BladesPod: QPR analysis and playoff odds

All advertisments are hidden for logged in members, why not log in/register?

Beans

Beans

Active Member
Joined
Mar 21, 2017
Messages
2,461
Reaction score
14,687
Nothing else going on in the Championship today, is there?

Right! Blades Analytics and I recap the win over QPR, talk about the merits or otherwise of Gary Madine (come on, everyone else has created threads about Madine), and then discuss the roadmap to making the top six... and higher.



https://itunes.apple.com/gb/podcast/bladespod/id1309198119?mt=2

As always, cheers for listening and UTB!
 



I've got an exam tomorrow at the stadium of rust...

I'll have my blades top on, my lucky red socks and boxers, and this podcast in my ears on the way.
 
Another great listen Beans and Blades Analytics , cheers gents.

Quick question, what is the twitter tag of the Wednesday stats guy that you mentioned. Interested to see the table you referenced about 2nd half season stats.

Cheers
 
It’s so interesting how you use statistical data to support your opinions.

It kind of proves which teams are in their genuine position and which have been lucky/ unlucky (over performing/ under performing).
 
Sheffsteel said:
It’s so interesting how you use statistical data to support your opinions.

It kind of proves which teams are in their genuine position and which have been lucky/ unlucky (over performing/ under performing).
Click to expand...

Cheers - yeah, that's the plan ;)

On that note of teams overperforming, I did notice that Derby have massively dropped off in the last month or so. Infogol's expected table has them as the 17th best team - imagine where they'd be without their superstar loanees :rolleyes: Good news for Wednesday though as they're no longer the worst-ranked team in the league, having risen all the way up to 21st.
 
A question regarding xg and xga - are the numbers based just on shots taken or more than that? What I mean is, if a player is in a position to shoot with an xg of 0.2 but chooses to pass instead is that included in the team stats?

Also, does a team xg take account of which player is getting the chance? If Harry Wilson has a shot from outside the box, is the team xg increased by the same as if Richard Keogh does?!

Even more detailed - is it different if Wilson hits it with right or left foot? If not, then surely this is where stats can be a bit misleading - Wilson can get himself into loads of good shooting positions wth high xg, but if the defender knows him and forces him onto his right foot all the time then the true xg should be much lower.

Overall though I find the close similarity between actual tables and xg tables fascinating - it’s particularly noticeable that by now the top and bottom of actual and xg tables are much more similar than the middle - I suppose showing that consistency (good or bad) generally gets its appropriate reward.

Keep up the good work - it adds to the Magic
 



ucandomagic said:
A question regarding xg and xga - are the numbers based just on shots taken or more than that? What I mean is, if a player is in a position to shoot with an xg of 0.2 but chooses to pass instead is that included in the team stats?
Click to expand...

Good questions! Most xG models are shots only. Some models (think it might be Opta... can't remember - paging Blades Analytics) record "dangerous moments" or similar which take into account things like a cross that should have been tapped in but the striker somehow fails to make contact with the ball, so it doesn't technically end with a shot.

Also, does a team xg take account of which player is getting the chance? If Harry Wilson has a shot from outside the box, is the team xg increased by the same as if Richard Keogh does?!
Click to expand...

It doesn't - xG is essentially an average of a large number of players who have taken shots from that position. You can get quite a bit of value out of this because you can see which players outperform their xG and draw the conclusion that they're an excellent finisher (Sharp is an obvious example, but if you were a performance analyst / scout you could look at data for less-well-known leagues or players and immediately get an idea of who could be an above-average finisher).

On the flip-side with McGoldrick he was massively underperforming his xG about 6 weeks ago and some of us (cough) suggested that, while he's not an amazing finisher, he had been a bit unlucky and would probably regress to something closer to his xG over the rest of the season. Lo, his xG for the season is currently just over 12, and he now has 10 goals - much closer to what he "should" have scored.

With Wilson, I'm not sure if he's absurdly good at long range shots or just whether he's just "above average" and takes a load of them. S2 Stats on Twitter pointed this out the other day, Derby actually take more shots from *outside* the box than they do inside! And Wilson himself leads the league in shots outside the box.



Even more detailed - is it different if Wilson hits it with right or left foot? If not, then surely this is where stats can be a bit misleading - Wilson can get himself into loads of good shooting positions wth high xg, but if the defender knows him and forces him onto his right foot all the time then the true xg should be much lower.
Click to expand...

Yes, I believe most if not all xG models account for shot quality which take into account the body part used to shoot, as well as the type of pass/move that led to the chance.

Overall though I find the close similarity between actual tables and xg tables fascinating - it’s particularly noticeable that by now the top and bottom of actual and xg tables are much more similar than the middle - I suppose showing that consistency (good or bad) generally gets its appropriate reward.

Keep up the good work - it adds to the Magic
Click to expand...

Same! It's also quite gratifying that United have been one of the top 2-3 teams on the expected tables pretty much from September onwards, and that we've matched that level of performance with the results.
 
Beans said:
Good questions! Most xG models are shots only. Some models (think it might be Opta... can't remember - paging Blades Analytics) record "dangerous moments" or similar which take into account things like a cross that should have been tapped in but the striker somehow fails to make contact with the ball, so it doesn't technically end with a shot.



It doesn't - xG is essentially an average of a large number of players who have taken shots from that position. You can get quite a bit of value out of this because you can see which players outperform their xG and draw the conclusion that they're an excellent finisher (Sharp is an obvious example, but if you were a performance analyst / scout you could look at data for less-well-known leagues or players and immediately get an idea of who could be an above-average finisher).

On the flip-side with McGoldrick he was massively underperforming his xG about 6 weeks ago and some of us (cough) suggested that, while he's not an amazing finisher, he had been a bit unlucky and would probably regress to something closer to his xG over the rest of the season. Lo, his xG for the season is currently just over 12, and he now has 10 goals - much closer to what he "should" have scored.

With Wilson, I'm not sure if he's absurdly good at long range shots or just whether he's just "above average" and takes a load of them. S2 Stats on Twitter pointed this out the other day, Derby actually take more shots from *outside* the box than they do inside! And Wilson himself leads the league in shots outside the box.





Yes, I believe most if not all xG models account for shot quality which take into account the body part used to shoot, as well as the type of pass/move that led to the chance.



Same! It's also quite gratifying that United have been one of the top 2-3 teams on the expected tables pretty much from September onwards, and that we've matched that level of performance with the results.
Click to expand...


If it’s primarily area-based, it presumably doesn’t also take into account whether a player is offbalance, whether they have players between them and the goal when getting a shot off...or does it?

Incidentally, an excellent episode – I couldn’t have put the Madine discussion any better.
 
Great work as always. One of the highlights of the working week when these drop.
 
Great listen as always.

I agree that we are more likely than not to make the top 6 at this point. The main thing that might stop us is injuries. We have done very well at getting our first XI on the pitch this season, with hardly any disruptions for injuries or suspension. Hopefully we will stay lucky - although fatigue is also an issue.

My impression is that we have a smaller, and weaker, reserve squad than our rivals for a top 6 spot.
 
Revolution said:
Great listen as always.

I agree that we are more likely than not to make the top 6 at this point. The main thing that might stop us is injuries. We have done very well at getting our first XI on the pitch this season, with hardly any disruptions for injuries or suspension. Hopefully we will stay lucky - although fatigue is also an issue.

My impression is that we have a smaller, and weaker, reserve squad than our rivals for a top 6 spot.
Click to expand...

I think we have decent cover for every position except left-sided centre back and left wing back.

If something major were to happen to O’Connell or Stevens, we’re relying on Bryan and Johnson as their deputies.

I’ve not seen enough of them to write them off but it’s a definite risk to our promotion chances.

We could really do with an experienced , proven defender who can cover both positions. Like a left-footed Cranie.

That’s the biggest remaining priority in January as far as I’m concerned.
 
Last edited:
LoughboroBlade said:
If it’s primarily area-based, it presumably doesn’t also take into account whether a player is offbalance, whether they have players between them and the goal when getting a shot off...or does it?
Click to expand...

Taking into account whether a player was off-balance, hmm - sort of, with some models. Some of them track how the shot was created, eg did it come after a dribble by the player, or a through-ball, or is it a rebound off a blocked shot etc.

Same with defenders - some models account for defender and keeper positioning, how much pressure a player is under when he shoots, etc.

Just found this which gives a bit more detail on the factors involved: https://www.pinnacle.com/en/betting-articles/Soccer/expected-goals-model-analysis/MEP2N9VMG5CTW99D
 
You must log in or register to reply here.

Similar threads

Deadbat
United 2 Wednesday 1 - report/ratings
2
Replies
30
Views
3K
Camden Blade
C
ucandomagic
All Stoked Up - Matchday 12 Graphs, Charts and XG-raphs - Combined Posts
Replies
6
Views
904
bakeblade
B
ucandomagic
Matchday 15 Graphs, XG-raphs, Charts and Chat – After the SnortFest
Replies
8
Views
1K
ThirdGenBlade
T
ucandomagic
Matchday 14 Graphs, XGraphs and Porcine Forecast
Replies
6
Views
1K
Ldn_Rd_Blades
Ldn_Rd_Blades
Beans
BladesPod: Swindon 5-5 Sheffield United, 2015 playoff
Replies
18
Views
2K
Houston_Blade
Houston_Blade

All advertisments are hidden for logged in members, why not log in/register?

All advertisments are hidden for logged in members, why not log in/register?

Back
Top Bottom