League Two Predictions (1 Viewer)

[kapowaz]

So, inspired by thread about The Run In I decided to have a go at creating a tool for predicting the outcome of the league table based on the remaining fixtures, which you can find here: http://predictor.ccfc.nu/. It now has a go at predicting the actual results for the remainder of the season, and tabulating the results.

As before, you can enter your own predictions, but if you scroll down and click the auto-predict tab, which will have a guess as to the likely results based on the past 10 match form for each team. Of course, these guesses are extremely naïve, so it’s likely to be quite wrong, but you can switch back to the Make predictions tab and edit the guesses yourself to see how you like it.

Obviously take it with big pinch of salt, but the predictor currently thinks we’ll finish 4th:

The way it works out how matches might resolve is pretty bad and I’d love to have something more accurate that takes into account goals scored/conceded etc. but this is pretty deep statistical analysis. It’s a little bit feedback-loopy right now (so if you are on a losing streak it makes your form worse so you’re more likely to lose again… see Grimsby and Port Vale). An earlier version that was even more naïve actually had us finishing 3rd, but it also had Luton in 5th so I knew it wasn’t quite right!

 

---

[Otis]

So, inspired by thread about The Run In I decided to have a go at creating a tool for predicting the outcome of the league table based on the remaining fixtures, which you can find here: http://predictor.ccfc.nu/. It now has a go at predicting the actual results for the remainder of the season, and tabulating the results.

As before, you can enter your own predictions, but if you scroll down and click the auto-predict tab, which will have a guess as to the likely results based on the past 10 match form for each team. Of course, these guesses are extremely naïve, so it’s likely to be quite wrong, but you can switch back to the Make predictions tab and edit the guesses yourself to see how you like it.

Obviously take it with big pinch of salt, but the predictor currently thinks we’ll finish 4th:

The way it works out how matches might resolve is pretty bad and I’d love to have something more accurate that takes into account goals scored/conceded etc. but this is pretty deep statistical analysis. It’s a little bit feedback-loopy right now (so if you are on a losing streak it makes your form worse so you’re more likely to lose again… see Grimsby and Port Vale). An earlier version that was even more naïve actually had us finishing 3rd, but it also had Luton in 5th so I knew it wasn’t quite right!

Hmmm. You have Grimsby losing every single match and not even picking up a single point. Unlikely I would say.

Be very, very surprised as well if Luton only manage to get 7 points from 7 games.

 

[Liquid Gold]

I clicked random and it has us winning 2-0 tomorrow. Squeaky is only looking 1 game at a time so I will too.

 

[speedie87]

I did the predictor on 1 st feb and had us finishing 6th on 79 points.
Top 7 were

Luton
Stanley
Wycombe
Mansfield
Exter
Us
Notts county

 

[Sky Blue Harry H]

If we finish in the play off places, I'd like Exeter and Notts County to be in with us. Not sure about the other team, though I'd prefer Mansfield to either Swindon or Lincoln.

 

[kapowaz]

Hmmm. You have Grimsby losing every single match and not even picking up a single point. Unlikely I would say.

Be very, very surprised as well if Luton only manage to get 7 points from 7 games.

Yeah, like I say the predictor is extremely naïve. I’d be interested in experimenting with some more realistic predictions, but this is basically trying to do the kind of stuff bookmakers do to set odds, which is extremely non-trivial statistics. This is just a fun little project really

Try monkeying around with the values and you should hopefully get something realistic based on your own biases.

 

[Otis]

Yeah, like I say the predictor is extremely naïve. I’d be interested in experimenting with some more realistic predictions, but this is basically trying to do the kind of stuff bookmakers do to set odds, which is extremely non-trivial statistics. This is just a fun little project really

Try monkeying around with the values and you should hopefully get something realistic based on your own biases.

What I tend to do and what makes it a little more accurate, though not at all perfect, is to also look at individual teams and trends and not just at each fixture.

So easy to put Accrington v Grimsby - win, Accrington v Yeovil - win, Accrington v Morecambe - win etc. On paper it looks the most likely outcome.

However teams tend not to win every game and likewise, others, not lose every game, so you do have to factor that in and make allowances, and also too predict some surprises, because there is nearly always at least one match every week where a bit of a shock is pulled.

 

[ceetee]

Look at the goal differences, you won't be far out

 

[Johhny Blue]

I predict the predictor will be wrong

 

[kapowaz]

I predict your prediction that the predictor will be wrong will be right!

 

[kapowaz]

Had a go at manually predicting every result. I’m sure this will be equally wrong, but it was a fun exercise:

I tried mixing up the results for the top 8 or so a bit and it seems we’re almost always going to come out around 3-4th provided we have a good run. It’s really down to how much Wycombe messes up…

 

[kapowaz]

I’ve updated the predictor with today’s results. Should make guessing the final games easier as there are fewer and fewer games to play — I suspect with 4-5 matches left we’ll have some pretty accurate guesses out there.

 

[Otis]

I’ve updated the predictor with today’s results. Should make guessing the final games easier as there are fewer and fewer games to play — I suspect with 4-5 matches left we’ll have some pretty accurate guesses out there.

Grimsby got a point.

Just saying, like.

 

[Astute]

Grimsby got a point.

Just saying, like.

My wife doesn't listen to me when I have a point either.

 

[kapowaz]

Grimsby got a point.

What’s your point?

 

[Otis]

5 and half inches with a slight kink in it.

 

[wingy]

Had a go at manually predicting every result. I’m sure this will be equally wrong, but it was a fun exercise:

I tried mixing up the results for the top 8 or so a bit and it seems we’re almost always going to come out around 3-4th provided we have a good run. It’s really down to how much Wycombe messes up…

Had a go at manually predicting every result. I’m sure this will be equally wrong, but it was a fun exercise:

I tried mixing up the results for the top 8 or so a bit and it seems we’re almost always going to come out around 3-4th provided we have a good run. It’s really down to how much Wycombe messes up…

I'd gladly take that scenario.
Just seven wins from seven it is then. B-)

 

[mds]

If we finish in the play off places, I'd like Exeter and Notts County to be in with us. Not sure about the other team, though I'd prefer Mansfield to either Swindon or Lincoln.

Much prefer autos but if we dont quite make it there isnt a team that on our day we cant beat, id be more worried about Exeter than any of the others as they have managed to do the double over us something i dont think any of the other playoff contenders have managed. We cant be far from top in the top 8 v top 8, way too many tops.

 

[Otis]

Much prefer autos but if we dont quite make it there isnt a team that on our day we cant beat, id be more worried about Exeter than any of the others as they have managed to do the double over us something i dont think any of the other playoff contenders have managed. We cant be far from top in the top 8 v top 8, way too many tops.

Errr? Eh? What? The double?

You obviously don't remember this then.

Saturday 23rd September

- Coventry City 2-0 Exeter City

 

[mds]

Errr? Eh? What? The double?

You obviously don't remember this then.

Saturday 23rd September

- Coventry City 2-0 Exeter City

Late night brain fart, beer brain For some reason i thought they beat us twice 1-0

 

[Irish Sky Blue]

Yeah, like I say the predictor is extremely naïve. I’d be interested in experimenting with some more realistic predictions, but this is basically trying to do the kind of stuff bookmakers do to set odds, which is extremely non-trivial statistics. This is just a fun little project really

Try monkeying around with the values and you should hopefully get something realistic based on your own biases.

What do the bookmakers say? I had ten pounds on us to win the league and ten on us to get promoted too. I think the first one has gone.

 

[Mucca Mad Boys]

Had a go at manually predicting every result. I’m sure this will be equally wrong, but it was a fun exercise:

I tried mixing up the results for the top 8 or so a bit and it seems we’re almost always going to come out around 3-4th provided we have a good run. It’s really down to how much Wycombe messes up…

This was a fun tool to play about with the results. Football is obviously unpredictable so it’s just that, fun.

However, you have our GD at +31?! You’re quite the optimist! That’s more or less winning every game 2-0 from now!

I think we’ll finish on 80-84pts at the minute. 5 wins and 2 draws has on 82pts. I think 83-84pts gives us a good chance of promotion.

Credit to your predictor, I genuinely think Luton could fall out of the automatic places — their form has been terrible of late. They have 6 games to sort it out, but I’d like to see them falter because it gives us a lifeline if Wycombe outperform us. GD is an obvious concern in an head to head though.