 54.37%
I only play LLM. It works across the board but it is harder to find players at lower levels. You need to reduce your expectations
Pulling the data from GS is very awkward, however. Cant find a simple way to copy and paste the data and exporting doesn't give the position data (x/20) for whatever reason. Having to screenshot and use a "picture to table" conversion.
Yes, unfortunately that aspect is manual. I haven't been able to find the info in the game itself to be able to download it and it is one of many fields not downloadable thru third party applications due to FM licencing.
If it helps, you don't need to include any ratings below 10
Been lurking for a while now, appreciate all that you (and the other great people on here) do for the community! Was just curious as I looked at the above post - Is there any way to change the positions that GS uses for these formations? I.e in the 4-1-3-2, switching TS (Default) to FS, and MC x3 to AM R/C/L etc in line with the @ZaZ formation for example.
Purely the lazy side of me would find it useful at a glance for seeing Best XI (Unless there another way in GS that I am unaware of) as well as the weakest members of the Best XI
Thanks again mate
I cant see any easy way to do this. There is no functionality I can see in GS to facilitate it. You can maybe look for other structures that use the position you are after and see which players are picked.
I use @ZaZ ones. They appear to be the best now
Does the 'Probability Of Development' mean anything?
I always look at current first. I do look at agility difference and potential ability when looking at younger players but cant say I have looked at Probability of Development.
I use Fast Striker calculations for False 9.
I am unsure what you mean by use the same number for every position, does that mean the same calculation?
Look at post #245 above. The player has a different Genie Rating for each position and a different position score out of 20 for each position. And resulting calculation is a different final rating for each position.
For example player A has 80% positional rating in AMR with 20 position score and player B has 85% positional rating in AMR with 19 position score. With your formula i mention above we can determine who is the better right?
Correct
I put 48 for every position but i saw your full formula here and you use different number each time. Please clarify this a little bit further. Thank you!
That link is for the general rating not the positional rating.
If I understand correctly, number 46 is the total attributes included within the formula. Is that correct?
It is not quite that simple. I looked at the percentage impact in the FM Arena Attributes testing table for Playing Position Rating both in this year's and last year's games. It worked out to be very close to 2.5% per point or 1/40th. Then I needed to calculate what Genie Scout uses for a reduction, which is quite a minor adjustment. I took a download of around 1000 wingers and looked at the differences in ratings against their positional ratings out of 20. Moving from 1/40th to 1/46th was the nearest to the adjustment.
Hope this helps
Here is a screenshot of one of my players from Genie Scout. Top left you can see the score out of 20 for the 4 attacking positions. Further along you see a rating for each position.
Now here is the screenshot of my spreadsheet having calculated the ability for each position using the algorithm I provided above.
So the calculation for AML is the GS rating of 47.22% * (1-(20-16)/46). 20-16 = 4. 4/46 = 0.087. 1-0.913 = 0.913. 47.22% * 0.913 = 43.11%.
I am hoping this clarification helps you understand
It depends on your tactic. I generally only use AFs so I use Fast Striker
Middleweight165 said: I'm catching up on this thread so lots of questions all at once
If I have 2 players, A with a rating of 80% and a positional score of 20 and B with a rating of 85% and a positional rating of 18. Player B's adjusted scorer rating is around 80.5%, on paper he would perform better? I know the margins are very small so might not result in real world better performance
My calculation would be that Player B would rate 81.3% so I would go with him. Training and playing him in that position may even lift the position rating up to 19 or 20.