Topic: The Alternative Premier League Table 2018-19

[drmick]

If we were only as good in the second half of the season, as Utd have been in this first half- we will finish on 85 points. That's mental.

[Mr_Shane]

Does it mean that since we have more 3 pointers left to play that we have some easier games in the long run than city??

indeed our end of the season run-in is relatively straightforward. southampton, chelsea, Cardiff Huddersfield Newcastle (Par 1), Wolves= last 6 games. Man City's run-in is against  teams they should beat as well, but they also got spurs and Burnley and Palace are their par 1s in their run-in

[rscanderlech]

If you believe the model, yes

I may be getting ahead of myself, but perhaps we can review at the end of the season whether it would be wise to tweak the model only very slightly to make the targeted point total maybe 93-95. But, as I say, this is a discussion for May.

[joe ®]

indeed our end of the season run-in is relatively straightforward. southampton, chelsea, Cardiff Huddersfield Newcastle (Par 1), Wolves= last 6 games. Man City's run-in is against  teams they should beat as well, but they also got spurs and Burnley and Palace are their par 1s in their run-in

Hopefully all of Southampton, Cardiff, Huddersfield and Newcastle are mathematically safe or relegated by the time we play them.

[Qston]

Hopefully all of Southampton, Cardiff, Huddersfield and Newcastle are mathematically safe or relegated by the time we play them.

That's unlikley to be honest

[Suareznumber7]

This graphic has been a great addition to the thread.  Looks like we are 5 wins and a draw away from reaching the goal for away points for the season. 

[RooiBefok]

APLT Full Table : Matchweek 17

[RooiBefok]

APLT Analysis & Extrapolations : Matchweek 17

[RooiBefok]

APLT Revolving Par : Matchweek 17

[RooiBefok]

LFC Current & Projected Form : Matchweek 17

[Hunts Cross]

Looking at the latest APLT graph, I was thinking ... If, at the start of the season, after a few ales and when I was in a really optimistic mood, I'd had a go at sketching out how I thought the graph would pan out to mid-season, in my wildest dreams I would not have imagined a spread like that. That is what it is ... unimaginable. That is what this team is delivering. +5 above City, +8 above 90-line, 14 above 4th place.

... still it will all likely crash and burn against Wolves next week /s.

[johnny74]

Is there a quick link to last season's final APLT table? Wondering what City finished with. They must have had +12?
And they were around +8 at this stage?

EDIT - found it. So they were at +8 and then about game 22 started to fall and hit +3 then from game 28ish went on a mad run and ended up at +10.

So, wondering how our form and their form are going to be this year. They are currently on +3 now just like they were at a later point last season.

I wonder if this season  the priority for them (if they have to prioritise resources) will be the CL. They have the league title, and of course they'd want to keep it, but I think given the choice of extra training and tactical preparation for a CL knockout game or for a PL game they might be favouring the former. Ditto with considerations of player fitness. Whereas I suspect that Klopp has already shown his hand here, preferring to play his strongest defence in the PL.

[Nessy76]

Is there a quick link to last season's final APLT table? Wondering what City finished with. They must have had +12?
And they were around +8 at this stage?

It's in the OP.



Finished on +10. 90+10=100.

[G Richards]

I think that historically the title winning trend line is spot on, but I suspect we are entering a new paradigm and agree with the need for a summer adjustment. Mind you, we're not at the half way mark yet, so there's a lot more data that needs to be gathered.

PS - I'm delighted with our comfortable win against Man Utd. I really thought they would come and spoil it for us, but we swatted them aside with great authority.

[Hymer Red]

It's in the OP.



Finished on +10. 90+10=100.

I see from game 17 last season only City finished with a higher position on the APLT that gives us a measure of how we need to keep going to win the title. They did have a dip around game 26-27 so if they do that again we can put them under some serious pressure. We finished 5 down on our game 17 total.

[latortuga]

I wonder if this season  the priority for them (if they have to prioritise resources) will be the CL. They have the league title, and of course they'd want to keep it, but I think given the choice of extra training and tactical preparation for a CL knockout game or for a PL game they might be favouring the former. Ditto with considerations of player fitness. Whereas I suspect that Klopp has already shown his hand here, preferring to play his strongest defence in the PL.

I see this point made a lot and it seems more hopeful than realistic.  Look at the graph of last season's APLT - aside from the derby collapse - they probably played their best football at the end of the season, much like we did. 

There's also a good chance we'll be right there with them in a semi final, so any small advantage would be nullified. 

[Brian Blessed]

Prof, you probably explained and I missed it, but on that third one, what do the blue lines indicate? I know combined they represent 90 points, just not sure why the of points isn't all on the aways.

[sminp]

Prof, you probably explained and I missed it, but on that third one, what do the blue lines indicate? I know combined they represent 90 points, just not sure why the of points isn't all on the aways.

I believe it’s more in keeping with what would be seen in real life. Whilst the APLT suggests a perfect home record in reality a team will drop a few points at home so the blue lines are actually more representative of the points totals that would be hit home and away

[Prof]

Prof, you probably explained and I missed it, but on that third one, what do the blue lines indicate? I know combined they represent 90 points, just not sure why the of points isn't all on the aways.

The blue line indicates the average points won at home and away based on a total of 90 points for league winners.  I did the calculation a couple of years ago, but it's still going to be the same ball park if not the same.  As a display, it actually has nothing to do with the aplt, but it seems to compliment it quite well.

If I get the time and inclination I might recalculate it and show my working for any of you who are interested.

The theory behind the aplt is based on the idea that winning at home is par, even though law of variance says that sometimes you won't, you can't reasonably predict which matches will be the ones you will drop points if you're going to win the league.  If I look at where the league winners do drop points at home, there's no real pattern as to who they are dropped against.

[Prof]

The theory behind the aplt is based on the idea that winning at home is par, even though law of variance says that sometimes you won't, you can't reasonably predict which matches will be the ones you will drop points if you're going to win the league.  If I look at where the league winners do drop points at home, there's no real pattern as to who they are dropped against.

Just looking at the last couple of seasons...

2017/18 - Man City drew two and lost one home game

They lost to Man Utd
Drew to Everton
Drew to Huddersfield

So they didn't drop home points to any decent team 

2016/17 - Chelsea lost two home games

Lost to Liverpool
Lost to Crystal Palace

2015/16 - Leicester won the league with 81 points, drawing six and losing one.  Very much an outlier

2014/15 - Chelsea drew four home games

Drew to Man City
Drew to Liverpool
Drew to Burnley
Drew to Southampton

[zabadoh]

APLT Analysis & Extrapolations : Matchweek 17

I love your table, and how it shows why we're so far ahead of City in the APLT:  We're +5 in the 7 x Par 1 matches that we've played.

Granted that City have only played 5 x Par 1 matches, so they have their chances to catch up with us in that regard.

[latortuga]

Just looking at the last couple of seasons...

2017/18 - Man City drew two and lost one home game

They lost to Man Utd
Drew to Everton
Drew to Huddersfield

So they didn't drop home points to any decent team 

2016/17 - Chelsea lost two home games

Lost to Liverpool
Lost to Crystal Palace

2015/16 - Leicester won the league with 81 points, drawing six and losing one.  Very much an outlier

2014/15 - Chelsea drew four home games

Drew to Man City
Drew to Liverpool
Drew to Burnley
Drew to Southampton

The predictability isn't in who you drop points to but when you drop those points that seems consistent.

If you look at the above points dropped, most of those games were near the start or end of the season.

[Brian Blessed]

I believe it’s more in keeping with what would be seen in real life. Whilst the APLT suggests a perfect home record in reality a team will drop a few points at home so the blue lines are actually more representative of the points totals that would be hit home and away

The blue line indicates the average points won at home and away based on a total of 90 points for league winners.  I did the calculation a couple of years ago, but it's still going to be the same ball park if not the same.  As a display, it actually has nothing to do with the aplt, but it seems to compliment it quite well.

If I get the time and inclination I might recalculate it and show my working for any of you who are interested.

The theory behind the aplt is based on the idea that winning at home is par, even though law of variance says that sometimes you won't, you can't reasonably predict which matches will be the ones you will drop points if you're going to win the league.  If I look at where the league winners do drop points at home, there's no real pattern as to who they are dropped against.

Thank you both

[latortuga]

In my mind we're +10 now because Wolves is for sure a par 1 game.

[Blinis]

In my mind we're +10 now because Wolves is for sure a par 1 game.

So Man City is at +5 then

[Ycuzz]

*Patiently awaits happy graffikksses*

[latortuga]

So Man City is at +5 then

Oh yeah...damn it! 

[McAlvis]

I feel Wolves away being a Par 3 balances out Burnley away being a Par 1, us being on +8 is now a little bit more 'correct' than it was before the match.

It is a weakness of the APLT that it doesn't take this season's form into account, but I don't think it is something you can fix without taking away two of the APLT's strengths, its simplicity and that it matches the real table after 38 games.

[TheShanklyGates]

Probably the most difficult of the away par 3s out of the way. Great stuff.

[Always_A_Red]

I feel Wolves away being a Par 3 balances out Burnley away being a Par 1, us being on +8 is now a little bit more 'correct' than it was before the match.

It is a weakness of the APLT that it doesn't take this season's form into account, but I don't think it is something you can fix without taking away two of the APLT's strengths, its simplicity and that it matches the real table after 38 games.

But its the same for all teams so there's no difference.

[drmick]

But its the same for all teams so there's no difference.

I'd say this season Burnley away will be a much harder fixture at the end of the season than at the beginning.

[latortuga]

But its the same for all teams so there's no difference.

So what your saying is you might as well just follow the real table and not the APLT?  I think his point has some validity.

[Prof]

The trouble is, the pars are set before the season and are based on objective info.  I also don't want to readjust the pars later in season because it'd mean the table earlier in the season is irrelevant when you're looking at it.

There will always be better and worse times to play teams.  Injuries, suspensions, fixture congestion, time of kick off etc. It comes down to how much you want to factor these variables in.

[drmick]

We have this debate every year. Didn't somebody even have a 5 game form rolling APLT at one stage as a suggestion?

The beauty of the APLT is its simplicity, and it's objectivity. Burnley being unworthy of Par1 applies to all the teams.

I keep going on about this, apologies, but the loss of that secure mid-table group of teams (like West Brom, Stoke, Southampton) leaving only Everton, and maybe Leicester is what has harmed the APLT.

That loss is probably a reflection on what has happened in the league though (kinda big 6 vs everyone else). If it continues in real life maybe the APLT has to adapt too?

[IanZG]

Just a quick question, is it possible to display how many points a team is expected to win until the end of the season? I'd expect we have 98 (our current standing in the APL table) - 48 (our current actual points) = 50 points to go, while City if they win have 93 - 47 = 46 points, which would mean our fixtures until the end of the season should be easier, right? Does any of this make sense?

[McAlvis]

The trouble is, the pars are set before the season and are based on objective info.  I also don't want to readjust the pars later in season because it'd mean the table earlier in the season is irrelevant when you're looking at it.

There will always be better and worse times to play teams.  Injuries, suspensions, fixture congestion, time of kick off etc. It comes down to how much you want to factor these variables in.

I think you are entirely correct not to do any readjusting for form over the season. The APLT is a nice compromise in that it provides some weighting to account for the difficulty of the fixtures a team has played which helps give a better insight into how the title race is going than the real table often does; without getting bogged down trying to account for form and the numerous other factors that affect the difficulty of each match. The success of this thread over so many seasons is a testament to APLT's usefulness.

As drmick wrote, the issue I raised gets discussed every season and I didn't mean to bring it up again, I was just adding to the discussion latortuga and Blinis were having about Wolves away being too difficult to be considered a par 3 by highlighting that by that logic we have already had an 'easy' par 1 away to Burnley to balance it out and that this is just something we have to accept as part of the APLT.

[G Richards]

The APLT is a great way of weighting the various fixtures over a season, and as such it gives a very useful indicator of progress. I love it!

Personally I think it’s impossible to track all the variables from week to week, as you would need to write a new statistical model every week, and even then it would be insufficient. Mourinho sacked? Burnley hungover in Europe? Southampton new manager? 48 hours recovery v 72 hours? Not to mention the impact of injuries and form. And even if you had the perfect model, very weak teams and players do things they aren’t supposed to do, for good and bad.

As I type Andros Townsend just hit a 30+ yard volley in against Man City. That’s not the sort of thing you see all the time! Man City are down 2-1 at halftime, and while I still expect them to win, they have work to do and a straightforward home game has all of a sudden become a puzzle for them to solve.

I see the APLT as an overview of where you stand in the grand scheme of things, as the season unfolds, and a really like it as is. I haven’t tracked it year on year, but presumably any discrepancies with the actual table are ironed out as the season progresses.

[RogerTheRed]

Looking forward to seeing this later today 👍

[Brian Blessed]

Looking forward to seeing this later today 👍

Not until after Spurs play tomorrow.

Let’s not get ahead of ourselves

[Redwhiteandnotblue]

Not until after Spurs play tomorrow.

Let’s not get ahead of ourselves

We're not letting anyone get ahead of us...