(Completed in stages)FM27 New Player Attribute (1,4,8,12,16,20) Test (Updated 22/11/2026)

harvestgreen22

New member
Joined
Nov 11, 2026
Messages
11
Reaction score
18

=====​

updated 2026-11-22​

FULL EXCEL

球员属性New Player Attribute (20261122 ver4).xlsx​

=====​

"github.com/lqf0624/fm-tactics-lab/releases"
发布 · lqf0624/fm-tactics-lab
感谢lqf0624 的工具Fm Tactics Lab tool。本次测试使用该工具进行。
Thanks to lqf0624 for the Fm Tactics Lab tool. This test was conducted using this tool.

1.利用工具导出AI主教练为他们的球员设计的真实战术、指令
Use tools to export the actual tactics and instructions designed by AI head coaches for their players.

Fm Tactics Lab 可以从存档中导出教练的真实战术策略。
这项新测试包含了19 位英格兰超级联赛教练的真实战术策略,
这些策略是从使用“fm-tactics-lab”存档中导出的,

包括他们在主场和客场比赛中的不同战术倾向、调整和指示。以及主教练根据他们球队员的特性来设置适合球员的战术
因此,对手的范围扩大了,与“预设战术”相比,“现实战术”更能描述游戏中教练将采用的应对方法。


Fm Tactics Lab allows for the extraction of authentic tactical strategies used by managers from save files.
This latest test incorporates the real-world tactical strategies of 19 Premier League managers,
extracted from save files using "fm-tactics-lab."

These include the specific tactical approaches, adjustments, and instructions employed by these managers in both home and away matches.
as well as the tactical setups they devise to best suit the individual attributes of their players.

Consequently, the range of opponents is broadened; compared to "preset tactics," these "realistic tactics" provide a more accurate representation of the strategies managers will actually employ within the game.


2.测试使用的战术
Tactics used for testing

STORMBRINGER 27 edit for test2.fmf​



“玩家”队相对之前的属性测试,不再使用预设战术。
我使用的是“STORMBRINGER 27 BBMIW 424 P102 FA ( STORMBRINGER 27 BBMIW 424 P102 FA )”战术的修改版。
如有需要,您可以下载查看。这是424-442战术。

For the "Player" team, I have moved away from the preset tactics used in the previous attribute tests.
I am now using a modified version of the "STORMBRINGER 27 BBMIW 424 P102 FA" tactic. “fm-arena.com/threads/stormbringer-27-bbmiw-424-p102-fa.111/”
Feel free to download and examine it if you’d like. It is a 4-2-4/4-4-2 hybrid tactic.

3.不同的战术不会对属性需求有很大变化
Different tactics do not significantly alter attribute requirements.

根据caiograco的帖子FM27 match engine: what we know so far
战术调整带来的属性变化不会有显著差异。
总体结论仍然相似,由于测试环境条件变化,细节有所不同。
本次测试的样本量比我之前的测试更大,因此准确率更高。

Based on the findings from caiograco’s post, "FM27 Match Engine: What We Know So Far,"
"fm-arena.com/threads/fm27-match-engine-what-we-know-so-far.122/"
“github.com/caiograco289/Fm-Match-Lab/tree/main/docs/studies”

tactical adjustments do not result in significant differences in attribute-related outcomes.
While the overall conclusions remain similar, there are some variations in the details due to differences in the testing environment.
This test utilized a larger sample size than my previous tests, leading to a higher degree of accuracy.


4.测试条件设置Test Condition Settings

测试条件,19名英格兰超级联赛球员的属性均保持不变。
玩家取代了“赫尔城”队。该队所有球员的属性(包括隐藏属性)均统一设置为12,身高统一设置为180厘米。
测试中改变全体非门将球员的一项属性,测试了当该属性分别为 1、4、8、12、16 和 20 时,球队的胜率/负率/平局率、预期进球数 (xG) 和积分。测试中主客场情况互换。
其中,将“战术理解”、“球员投入”和“下一场比赛信息”设置为 75。“球员能力”会根据球员的属性自动计算。
原因是当前游戏中很难达到较高的战术效率。

For the test conditions, the attributes of players across the English Premier League remained unchanged.
The user took control of Hull City. All players on the team—including their hidden attributes—were set to a uniform value of 12, with a standardized height of 180 cm.
The test involved isolating a single attribute for all non-goalkeeper players and observing the team’s win/loss/draw rates, expected goals (xG), and total points when that attribute was set to 1, 4, 8, 12, 16, and 20, respectively. The test accounted for both home and away fixtures.

Additionally, “Tactical Understanding”, “Player Buy-In”, and “Next Match Info” were set to 75.
"Player Ability" was calculated automatically based on the players' attributes.
This approach was chosen because achieving high tactical efficiency is notoriously difficult in the current version of the game.


5.样本数sample size
总样本数256,500场
原计划更多但是太耗时。
每个属性、条件、误差,对应的样本数显示在excel中

Total sample size: 256,500 instances.
More were planned, but the process is too time-consuming.
The sample size corresponding to each attribute, condition, and error is shown in the Excel file.


6.主要使用主场表来衡量球员
Player evaluation is primarily based on performance in home games.

主场胜率≈40%,客场胜率≈12%
显然,由于客场比赛的胜率极低,几乎没有进一步下降的空间。
一些重要属性在统计数据中可能被严重低估,因为它们的胜率太低,无法再进一步降低。
因此,从主场比赛中得出的结论相对更可靠。

Player evaluation is primarily based on performance in home games.
The win rate in home games is approximately 40%, compared to about 12% in away games.
Clearly, given the already extremely low win rate in away games, there is almost no room for it to drop any further.
Consequently, certain key attributes may be significantly undervalued in the statistics, as the win rate is already too low to reflect any further decline.
Therefore, conclusions drawn from home-game performance are relatively more reliable.


一些属性会在客场有轻微的削弱。个别属性在主场、客场都表现很好。
但程度不明显,因此,我不打算具体分析。其中,AI分析显示在主场、客场都很好的是“工作投入”

Some attributes see a slight reduction when playing away from home.
Certain attributes perform well both at home and away.
However, the effect is not significant, so I do not intend to analyze it in detail.
Notably, AI analysis indicates that "Work Rate" performs well in both home and away matches.


6.主要表格,比较属性1->20的影响力
Main table, comparing the influence of attributes 1->20

80.webp

值得注意的细节
抗压能力排第 6,它是隐藏属性,和预判几乎完全并列,但游戏界面上看不到
稳定排第 14,也是隐藏属性。
大赛则完全不显著。在测试中,选用的是“联赛”,其在代码显示的比赛重要性是13。这个比赛重要性可能会影响“大赛“的效果

Notable details:
"抗压Pressure" ranks 6th; it is a hidden attribute—nearly tied with "Anticipation"—but is not visible in the game interface.
"稳定Consistency" ranks 14th; this is also a hidden attribute.
"大赛ImportantMatches" (Big Game Player) is entirely insignificant.The tests were conducted using "League" matches, which have an "match importance" value of 13 in the game's code. This "match importance" level may influence the effect of the "大赛ImportantMatches" attribute.

7.分支表格
Branch Table


81.webp


这是测试的主要成果。
由于未到达预定的40万场以上的测试样本数,部分数据会有一些统计误差噪声
This is the main result of the test.
As the target of over 400,000 test samples has not been reached, some of the data may contain statistical noise or errors.

从表格里非常容易看出,一些属性有不同的贡献分布。
It is very easy to see from the table that some attributes have different contribution distributions.


(7.1)
工作投入WorkRate、抗压Pressure
是非常典型的,他们只要≥4,就能享受绝大部分的胜率贡献。
又或者说,这3个属性只需要满足“4“点。更多的值影响相对很小。

Work Rate, Pressure, and Vision—these are classic examples.
As long as each of these attributes is ≥ 4, you gain the vast majority of the potential win-rate benefit.
In other words, you only need a value of 4 for these three attributes; anything higher has a negligible additional impact.

抗压 1→4 一段就拿下全部收益的 55%;工作投入更是 81%
Pressure (levels 1–4): the first level alone accounts for 55% of the total benefit, while for Work Rate, that figure reaches 81%.

(7.2)
越往上越值的属性
Attributes that become increasingly valuable at higher levels

弹跳JumpingReach、集中Concentration

弹跳 16→20 的 +7.0 是全部 170 个格子里的最大值之一,唯一“堆到 20 反而最赚”的属性。
The +7.0 increase in JumpingReach (from 16 to 20) is one of the highest values across all 170 slots—the only stat for which "stacking it up to 20" actually yields the greatest benefit.

(7.3)
全线高价值型
High-Value Across the Board

速度Pace
爆发Acceleration

唯一两个“无死区”的属性。速度的峰值在 8→12(+12.6),爆发在 8→12(+11.5)。
Possesses the unique distinction of having no "dead zones" in any attribute. Peak values for Pace (8→12, +12.6) and Acceleration (8→12, +11.5) are both exceptional.

(7.4)
属性强度分段(T0–T7)
Tiers (T0–T7)

T0 断层 peak
速度 Pace
爆发 Acceleration

T1 核心 core
耐力 Stamina
弹跳 Jumping Reach
预判 Anticipation
抗压 Pressure

T2 强 strong
强壮 Strength
集中 Concentration

T3 中 mid
盘带 Dribbling
射门 Finishing
平衡 Balance
工作投入 Work Rate

T4 弱 weak
灵活 Agility
稳定 Consistency
视野 Vision
远射 Long Shots

T5 边缘 Marginal
意志力 Determination
左脚 Left Foot
决断 Decisions
镇定 Composure
头球 Heading
盯人 Marking
大赛 Important Matches

T6 无影响 No impact
勇敢 Bravery
防守站位 Positioning
传球 Passing
抢断 Tackling
侵略性 Aggression
技术 Technique
团队合作 Teamwork
停球 First Touch

T7 负向 Negative
传中 Crossing
无球跑动 Off The Ball
才华 Flair

注意,我不使用xG(预期进球数)作为衡量标志,
因为有些属性不贡献xG,但能贡献胜率
Note that I do not use xG (expected goals) as a metric,
because certain attributes may not contribute to xG, yet they do contribute to the win probability.

测试设置的局限性:本次测试中玩家球员属性对英超球员处于劣势,
因此一些属性的重要幅度可能和“实力持平”、“实力优势”时不同。不过,我认为劣势时的属性更有价值。
Limitations of the test setup: In this test, the player's player attributes were at a disadvantage compared to Premier League players.
Therefore, the significance of some attributes may differ from when the players were "equal in strength" or "superior in strength."However, I believe attributes are more valuable when the player is at a disadvantage.


8.对手强度分析
Attribution Analysis of Opponent Strength


对手球队的实力是主要被属性决定的,还是由战术实力决定的?
Is opponent strength driven by player attributes or by tactical strength?


分析对手的强度是由球员属性造成的还是教练战术强度造成的
An analysis of whether opponent strength comes from player attributes or from manager tactics.

——结果显示在"对手强度 Opponent strength"表格中

— the results are shown in the "对手强度 Opponent strength" worksheet.

首先用两种方法交叉验证:相关性(19 支球队的自然差异)+ 对照实验(交换战术)
Two independent methods are used to cross-check each other: correlation (the natural variation across the 19 teams) and a controlled experiment (swapping tactics).


一:相关性(19 支球队)
Correlation (19 teams)
把对手的首发 CA、30 项属性总和,与「我们对其的主场胜率」做相关:
The opponents' starting-XI CA and the sum of their 30 attributes are correlated with our home win rate against them:

球员能力解释了约 79% 的对手强度差异,剩下 21% 才是战术 + 噪声。
Player quality explains about 79% of the variation between opponents; the remaining 21% is tactics plus noise.

二:战术交换实验(2.6 分钟,360 场)
Tactic-swap experiment (2.6 minutes, 360 matches)
把富勒姆和曼城的战术互换,看胜率怎么变。得到战术的效果约为阵容效果的 1/14 到 1/6。
The tactics of Fulham and Man City were swapped to see how the win rate changes. The effect of tactics comes out at roughly 1/14 to 1/6 of the effect of the squad.

结论:球员属性是主因,战术是次要因素
Conclusion: player attributes are the main driver; tactics are secondary.


三:进一步确认,把 19 队的战术全部替换成「曼城的战术」,看胜率方差缩小多少
Further confirmation: replace all 19 opponents' tactics with Man City's, and see how much the variance shrinks

① 战术解释了约 7–8% 的方差
① Tactics explain about 7–8% of the variance

去掉战术差异后,CA 的解释力从 79% 升到 86%。多出来的 7 个百分点,就是原来被「各队战术差异」占掉的份额。
Once the tactical differences are removed, CA's explanatory power rises from 79% to 86%. The extra seven percentage points are exactly the share that was previously occupied by "differences between teams' tactics".

这与实验二的结论一致:球员属性 ≈ 85%,教练战术 ≈ 8%。
This is consistent with experiment 2: player attributes ≈ 85%, manager tactics ≈ 8%.

② 各队自己的战术更强,更换了曼城的战术以后,各队变弱
② Each team's own tactic is stronger; after being forced onto Man City's tactic, the teams got weaker

统一成曼城战术后,我们对 19 队的平均胜率从 39.4% 升到 45.8%(+6.4pp)——说明曼城这套打法对我们是更有利的。
After switching everyone to Man City's tactic, our mean win rate against the 19 teams rises from 39.4% to 45.8% (+6.4 pp) — meaning City's approach is more favourable to us .

原因:可能是各队教练的战术是为他们的球员量身定制的。而在工具中,战术是根据每个球队和当前主教练的风格来生成的,强行更换到其他球队以后会有比如位置不合适的情况。
The reason: each manager's tactic is probably tailored to his own players. In this tool the tactic is generated from each club's squad and the current manager's style, so forcing it onto another club can produce mismatches such as unsuitable positions.

③ 战术的影响高度不均
③ The impact of tactics is highly uneven

对列表中的弱队(CA 130–145)影响最大(如果更换,−12% ~ −18% 胜率)
——他们的原主教练战术把有限的球员能力发挥到了极致,换成曼城的战术反而浪费了;
The weak teams on the list (CA 130–145) are affected most (switching costs them 12–18 pp of win rate) — their own manager's tactic squeezed the utmost out of limited players, so forcing City's tactic on them wastes what they have.

对强队(CA 152–166)几乎没影响,因为他们的球员本身就够强,什么战术都能打。
The strong teams (CA 152–166) are barely affected, because their players are good enough to play any system.

一句话总结
One-sentence summary

对于玩家的对手,对手球员质量决定对手胜负,对手教练战术只能多拿 5–8% 的胜率。
For a player’s opponents, the quality of the opposing players is the primary factor determining the outcome, while the opposing coach’s tactics can only contribute an additional 5–8% to their win probability.


9.位置熟练度
position familiarity


69.webp

战术424-442
在有球阶段,有2名边锋。在无球阶段,有2名边前卫。
Tactic: 424-442.
During the possession phase, there are 2 man , AML and AMR. During the off-possession phase, the AML and AMR transform into ML and MR

测试只改变这2个球员的位置熟练度
The test only changed the positions Familiarity levels of the 2 players.
70.webp


结论:
Conclusion,
71.webp

一个只有前锋或者中后卫位置熟练度的球员,
只要他们拿到“8点“位置熟练度,有球或者无球有一个就行,他们就能踢这个位置,不吃惩罚(这个和FM26不一样,要求很低)
A player who is only "ST" or only "DC" .
As long as there is proficiency in the 8-point AMR or AML or MR or ML positioning,They can basically play this role without being punished.
(This is completely different from FM26. FM26 has very strict requirements for this.)

8点的位置熟练度非常容易练,让球员在友谊赛或者比赛踢这个位置就行
And 8 points is actually relatively easy to growth (for young players aged 22 or younger) , by friendly matches or match

之前的测试表明:一个球员如果在19岁和19岁之前,练位置熟练度会容易得多。但是到了25岁开始就非常困难。最后期限是24岁
Previous tests have shown that: It is much easier for a player to develop proficiency in a specific position at age 19 or younger.
By the time they reach 25, it becomes extremely difficult.
The cutoff point is age 24.

73.webp

82.webp

83.webp


注意,在游戏里,熟练度5-8,会显示为“力不从心”。
为了保险,可以练到9点,会显示为“勉为其难”

Note that in the game, a proficiency level of 5–8 will be displayed as "Awkward".
To be on the safe side, you can raise it to 9, at which point it will be displayed as "Unconvincing" .
 
Last edited:
Great work again harvestgreen, big fan since FM24! Could you also run experiments with hidden attributes? Is Big match and Consistency still top? And are their delta comparable to pace accel?
 
When you say 'gears' do you mean that attribute value itself? 16 gears in anticipation = 16 anticipation as displayed in game?
 
Great work again harvestgreen, big fan since FM24! Could you also run experiments with hidden attributes? Is Big match and Consistency still top? And are their delta comparable to pace accel?
The program currently cannot effectively support the testing of hidden attributes, so these hidden attributes will be ignored for now.

Based on the results of the previous tests, it can be estimated that:
”
consistency ≈ important Matches ≈ 30% * pressure
pressure ≈ Work Rate
“
When you say 'gears' do you mean that attribute value itself? 16 gears in anticipation = 16 anticipation as displayed in game?
right. gears = attribute value itself
This is a machine translation, so there were some inaccuracies in the translation of certain words.
 
68.webp

position familiarityTactic: 424-442.
During the possession phase, there are 2 man , AML and AMR. During the off-possession phase, the AML and AMR transform into ML and MR

The test only changed the positions Familiarity levels of the 2 players.

Conclusion,
A player who is only "ST" or only "DC" .
As long as there is proficiency in the 8-point AMR or AML or MR or ML positioning,They can basically play this role without being punished.
(This is completely different from FM26. FM26 has very strict requirements for this.)

And 8 points is actually relatively easy to growth (for young players aged 22 or younger) , by friendly matches or match
 
Here is a genie scout file based on this, purity, and caiograco's combined testing. Did not include work rate.

 
Here is a genie scout file based on this, purity, and caiograco's combined testing. Did not include work rate.

It would be great to see their attribute weights adapted for FMST or FMSS, especially given how buggy GS is. Their FM27 findings seem incredibly promising and could be a real game-changer if the weights were accurately converted and properly implemented.
 
It would be great to see their attribute weights adapted for FMST or FMSS, especially given how buggy GS is. Their FM27 findings seem incredibly promising and could be a real game-changer if the weights were accurately converted and properly implemented.

ST: Pace 100, Acceleration 87, Jumping Reach 78, Finishing 41, Stamina 33,
Balance 31, Anticipation 31, Concentration 21, Strength 21, Dribbling 16,
Pressure 15

AM/W: Pace 100, Acceleration 80, Stamina 48, Anticipation 40, Strength 39,
Dribbling 30, Jumping Reach 29, Finishing 27, Concentration 25, Balance 17,
Long Shots 16

FB/WB: Pace 100, Acceleration 86, Stamina 54, Anticipation 39, Strength 29,
Jumping Reach 29, Dribbling 28, Concentration 26, Balance 23, Pressure 17

CB: Jumping Reach 100, Acceleration 95, Pace 75, Stamina 41, Anticipation 36,
Balance 32, Strength 25, Concentration 25, Pressure 16

DM: Acceleration 100, Pace 70, Stamina 57, Anticipation 39, Strength 37,
Dribbling 27, Jumping Reach 26, Pressure 24, Determination 19,
Imp. Matches 16

I removed work rate as you can just make sure it's the minimum that matters (around 10) and long shots from FB/WB which was around 15, but I didn't want it in there. There wasn't really any testing on CM's so I left it out, but I imagine it's similar to DM.
 
It would be great to see their attribute weights adapted for FMST or FMSS, especially given how buggy GS is. Their FM27 findings seem incredibly promising and could be a real game-changer if the weights were accurately converted and properly implemented.
You can do this in FMST by creating custom metrics, in the statistics section. Like the one below for a striker, you can probably mess around with some of the options to get a percentage as well.

ST(acceleration * 0.87 + pace * 1 + jumping_reach * 0.78 + stamina * 0.33 + anticipation * 0.31 + strength * 0.21 + balance * 0.31 + dribbling * 0.16 + concentration * 0.21 + pressure * 0.15 + finishing * 0.41)
AM/W(acceleration * 0.8 + pace * 1 + jumping_reach * 0.29 + stamina * 0.48 + anticipation * 0.4 + strength * 0.39 + balance * 0.17 + dribbling * 0.3 + concentration * 0.25 + finishing * 0.27 + long_shots * 0.16)
FB/WB(acceleration * 0.86 + pace * 1 + jumping_reach * 0.29 + stamina * 0.54 + anticipation * 0.39 + strength * 0.29 + balance * 0.23 + dribbling * 0.28 + concentration * 0.26 + pressure * 0.17)
CB(acceleration * 0.95 + pace * 0.75 + jumping_reach * 1 + stamina * 0.41 + anticipation * 0.36 + strength * 0.25 + balance * 0.32 + concentration * 0.25 + pressure * 0.16)
DM(acceleration * 1 + pace * 0.7 + jumping_reach * 0.26 + stamina * 0.57 + anticipation * 0.39 + strength * 0.37 + dribbling * 0.27 + pressure * 0.24 + determination * 0.19 + (important_matches * 0.16)


FMST.webp
 
Last edited:
Back
Top