Read on Twitter
“Pick Salah at home and Mane away” This statement seems to have become common consensus around large parts of FPL Twitter
Looking into the stats I found that there is no truth to this statement at all.
I wrote a short [THREAD] to explain
This theory seemingly originates from looking at goals scored at Anfield last PL season (19/20)
Salah 15 G
Mane 11 G
By looking at these stats Salah seems like the superior player at home
Salah 15 G Mane 11 GBy looking at these stats Salah seems like the superior player at home
Similarly, Mane looks like the best pick away from home
last PL season (19/20)
Salah 4 G
Mane 7 G
And by that a theory is born
“Pick Salah at home and Mane away”
last PL season (19/20)
Salah 4 G Mane 7 GAnd by that a theory is born
“Pick Salah at home and Mane away”
Then 20/21 opens with the following results
GW 1 LEEDS (H) Salah score 3 goals
GW 2 CHE (A): Mane score 2 goals
Perfectly in line with our theory.
The 19/20 pattern seems to predict this seasons outcome!
GW 1 LEEDS (H) Salah score 3 goals GW 2 CHE (A): Mane score 2 goalsPerfectly in line with our theory.
The 19/20 pattern seems to predict this seasons outcome!
Let us rewind back one year and look at the 18/19 PL season.
How well did the Salah and Mane's home/away record in 18/19 predict the subsequent season?
How well did the Salah and Mane's home/away record in 18/19 predict the subsequent season?
PL 2018/19 goals scored:
Away
Salah 9 G
Mane 4 G
Home
Salah 13 G
Mane 18 G
Mane actually scored 82% of his goals at home in 18/19!
Away
Salah 9 G Mane 4 GHome
Salah 13 G Mane 18 GMane actually scored 82% of his goals at home in 18/19!
If we were to make a theory before the 19/20 season it would’ve been:
“Pick Salah away and Mane at home”
Following this theory would be ill advised. You would end up with 15 goals in total, less than if you had just stuck with either one of them.
“Pick Salah away and Mane at home”Following this theory would be ill advised. You would end up with 15 goals in total, less than if you had just stuck with either one of them.
This doesn't bode well for our original theory
Why this difference from one season to another?
Did something change tactically that could cause this?
Maybe 18/19 failed to predict 19/20 because Klopp changed tactics to accommodate Manes strengths away from home?
Why this difference from one season to another?Did something change tactically that could cause this?
Maybe 18/19 failed to predict 19/20 because Klopp changed tactics to accommodate Manes strengths away from home?
Let's look at the timing of the away goals during the 19/20 season
Mane scored 4 goals in the first 7 games
Over the next 12 games Salah outscored Mane
Any tactical change was only effective in the first 7 games
Mane scored 4 goals in the first 7 games Over the next 12 games Salah outscored ManeAny tactical change was only effective in the first 7 games
A tactical change benefitting Mane should also be noticed in the expected stats for 19/20 (xG per 90)
Away
Salah: 0.57 xG
Mane: 0.35 xG
Turns out Salah actually performed better than Mane away from home!
Away
Salah: 0.57 xG Mane: 0.35 xGTurns out Salah actually performed better than Mane away from home!
This clearly demonstrate the dangers of looking at patterns emerging in small samples.
Looking at single season away/home performance looks to be a really bad idea.
When we see theories like this happen, we should always be critical and ask questions.
Looking at single season away/home performance looks to be a really bad idea.
When we see theories like this happen, we should always be critical and ask questions.
Then who is the best home/away player out of Salah and Mane?
Let’s increase the sample size and take a look at the averages over two years (18/19 and 19/20 seasons combined)
Since we are looking for Fantasy potential, let's add assists by looking at Goal Involvements (GI)
Let’s increase the sample size and take a look at the averages over two years (18/19 and 19/20 seasons combined)
Since we are looking for Fantasy potential, let's add assists by looking at Goal Involvements (GI)
18/19 and 19/20 combined (GI per 90)
Home
Salah 1.02 GI
Mane 0.95 GI
Away
Salah 0.69 GI
Mane 0.50 GI
Salah has a small edge at home, and even more so away.
Now let us check how the expected stats line up with the output
Home
Salah 1.02 GI Mane 0.95 GIAway
Salah 0.69 GI Mane 0.50 GISalah has a small edge at home, and even more so away.
Now let us check how the expected stats line up with the output
18/19 and 19/20 combined (xGI per 90)
Home
Salah 0.87 xGI
Mane 0.74 xGI
Away
Salah 0.72 xGI
Mane 0.52 xGI
Both players outscored their xG at home.
Salah performed better than Mane both home and away.
Home
Salah 0.87 xGI Mane 0.74 xGIAway
Salah 0.72 xGI Mane 0.52 xGIBoth players outscored their xG at home.
Salah performed better than Mane both home and away.
This makes sense as Salah outperformed Mane in general (disregarding home/away).
18/19 and 19/20 combined
Salah 0.80 xGI
Mane 0.64 xGI
Salah 0.87 GI
Mane 0.74 GI
18/19 and 19/20 combined
Salah 0.80 xGI Mane 0.64 xGI Salah 0.87 GI Mane 0.74 GI
Both players perform better at Anfield than on the road
18/19 and 19/20 combined
Salah had 60% of his GI’s at home
Mane had 66% of hi GI’s at home
This is likely correlated to Liverpool scoring 60% of their goals at home in the same period
18/19 and 19/20 combined
Salah had 60% of his GI’s at home Mane had 66% of hi GI’s at homeThis is likely correlated to Liverpool scoring 60% of their goals at home in the same period
CONCLUSION
The “Pick Salah at home and Mane away” theory is false
Neither player performs markedly better at home/away than the other
Both players align their home goals percentage to that of the team
The “Pick Salah at home and Mane away” theory is false Neither player performs markedly better at home/away than the other Both players align their home goals percentage to that of the team
If you liked this thread, you might also like the piece I wrote earlier this week: https://twitter.com/russpuckney/status/1308488258914004994?s=20
Source: http://FBref.com
