2020/21 - The Run In and Top 4 chances - an Ongoing Analysis

[mozartfox]

Screw all these projections. Just need to take care of our performances and results and we will be there in 4th , come final whistle.

 

New Injuries, Players returning from Injury, Match Officials and VAR are massive variables ignored in all these percentages.

[Chrysalis]

23 hours ago, davieG said:

The Independent's view.

 

https://www.independent.co.uk/sport/football/premier-league/premier-league-top-four-fixtures-man-utd-liverpool-chelsea-leicester-b1817494.html

Really hoping arsenal do us a favour, west ham have a favourable run in.

[HereBeFoxes]

4 hours ago, Vardinio'sCat said:

 

Rhythm

And come on you Blues!!!

[HereBeFoxes]

Some more output from my (yes, flawed) monte carlo simulation; points needed for (at least) 4th.  Eg 68 points will get you 4th (or higher) in 84% of the simulation runs.  Remember - it's just a bit of fun (but these are real actual theoretical possibilies from possible combinations of real world results - ie it could happen that 58 points is enough for 4th.  Don't lump on it though).

 

Points    Frequency % (cumulative)
 58        0.04%
 59        0.2%
 60        1%
 61        2%
 62        6%
 63        13%
 64        25%
 65        40%
 66        56%
 67        71%
 68        84%
 69        92%
 70        96%
 71        99%
 72        99.8%
 73        99.96%
 74        99.99%
 75        100.00%

[HereBeFoxes]

And the same again for 3rd:

 

Points    Frequency % (cumulative)
62       0.04%
63       0.5%
64       2%
65       6%
66       13%
67       24%
68       39%
69       56%
70       70%
71       83%
72       92%
73       96%
74       99%
75       99%
76       99.8%
77       99.96%
78       100.00%

[Ric Flair]

1 hour ago, Deeg67 said:

66 points going into the final three matches is the magic number for me.  Get there and I think we’re in.  That means 10 in the next 6, which really should be doable even if we lose to Man City.

I think 12 from the next 6 is achievable, possibly even 13-15. If we get to that then we will be about there. It would be so good to go in to those 3 games with the confidence to just attack all 3 and pick them off.

[coolhandfox]

I think we should be looking upwards our target should be 2nd.

 

We have pretty much matched Man U pace for pace.

 

Forget 3rd, Forget 4th.

 

Runners up, here we come..........

 

 

[Les-TA-Jon]

32 minutes ago, HereBeFoxes said:

Some more output from my (yes, flawed) monte carlo simulation; points needed for (at least) 4th.  Eg 68 points will get you 4th (or higher) in 84% of the simulation runs.  Remember - it's just a bit of fun (but these are real actual theoretical possibilies from possible combinations of real world results - ie it could happen that 58 points is enough for 4th.  Don't lump on it though).

 

Points    Frequency % (cumulative)
 58        0.04%
 59        0.2%
 60        1%
 61        2%
 62        6%
 63        13%
 64        25%
 65        40%
 66        56%
 67        71%
 68        84%
 69        92%
 70        96%
 71        99%
 72        99.8%
 73        99.96%
 74        99.99%
 75        100.00%

 

31 minutes ago, HereBeFoxes said:

And the same again for 3rd:

 

Points    Frequency % (cumulative)
62       0.04%
63       0.5%
64       2%
65       6%
66       13%
67       24%
68       39%
69       56%
70       70%
71       83%
72       92%
73       96%
74       99%
75       99%
76       99.8%
77       99.96%
78       100.00%

Out of the loop here - what is the Monte Carlo method? And what do all these numbers mean? 

[HereBeFoxes]

29 minutes ago, Les-TA-Jon said:

 

 

Quote

Out of the loop here - what is the Monte Carlo method? And what do all these numbers mean? 

Sorry - Monte Carlo is a method of simulating and quantifying the possible range of outcomes from a series of probabilistic events; in this case each remaining match in the Premier League can either be a home win, draw, or away win.  One assigns a probability to each of these outcomes, and then runs the remaining 92 matches (in this case) 10,000 times, with the result of each match randomly determined according to the assigned probabilities.  You then get 10,000 final tables, each one of which is a real possible outcome.

 

These numbers are the percentage of times in the 10,000 final tables that the listed number of points is enough to get 4th or 3rd place (at least); eg for 3rd - there was no final table where 61 points was enough to get 3rd or higher, whereas 70 points got 3rd or higher in 70% of the final tables.

 

Edit to add:  You can also look at it the other way round: in 30% of cases 70 points is not enough to get you 3rd place.

Edited 17 March 2021 by HereBeFoxes

[Les-TA-Jon]

48 minutes ago, HereBeFoxes said:

 

Sorry - Monte Carlo is a method of simulating and quantifying the possible range of outcomes from a series of probabilistic events; in this case each remaining match in the Premier League can either be a home win, draw, or away win.  One assigns a probability to each of these outcomes, and then runs the remaining 92 matches (in this case) 10,000 times, with the result of each match randomly determined according to the assigned probabilities.  You then get 10,000 final tables, each one of which is a real possible outcome.

 

These numbers are the percentage of times in the 10,000 final tables that the listed number of points is enough to get 4th or 3rd place (at least); eg for 3rd - there was no final table where 61 points was enough to get 3rd or higher, whereas 70 points got 3rd or higher in 70% of the final tables.

 

Edit to add:  You can also look at it the other way round: in 30% of cases 70 points is not enough to get you 3rd place.

OK thanks for explaining. So my understanding is that you personally looked at each of the remaining 92 fixtures and then (arbitrarily?) assigned a probability to home win, draw, away win? Or do the probabilities come from somewhere else? 

[HereBeFoxes]

3 minutes ago, Les-TA-Jon said:

OK thanks for explaining. So my understanding is that you personally looked at each of the remaining 92 fixtures and then (arbitrarily?) assigned a probability to home win, draw, away win? Or do the probabilities come from somewhere else? 

Yes - for the run those tables come from not entirely arbitrarily - they are based on the performance of each team over the last 8 matches - one can put any probabilities one wants, in, I'm just lazy and there are 90+ of them so I make an algorithm to look at the WLD probabilities for each team in a tie and calculate some liklihoods, quite simplistically I admit.  An example, Villa-Spurs it calculates as 33% chance home win, 25% draw, 42% away win.  Doesn't seem insane, but really who knows.

 

If I could get the probabilities from somewhere else (although obviously they're all going to be arbitrary to some extent, as they are prediciting the future) I could look at the outcome set for that too.  Someone suggested betfair as a source, but I don't know how to do that...

[Les-TA-Jon]

1 hour ago, HereBeFoxes said:

Yes - for the run those tables come from not entirely arbitrarily - they are based on the performance of each team over the last 8 matches - one can put any probabilities one wants, in, I'm just lazy and there are 90+ of them so I make an algorithm to look at the WLD probabilities for each team in a tie and calculate some liklihoods, quite simplistically I admit.  An example, Villa-Spurs it calculates as 33% chance home win, 25% draw, 42% away win.  Doesn't seem insane, but really who knows.

 

If I could get the probabilities from somewhere else (although obviously they're all going to be arbitrary to some extent, as they are prediciting the future) I could look at the outcome set for that too.  Someone suggested betfair as a source, but I don't know how to do that...

But where are those numbers coming from? Just curious. The whole system seems to be built on those, so where they come from/how they're calculated seems pertinent. 

[Koke]

Problem is if we lose to Man City and West Ham, which is entirely possible, the insecurities will be back. Not losing to West Ham is so vital.

[Ric Flair]

9 hours ago, Koke said:

Problem is if we lose to Man City and West Ham, which is entirely possible, the insecurities will be back. Not losing to West Ham is so vital.

I'm telling you now mate, we will not lose to West Ham. It will be vintage Leicester. I'm not even sure Man City beat us either, think it will be a draw.

[Foxes96]

Conscious it still seems like a long shot but anyone got the odds on Liverpool and Arsenal winning in Europe and finishing outside the top 4, such that 4th place wouldn’t get champions league? 
 

Both got reasonably favourable draws it seems...

 

Didn’t count Chelsea because I can see them getting top 4 now, although still a possibility they could win the CL and finish 5th or below. 

[Simmo86]

Wouldn't say Liverpool's is favourable, likely going to have to beat Real Madrid, Chelsea and either Man City or Bayern to win it.

 

Fairplay to them if they do!

[maddisonenthusiast]

Our next three games (https://www.wheresthematch.com/Football/Leicester-City.asp) are crucial, think winning the FA Cup will really help us cement our position in the countries elite! We should be taking it serious and I think Rogers will field a strong team for that now we don't have the Europa League to worry about. 

 

The Man City game will be tough, however, if we can sneak a point from that we will be well on our way to champions league football next season. 

 

West Ham is going to a close one, but fancy us to win against them and then we play West Brom, which should be pretty routine.

 

Overall, I think these next three games will shape our season. If we lose both of the league games, we could be chasing the likes of Spurs & Liverpool.

[majaco]

On 17/03/2021 at 09:25, HereBeFoxes said:

Some more output from my (yes, flawed) monte carlo simulation; points needed for (at least) 4th.  Eg 68 points will get you 4th (or higher) in 84% of the simulation runs.  Remember - it's just a bit of fun (but these are real actual theoretical possibilies from possible combinations of real world results - ie it could happen that 58 points is enough for 4th.  Don't lump on it though).

 

Points    Frequency % (cumulative)
 58        0.04%
 59        0.2%
 60        1%
 61        2%
 62        6%
 63        13%
 64        25%
 65        40%
 66        56%
 67        71%
 68        84%
 69        92%
 70        96%
 71        99%
 72        99.8%
 73        99.96%
 74        99.99%
 75        100.00%

Great stats again.  Thank you.

 

At the start of the season, I thought 70 points would be good progress.  Looking at these stats, it should do the trick.

[majaco]

On 17/03/2021 at 11:27, HereBeFoxes said:

Yes - for the run those tables come from not entirely arbitrarily - they are based on the performance of each team over the last 8 matches - one can put any probabilities one wants, in, I'm just lazy and there are 90+ of them so I make an algorithm to look at the WLD probabilities for each team in a tie and calculate some liklihoods, quite simplistically I admit.  An example, Villa-Spurs it calculates as 33% chance home win, 25% draw, 42% away win.  Doesn't seem insane, but really who knows.

 

If I could get the probabilities from somewhere else (although obviously they're all going to be arbitrary to some extent, as they are prediciting the future) I could look at the outcome set for that too.  Someone suggested betfair as a source, but I don't know how to do that...

Your system seems good to me.  It may have shortcomings but it seems really considered.   I have decided it is biblical.  If you are wrong, you deserve a harder to e on this forum than was given to Ben Chilwell or Demarai Gray.

[Les-TA-Jon]

UPDATE March 22nd 

 

I've changed some things around. Removed Man City - because they are an irrelevance. 

 

Previously we were looking a historic 4th place finishes. But of course that doesn't give a 'true' reflection of what was required for 4th place, because there could be a big gap. So I looked at '5th place + 1 point' (which doesn't account for Goal Difference being enough to finish 4th, but screw that ). So the new columns marked "****" are for that. 

 

Historical 4th place finishes: Lowest 60, Average 68, Highest 79

Historical 5th place + 1 finishes: Lowest 57, Average 65, Highest 76

 

The gap between our current form/points per game and what we need to finish on 65-68 points is very promising. We can can drop down to 1 point per game. 

 

Chelsea and West Ham need to maintain form. Everton and Spurs need a big uptick. Liverpool, Arsenal and Villa need to do something crazy.

 

Points Required

 

Droppable Points

 

Fixtures

 

 

Edited 22 March 2021 by Les-TA-Jon

[Ric Flair]

I wonder when the West Brom game will get rearranged for? Could it actually be brought forward between the Man City and West Ham games?

 

Either way, we just have to keep our performance levels and confidence high and we'll keep pushing on.

[HighPeakFox]

Caught between being delighted that a number of players get some time to recover and recuperate, and worried that the break has the potential to disrupt matters.

 

I think on balance, the rest time wins, for me.

[foxfanazer]

Just now, HighPeakFox said:

Caught between being delighted that a number of players get some time to recover and recuperate, and worried that the break has the potential to disrupt matters.

 

I think on balance, the rest time wins, for me.

Absolutely! Get a well earned rest and then get ourselves geared up for Man City. Might work out well to stop their momentum also

[Les-TA-Jon]

@Ric Flair I was still fiddling with it. Full update now posted

 

[HighPeakFox]

3 minutes ago, foxfanazer said:

Absolutely! Get a well earned rest and then get ourselves geared up for Man City. Might work out well to stop their momentum also

I think yesterday was enormous in so many ways - I think BR will want to get it right against MC too, not regard it as a free hit.