Hi, @Fungi Bal - I can, it’s a little fugly right now because I’m experimenting. Also, the XLS also includes my ykykyk weights and ratings because I’m still using those as a “tiebreaker” decision factor. In other words, if I’m comparing 2 offensive players with about 2.5 GD over a player of all 12s - then I’ll look at how large the aggregate YKYKYK diff is and pick the larger YKYKYK and best $ deal for my acquisition.
1- Just to repeat - for Acceleration - 49 Goal Diff.
2 - Divided by 12 to reach 4.08 (contributions per point for goals over 380 (38 matches in the simulated season* across 10 potential contributing positions - note this is the weakness with the approach IMHO - I don’t know how the goals were distributed across the team in the aggregate simulations).
3 - Divided by said 380 - .0107 per point, per player, per match.
EXAMPLE//
Balikwisha / Offense / Antwerp - creates .0654 total across all attribute. My “rating” for him is a 1.59
The 1.59 is (.0654 * 38)-( .0612 * 38) or 24.85- 23.256
The .0654 is the sum of all GD weights for all attributes (has 15/14 pace/acc) 38 is just a random number of matches - I suppose you could use anything here, I used 38 for consistency with the sims. .0612 is thee sum of all GD weights for all attributes = 12 (again, suppose you could use 10 if you’d like, I chose 12)
Also, I have to admit I did the multiplication by 38 for the final value only for my own eyes. In other words, aesthetically, the 1.59 just views better than .0042 in the XLS when comparing players.
So I tried a thing. It may be working. In an attempt to construct a roster optimized to the attribute testing results, I did the following (please note, I’ll be using terms like “per player” or “per match” or “per year” - these are more terms of convenience than a literal expectation, more on that later):
1) Took the Goal Diff for each attribute (positive and negative) and divided it by 12 to get Goal Diff per point of attribute - 12 comes from the difference between the “8” and the “20” used in the testing itself. - Acceleration had a Positive or Offensive Goal Diff of 49 going from 8 to 12, so each “point” is worth 4.08 2) Divided that by 380 (10 players, 38 matches) to arrive at per point, per match, per player. - Result is a number (for acceleration and offensive positions, the number was .0107).
I decided to use this number to compare players based on attributes as another “weighting” exercise. While it allows one to compare two players, after experimentation, it was easier for me to consume and compare when against an “average” - I decided to use a player with “12” as that average player.
So now I can compare the aggregate player attributes for (only) those attributes with clear, demonstrable impact in the attribute testing simulations (acceleration, pace, dribbling, anticipation, Jumping Reach, etc.) and create something closer to baseballs “Wins Above Replacement” using a player with all 12s in the attributes that matter.
For any one player explaining in terms of a spreadsheet -
The player is a row, the attributes are columns.
The cells at the intersection are the player’s attribute # from FM24.
There is another column that is a sum of each attribute multiplied against the attribute contribute (per point, per match, per player) in #2 above - I call this # Goals per Match (or Goals Prevented per Match for defense).
I multiply that # by 38 to get the aggregate Goals per Season.
I then subtract Goals Per Season for the player from the Goals Per Season for an “all 12” player to get my “Goals Above Replacement” (GAR) for Attacking Midfielders and Attackers/Strikers - or goals prevented (GPAR) for DL/DR/DC. I took the a similar approach for GK.
For DM - I compiled both the DMs GAR and their GPAR multiplied by -1 to create an aggregate positive number of Goal Contributions Above Replacement (a goal saved is a goal earned).
Ya Boi Eren Dinkci, in my save at age 25, is 3.4 Goals per year greater than a player with all 12s. Even with the haircut.
My man Phil Neumann, in my save at age 30, is 03.75 Goals Prevented per year greater than a player with all 12s.
Bradley Ibrahim as a 23 year old DM in my save is about 5.2 GCAR.
Some observations -
1) These numbers don’t include the impact of the formation (obviously). I’ve been using a meta 424 in the 3 saves I’ve been experimenting with this effort - it would be interesting to see what an “all 12s” 424 team produces and somehow compensate for that (not sure how).
2) I combined this in my XLS (I play on iPad so I my experience is console-ish) against my old ykykyk weights - there are differences in quality (Lucas Bretelle as a DM, for example, tends to have high ykykykyk based weights in my saves - surprised at how “meh” he scored in GCAR.
3) Totally helps financially. I was comparing a player named Connor Bradley for DR against one named Pierre Nadjombe. Similar scores, but the cost per GPAR was crazy in favor of Nadjombe. Even though Connor does seem to have better overall attribute numbers - many of them just don’t matter.
4) I don’t believe at all that if Dango Ouattara is a 27 goal per 38 matches player. I’m sure in some distribution of outcomes, he could be - but that number is useful because I can peg it against an all 12 (23.3 per 38). The labels I use and numbers shouldn’t be thought of as literal expectations.
5) In terms of lineup construction, it really helps to compare lineup A vs. lineup B vs. lineup C (etc.) when planning / scheduling tournaments vs. league play and creating depth.
6) Similar roster construction benefits and lineup comparison benefits when players could be multiple roles. Both Dinkci (AMR/STC) and Neumann (DR/DC) mentioned above qualify at different positions and have (albeit subtly) different GAR/GPAR scores for the positions. But over 38 matches, I’d like to believe that .3 GAR would be a difference that justifies Dinkci at AMR vs. STC.
Anyhow, happy to share more if anyone cares. It’s a fun hobby.
Also, and I hate to be dense here, but *if* I were to use this as part of talent evaluation - why (and I’m not trying to be facetious here), why would I care about points?
I feel like the process would be to understand impact on goals compared to some standard (Goals Above Replacement Player for FFM vs. WAR for baseball as an example)?
I created a few generic "standards" - all measured in an XLS using ykykyk weights for each position.
What I do is enter in the player score for the attributes identified in ykykyk.
Then I multiply that player's score against the weight for the attribute. So if Pace for a position is a .8 weight, and the player is a "10" - they are given an 8.
I then add them all up into a aggregate - Total Weighted Player Score. TWPS becomes the basis for other measurements/comparisons.
As an example, in my current SGE save, N'Dicka has a TWPS of 104.3 as a DC. Baidoo has a TWPS of 93.4 at the same position. Hardley an 86.8.
Next is "weighted player score above average" - where I compare the player in position to a generic "all 12s" player score. Technically, I guess "10" might be average, but I use 12 as my baseline.
So N'Dicka has a WPSAA of 11.9, Baidoo is 1, and poor Hardley is -5.6. N'Dicka is above average, Baidoo is about average, and Hardley, well, I hope I can loan him to a Bundasliga 2 side for his cost in salary.
I've also been playing Bromley from National to (just today) Premier for the 28/29 season. WPSAA has been useful in that journey. At National I was shooting for a WPSAA of -3 for each position in my starting 11. In Championship - +4 to +7 helped me get the highest xG and expected points in the league (Fulham overachieved and took top spot).
The other score I use is % of "Best In World" - when the top 50 in the world reports come out - I find the players with the top TWPS for each position. I then compare my player against the BIW. A simple %. It's tougher to upkeep this stat, and it is variable from save to save. But it can be nice to know how far a 22 year old Junior Kroupi is from an aged Mbappe.
Final use for TWPS is something I'm playing around with called "Positional Difficulty." An attempt to measure market scarcity - "PD" measures the TWPS for Best In World for the position against a TWPS of all "20's."
The theory being that the greater the gap between all 20 and BIW - or the higher the PD score - the more difficult it is to find high TWPS scoring players for that position. This could be total nonsense, of course, but I use it prioritize spending. DR and DM positions have had higher PD scores in a few saves I have - and based on playing style, I tend to spend more aggressively for those positions.
I appreciate the work in Python. For those asking "why don't" - well, some of us do a bit more. I have my own ratings (customized to my passing-heavy style of play), my own metrics (ratings against "average" or "best in the world" ) and my own views (what is my best squad for national competitions?).
Diversity of thought and effort is pretty good. But I'm sure if you just use Squirrel Plays or whatever, you'll be fine too
@skawkclsrn - GA/GS difference testing comes from the attribute testing on this site. It’s specious for some of the reasons @MeanOnSunday mentions. But I think it’s somewhat meaningful in that we’re looking not just at the inputs into the simulator (as ykykyk or Mark’s or SIs or whomever) but shows us how the simulator responds. It allows us to ask ourselves questions like, “if ‘Off the Ball’ isn’t meaningful in a controlled environment - how could it be meaningful in my environment - (a 4-2-4 in my case)?”
First, let me be the first to say I’m not good at analytics, soccer (I’m just a Yank who became interested in it at the tail end of COVID), nor FootballManager (I’m only on my 5th career here).
So negative feedback is EXACTLY what I’m looking for. I hadn’t thought about the correlation and additive nature of calculating.
RE: positional differences being averaged out - wouldn’t that also be a potential benefit? Going into this analysis, I thought about:
1) the fact that it’s not my tactic, and 2) it’s coarse on positional differences, and 3) it’s older (FM22), and 4) it doesn’t test all attributes, and 5) some of the differences may not be statistically significant (+2 goals over 913 simulations? Meh?).
So the compromise I’m offering is not weighting every attribute, but simply trying to combine those that the simulation seems to respond to, and then adding enough attributes that are NOT measured in the attribute testing with ykykyk weights.
Finally, I’m starting to think that any of these approaches are “good enough to be good enough.” Using @Mark weights - I’m dominating MLS on the Touch version. Like no scum save just went a season without losing. Would it even matter if I switched to ykykyk? Is my success because I “cheated” came onto this site, and emulated 4-2-4 Bombyte Tweak more than finding really great players?
I mean, I had success in Ligue 1 with Lorient (!) just measuring how many attributes weighted above 50 were above 16 for each player. Not nearly as granular as any of the above - but still successful.
As long as I know that P/A rule, and the stars are a lie, and I’m playing a proven tactic - is the whole XLS exercise just there for me to not ever overpay for Xavi Simmons again?
They both operate under the assumption that the information provided by Sports Interactive (SI) within the game is entirely accurate. Per position attributes are divided into three categories: Required, Preferred, and Irrelevant. Required attributes carry the most weight, preferred attributes a bit less, and irrelevant attributes are ignored in their calculations.
Right? That's one of the reasons I love the attribute testing on this site, but as you say, it could use what feels like an endless amount of work and variable manipulation to get closer to reality.
I love the python tooling. Will switch someday to it - if only because I want to weight statistical results in decision making more than the "attribute weight only" approach I have now.
This is probably one of the best marketing campaigns ever.. agree, and esp. if you want to get on the radar of enthusiasts.
My guess also (given Bromeley's stature) is that they want a better analytical program, but don't have the money to do more than find someone who does this in their spare time, hates their low-paying job, and will give them heart and soul for little compensation.
Hell, as a yank, I'll do it for free, part-time, if they pay for my expenses...
There's an appeal to me to make my experience a little less "reverse engineering the simulator" at this point. That said, I've thought about:
1) Playing on PC and getting a skin that blocks the attributes altogether. The issue with that approach is that I feel like if I were, say, actually taking the Bromely job, there would be some subjective measurements for things like "determination" or "composure" and objective measurements for things like pace and acceleration. So....
2) There are skins and whatnot that give ranges for the attributes on the PC I'm lead to understand - which might be closer to how an analytics department might work.
I thought I'd reach out here and see if anyone plays without the precise (or even showing) attibutes, why, what your set-up is like, etc.
I just type them in. Again, I've been playing on an iPad, and with it's limited capabilities, I will just type stuff from screen to screen. I suppose I could create a custom view based on the attributes for the individual position, but i'm lazy.
ABOUT THE JOB We are looking for a talented Football Manager™ gamer who wants to test their skills in the real world and take their first step in a potential career in professional football. This person will join Bromley’s tactical staff for the rest of the season. No CV or experience required. This is a backdoor to the backroom for someone who wants to explore a career at a professional club. We want a Football Manager™ expert, the guru of the group chat, the final score forecaster. In this role you will not have to make any executive decisions, but you’ll be shadowing and learning directly from all backroom positions to understand the ins and outs of managing a real football club.
KEY ROLES AND RESPONSIBILITIES Contribute to all 1st team analysis, pre, during and post-match, live training and notational analysis. Support with analysis, evaluation and feedback for 1st Learn to produce detailed 1st team opposition analysis, using subjective and objective data in line with the Club’s philosophy and expectations. Tap into your Football Manager™ knowledge to support the management and backroom team on all areas including team selection, tactics and scouting. Present detailed reports and feedback to the 1st Team Manager and key staff. Develop and maintain performance related databases for analytical reviews. Assist in target setting and developing individual player performance plans. Perform additional duties as required to meet the needs of the Club. Stream and vlog about your experience with Bromley FC in-line with the agreed terms and conditions.
Because I just have my little XLS - the depth of analysis you're putting in seems pretty cool. I started playing this game to learn more about soccer (yank here). Now I'm enjoying the ability to analyze the "game within the game" - and with 15 XLS and so many formulas - I feel like it's great to have a community of analytical minds.
I'm toying with a new rating system, and would love some feedback. It ignores GK - but bear with me on that. I think a different approach is warranted for that special position.
First question is - how much do we believe the underlying simulation algorithms change from year to year? Specifically,
Do we suppose that these would be meaningful in 24?
If so, then I like these more than weights, because it shows how the simulator responds to inputs - it's more "evidence-based."
So I did a little analysis last night, and compared the Attribute Testing results to ykykyk, in an effort to create a new roster prioritization / scoring approach that relies as much as possible on the Attribute Testing results.
Methodology -
1) Looked at the measured attributes and segmented them into whether there was a defensive impact, an offensive impact, or both.
Looking specfically at Goals Allowed (Defense) vs. Goals Scored (Attacking).
2) Next, created a basic Attack template for attacking heavy roles, and a Defense template for Defensive heavy roles.
The different shades of green are used to help me identify if the impact of the attribute was pretty strong, medium, or weak.
I then reckoned that I would limit the number of attributes I measured based on roughly 50% of total attributes available (so 18 attributes to be considered). That meant that for attackers - there were only 14 relatively meaningful attributes, and for defenders - 16.
4) I decided to "fill in the missing attributes" by comparing the Attribute Testing with ykykyk weights for each position. So, Finishing isn't represented in the Attribute testing, but sure seems important to attackers in terms of weights, so I added that.
5) This all resulted in the following:
So you see the attributes, the numbers in the first row represent the ykykyk weights, the second row numbers are the impact on Goals Allowed difference or Goals Scored difference.
The red numbers represent extrapolation (complete hogwash, just interested in looking at unmeasured attributes i added, and if the attribute wasn't measured, what is the possible impact largest impact based on attributes with similar ykykyk weights). I pretty much know the red numbers are hogwash because Off The Ball is weighted highly in some positions, but with no demonstrable impact in the attribute testing results.
From here, I would simply replace my current approach (weights-only used in ranking, with this new idea.
I want to measure a couple of things depending on the state of the club I'm managing.
What I do is try to come up with an aggregate rating - that may be just entering key attributes against ykykyk or somesuch. That's useful. But then I do two things -
1) Compare my player aggregate rating to a player with "all 12s" - in an attempt to recreate something like "Wins Above Replacement" as used in baseball. Basically - is my player better than the average flotsam I can pick up easily - and how much better? This helps in roster construction and transfer window opportunities, and esp if I'm playing a small or poor club (or the MLS).
2) Every time the game posts "Top 50 Best in World" to my inbox - I go and find the highest rated player in that top 50 and put them in the same ratings as I use for my players. Then I create a comparison metric that answers the question "how far away is my player from the best on the market?"
This is useful for clubs that are rich and / or when I want to dethrone PSG.
3) I'm playing with the idea of comparing my player's ability vs. how much I'm spending. I haven't come up with a good approach other than to divide the aggregate rating by salary here, which can be helpful, but I feel like there might be a better approach.
Finally, I'm playing with the idea of using/stressing the attribute testing on this forum:
Which I *love* because it shows me based on results vs. the weights. I feel like measuring based on how the simulators algorithms respond is more meaningful than just how it constructs the inputs into the simulator.
Why you don't use FMRTE? I think its more advanced scouting tool and also with much more flexibility on ratings etc. I was using Genie Scout too before some years but the developer of the app don’t update it anymore with new features, only fixing some bugs and make it work for each edition/patch.
Thank you again for your time and your effort you put. Keep going Expand
A couple of reasons:
First, I play on an iPad. So it’s easier for me to just use an XLS (which I really only use 3-4 times a season.
Second, I have custom metrics that help me evaluate a player. So if I decide to switch to PC or Mac, if FMRTE allows me to do custom analytics - that might be an attractive option.
Third, when I play with the numbers myself - I learn more about the game maybe?
@Mark - See below:
Attribute, Goal Difference per point of Attribute (negative for defense, positive for offense, DM get’s both into an aggregate as mentioned above):
GK -
Acceleration (-.022),
Agility (-.018)
Anticipation (-.0006)
Concentration (-.0006)
Positioning (-.0006)
Aerial Reach (-.0006)
Command of Area (-.0006)
Communication (-.0006)
1st Touch (-.002)
Handling (-.0006)
Kicking (-.0007)
1x1 (-.002)
Reflexes (-.022)
Defense
Acceleration / -.0105
Agility / -.0013
Balance. / -.0018
Jumping Reach / -.0018
Natural Fitness / -.0004
Pace / -.013
Stamina / -.0035
Strength / -.0011
Anticipation / —.002
Concentration / -.0018
Determination / -.0011
Dribbling / -.0007
1st Touch / -.0004
Offense
Acceleration / .0107
Agility / .0013
Balance / .0022
Jumping Reach / .0031
Natural Fitness / .0004
Pace / .0099
Stamina / .002
Strength / .0015
Anticipation / .0033
Composure / .0004
Concentration /.0009
Decisions / .0004
Determination / .0009
Work Rate / .0013
Crossing / .0011
Dribbling / .0055
Finishing / .0018
1st Touch / .00004
Long Shots / .0015
Heading / .0011
Passing / .0009
Technique / .0004
1- Just to repeat - for Acceleration - 49 Goal Diff.
2 - Divided by 12 to reach 4.08 (contributions per point for goals over 380 (38 matches in the simulated season* across 10 potential contributing positions - note this is the weakness with the approach IMHO - I don’t know how the goals were distributed across the team in the aggregate simulations).
3 - Divided by said 380 - .0107 per point, per player, per match.
EXAMPLE//
Balikwisha / Offense / Antwerp - creates .0654 total across all attribute. My “rating” for him is a 1.59
The 1.59 is (.0654 * 38)-( .0612 * 38) or 24.85- 23.256
The .0654 is the sum of all GD weights for all attributes (has 15/14 pace/acc)
38 is just a random number of matches - I suppose you could use anything here, I used 38 for consistency with the sims.
.0612 is thee sum of all GD weights for all attributes = 12 (again, suppose you could use 10 if you’d like, I chose 12)
Also, I have to admit I did the multiplication by 38 for the final value only for my own eyes. In other words, aesthetically, the 1.59 just views better than .0042 in the XLS when comparing players.
1) Took the Goal Diff for each attribute (positive and negative) and divided it by 12 to get Goal Diff per point of attribute
- 12 comes from the difference between the “8” and the “20” used in the testing itself.
- Acceleration had a Positive or Offensive Goal Diff of 49 going from 8 to 12, so each “point” is worth 4.08
2) Divided that by 380 (10 players, 38 matches) to arrive at per point, per match, per player.
- Result is a number (for acceleration and offensive positions, the number was .0107).
I decided to use this number to compare players based on attributes as another “weighting” exercise. While it allows one to compare two players, after experimentation, it was easier for me to consume and compare when against an “average” - I decided to use a player with “12” as that average player.
So now I can compare the aggregate player attributes for (only) those attributes with clear, demonstrable impact in the attribute testing simulations (acceleration, pace, dribbling, anticipation, Jumping Reach, etc.) and create something closer to baseballs “Wins Above Replacement” using a player with all 12s in the attributes that matter.
For any one player explaining in terms of a spreadsheet -
The player is a row, the attributes are columns.
The cells at the intersection are the player’s attribute # from FM24.
There is another column that is a sum of each attribute multiplied against the attribute contribute (per point, per match, per player) in #2 above - I call this # Goals per Match (or Goals Prevented per Match for defense).
I multiply that # by 38 to get the aggregate Goals per Season.
I then subtract Goals Per Season for the player from the Goals Per Season for an “all 12” player to get my “Goals Above Replacement” (GAR) for Attacking Midfielders and Attackers/Strikers - or goals prevented (GPAR) for DL/DR/DC. I took the a similar approach for GK.
For DM - I compiled both the DMs GAR and their GPAR multiplied by -1 to create an aggregate positive number of Goal Contributions Above Replacement (a goal saved is a goal earned).
Ya Boi Eren Dinkci, in my save at age 25, is 3.4 Goals per year greater than a player with all 12s. Even with the haircut.
My man Phil Neumann, in my save at age 30, is 03.75 Goals Prevented per year greater than a player with all 12s.
Bradley Ibrahim as a 23 year old DM in my save is about 5.2 GCAR.
Some observations -
1) These numbers don’t include the impact of the formation (obviously). I’ve been using a meta 424 in the 3 saves I’ve been experimenting with this effort - it would be interesting to see what an “all 12s” 424 team produces and somehow compensate for that (not sure how).
2) I combined this in my XLS (I play on iPad so I my experience is console-ish) against my old ykykyk weights - there are differences in quality (Lucas Bretelle as a DM, for example, tends to have high ykykykyk based weights in my saves - surprised at how “meh” he scored in GCAR.
3) Totally helps financially. I was comparing a player named Connor Bradley for DR against one named Pierre Nadjombe. Similar scores, but the cost per GPAR was crazy in favor of Nadjombe. Even though Connor does seem to have better overall attribute numbers - many of them just don’t matter.
4) I don’t believe at all that if Dango Ouattara is a 27 goal per 38 matches player. I’m sure in some distribution of outcomes, he could be - but that number is useful because I can peg it against an all 12 (23.3 per 38). The labels I use and numbers shouldn’t be thought of as literal expectations.
5) In terms of lineup construction, it really helps to compare lineup A vs. lineup B vs. lineup C (etc.) when planning / scheduling tournaments vs. league play and creating depth.
6) Similar roster construction benefits and lineup comparison benefits when players could be multiple roles. Both Dinkci (AMR/STC) and Neumann (DR/DC) mentioned above qualify at different positions and have (albeit subtly) different GAR/GPAR scores for the positions. But over 38 matches, I’d like to believe that .3 GAR would be a difference that justifies Dinkci at AMR vs. STC.
Anyhow, happy to share more if anyone cares. It’s a fun hobby.
I feel like the process would be to understand impact on goals compared to some standard (Goals Above Replacement Player for FFM vs. WAR for baseball as an example)?
(5.9 goal diff from 12 steps (8 to 20 or 5.9/12) and to quote earlier thread, the benefits seem linear in increase)
Reflexes for GK = roughly -10 GoalDiff for 1 player, for 38 matches
Pace = roughly -5.9 Goal Diff, for 1 player for 38 matches
Is that correct/valid conclusion?
Generally, budget permitting, I have a primary 11 and secondary 11 squad - and sub in from secondary.
What I do is enter in the player score for the attributes identified in ykykyk.
Then I multiply that player's score against the weight for the attribute. So if Pace for a position is a .8 weight, and the player is a "10" - they are given an 8.
I then add them all up into a aggregate - Total Weighted Player Score. TWPS becomes the basis for other measurements/comparisons.
As an example, in my current SGE save, N'Dicka has a TWPS of 104.3 as a DC. Baidoo has a TWPS of 93.4 at the same position. Hardley an 86.8.
Next is "weighted player score above average" - where I compare the player in position to a generic "all 12s" player score. Technically, I guess "10" might be average, but I use 12 as my baseline.
So N'Dicka has a WPSAA of 11.9, Baidoo is 1, and poor Hardley is -5.6. N'Dicka is above average, Baidoo is about average, and Hardley, well, I hope I can loan him to a Bundasliga 2 side for his cost in salary.
I've also been playing Bromley from National to (just today) Premier for the 28/29 season. WPSAA has been useful in that journey. At National I was shooting for a WPSAA of -3 for each position in my starting 11. In Championship - +4 to +7 helped me get the highest xG and expected points in the league (Fulham overachieved and took top spot).
The other score I use is % of "Best In World" - when the top 50 in the world reports come out - I find the players with the top TWPS for each position. I then compare my player against the BIW. A simple %. It's tougher to upkeep this stat, and it is variable from save to save. But it can be nice to know how far a 22 year old Junior Kroupi is from an aged Mbappe.
Final use for TWPS is something I'm playing around with called "Positional Difficulty." An attempt to measure market scarcity - "PD" measures the TWPS for Best In World for the position against a TWPS of all "20's."
The theory being that the greater the gap between all 20 and BIW - or the higher the PD score - the more difficult it is to find high TWPS scoring players for that position. This could be total nonsense, of course, but I use it prioritize spending. DR and DM positions have had higher PD scores in a few saves I have - and based on playing style, I tend to spend more aggressively for those positions.
Diversity of thought and effort is pretty good. But I'm sure if you just use Squirrel Plays or whatever, you'll be fine too
@skawkclsrn - GA/GS difference testing comes from the attribute testing on this site. It’s specious for some of the reasons @MeanOnSunday mentions. But I think it’s somewhat meaningful in that we’re looking not just at the inputs into the simulator (as ykykyk or Mark’s or SIs or whomever) but shows us how the simulator responds. It allows us to ask ourselves questions like, “if ‘Off the Ball’ isn’t meaningful in a controlled environment - how could it be meaningful in my environment - (a 4-2-4 in my case)?”
@MeanOnSunday
First, let me be the first to say I’m not good at analytics, soccer (I’m just a Yank who became interested in it at the tail end of COVID), nor FootballManager (I’m only on my 5th career here).
So negative feedback is EXACTLY what I’m looking for. I hadn’t thought about the correlation and additive nature of calculating.
RE: positional differences being averaged out - wouldn’t that also be a potential benefit? Going into this analysis, I thought about:
1) the fact that it’s not my tactic, and
2) it’s coarse on positional differences, and
3) it’s older (FM22), and
4) it doesn’t test all attributes, and
5) some of the differences may not be statistically significant (+2 goals over 913 simulations? Meh?).
So the compromise I’m offering is not weighting every attribute, but simply trying to combine those that the simulation seems to respond to, and then adding enough attributes that are NOT measured in the attribute testing with ykykyk weights.
Finally, I’m starting to think that any of these approaches are “good enough to be good enough.” Using @Mark weights - I’m dominating MLS on the Touch version. Like no scum save just went a season without losing. Would it even matter if I switched to ykykyk? Is my success because I “cheated” came onto this site, and emulated 4-2-4 Bombyte Tweak more than finding really great players?
I mean, I had success in Ligue 1 with Lorient (!) just measuring how many attributes weighted above 50 were above 16 for each player. Not nearly as granular as any of the above - but still successful.
As long as I know that P/A rule, and the stars are a lie, and I’m playing a proven tactic - is the whole XLS exercise just there for me to not ever overpay for Xavi Simmons again?
Right? That's one of the reasons I love the attribute testing on this site, but as you say, it could use what feels like an endless amount of work and variable manipulation to get closer to reality.
I love the python tooling. Will switch someday to it - if only because I want to weight statistical results in decision making more than the "attribute weight only" approach I have now.
My guess also (given Bromeley's stature) is that they want a better analytical program, but don't have the money to do more than find someone who does this in their spare time, hates their low-paying job, and will give them heart and soul for little compensation.
Hell, as a yank, I'll do it for free, part-time, if they pay for my expenses...
https://fmstag.com/fm-stags-custom-views-megapack-for-fm24/
There's an appeal to me to make my experience a little less "reverse engineering the simulator" at this point. That said, I've thought about:
1) Playing on PC and getting a skin that blocks the attributes altogether. The issue with that approach is that I feel like if I were, say, actually taking the Bromely job, there would be some subjective measurements for things like "determination" or "composure" and objective measurements for things like pace and acceleration. So....
2) There are skins and whatnot that give ranges for the attributes on the PC I'm lead to understand - which might be closer to how an analytics department might work.
I thought I'd reach out here and see if anyone plays without the precise (or even showing) attibutes, why, what your set-up is like, etc.
Anyone thinking/thought similar?
ABOUT THE JOB
We are looking for a talented Football Manager™ gamer who wants to test their skills in the real world and take their first step in a potential career in professional football. This person will join Bromley’s tactical staff for the rest of the season. No CV or experience required. This is a backdoor to the backroom for someone who wants to explore a career at a professional club. We want a Football Manager™ expert, the guru of the group chat, the final score forecaster. In this role you will not have to make any executive decisions, but you’ll be shadowing and learning directly from all backroom positions to understand the ins and outs of managing a real football club.
KEY ROLES AND RESPONSIBILITIES
Contribute to all 1st team analysis, pre, during and post-match, live training and
notational analysis.
Support with analysis, evaluation and feedback for 1st
Learn to produce detailed 1st team opposition analysis, using subjective and objective data in line with the Club’s philosophy and expectations.
Tap into your Football Manager™ knowledge to support the management and backroom team on all areas including team selection, tactics and scouting.
Present detailed reports and feedback to the 1st Team Manager and key staff.
Develop and maintain performance related databases for analytical reviews.
Assist in target setting and developing individual player performance plans.
Perform additional duties as required to meet the needs of the Club.
Stream and vlog about your experience with Bromley FC in-line with the agreed terms and conditions.
TO APPLY
Show your skills in FM24 by winning a domestic
league to gain the Xbox achievement “Championes”
Visit https://www.xbox.com/en-gb/promotions/the-everyday-tactician and follow the steps.
That's just badass.
First question is - how much do we believe the underlying simulation algorithms change from year to year? Specifically,
https://fm-arena.com/table/13-attribute-testing/#options
Do we suppose that these would be meaningful in 24?
If so, then I like these more than weights, because it shows how the simulator responds to inputs - it's more "evidence-based."
So I did a little analysis last night, and compared the Attribute Testing results to ykykyk, in an effort to create a new roster prioritization / scoring approach that relies as much as possible on the Attribute Testing results.
Methodology -
1) Looked at the measured attributes and segmented them into whether there was a defensive impact, an offensive impact, or both.
Looking specfically at Goals Allowed (Defense) vs. Goals Scored (Attacking).
2) Next, created a basic Attack template for attacking heavy roles, and a Defense template for Defensive heavy roles.
The different shades of green are used to help me identify if the impact of the attribute was pretty strong, medium, or weak.
I then reckoned that I would limit the number of attributes I measured based on roughly 50% of total attributes available (so 18 attributes to be considered). That meant that for attackers - there were only 14 relatively meaningful attributes, and for defenders - 16.
4) I decided to "fill in the missing attributes" by comparing the Attribute Testing with ykykyk weights for each position. So, Finishing isn't represented in the Attribute testing, but sure seems important to attackers in terms of weights, so I added that.
5) This all resulted in the following:
So you see the attributes, the numbers in the first row represent the ykykyk weights, the second row numbers are the impact on Goals Allowed difference or Goals Scored difference.
The red numbers represent extrapolation (complete hogwash, just interested in looking at unmeasured attributes i added, and if the attribute wasn't measured, what is the possible impact largest impact based on attributes with similar ykykyk weights). I pretty much know the red numbers are hogwash because Off The Ball is weighted highly in some positions, but with no demonstrable impact in the attribute testing results.
From here, I would simply replace my current approach (weights-only used in ranking, with this new idea.
Thoughts?
I want to measure a couple of things depending on the state of the club I'm managing.
What I do is try to come up with an aggregate rating - that may be just entering key attributes against ykykyk or somesuch. That's useful. But then I do two things -
1) Compare my player aggregate rating to a player with "all 12s" - in an attempt to recreate something like "Wins Above Replacement" as used in baseball. Basically - is my player better than the average flotsam I can pick up easily - and how much better? This helps in roster construction and transfer window opportunities, and esp if I'm playing a small or poor club (or the MLS).
2) Every time the game posts "Top 50 Best in World" to my inbox - I go and find the highest rated player in that top 50 and put them in the same ratings as I use for my players. Then I create a comparison metric that answers the question "how far away is my player from the best on the market?"
This is useful for clubs that are rich and / or when I want to dethrone PSG.
3) I'm playing with the idea of comparing my player's ability vs. how much I'm spending. I haven't come up with a good approach other than to divide the aggregate rating by salary here, which can be helpful, but I feel like there might be a better approach.
Finally, I'm playing with the idea of using/stressing the attribute testing on this forum:
https://fm-arena.com/table/13-attribute-testing/#options
Which I *love* because it shows me based on results vs. the weights. I feel like measuring based on how the simulators algorithms respond is more meaningful than just how it constructs the inputs into the simulator.
Why you don't use FMRTE? I think its more advanced scouting tool and also with much more flexibility on ratings etc. I was using Genie Scout too before some years but the developer of the app don’t update it anymore with new features, only fixing some bugs and make it work for each edition/patch.
Thank you again for your time and your effort you put. Keep going
A couple of reasons:
First, I play on an iPad. So it’s easier for me to just use an XLS (which I really only use 3-4 times a season.
Second, I have custom metrics that help me evaluate a player. So if I decide to switch to PC or Mac, if FMRTE allows me to do custom analytics - that might be an attractive option.
Third, when I play with the numbers myself - I learn more about the game maybe?