[Albion] To Avoid Hürzeler Bashing - The Staying Up and Europe Tracker - season 2024-25 Game 38 Update

[Kalimantan Gull]

I disagree, @Giraffe first post says the "easiest way" (at least on paper) to get the points is to win the home games against the bottom 10 and draw the away games and so that is the basis for allocating points on the tracker. BUT it does not say they are the predicted results- I dont think anyone would predict we wouldnt get any away wins, or only beat the bottom ten at home-indeed would anyone ever actually predict we would beat any of the bottom 3 at home let alone all of the bottom 10

Its simply a (very good) guide as to overall progress not a wanna be crystal ball into actual results

Well exactly, it's a predictor. If it was a pure tracker it would have 1.6 points per game plugged in and it would track our progress against that, but instead it attempts to pick the easiest games and assume we win points in those, and the hardest games and predict we don't win points in those, and then tracks our progress against those predictions.

Ergo, it's a predictor. It's not a GOOD predictor, but it's still a predictor.

 

[GT49er]

Best thread @Giraffe

My view is that the Euro target should be increased (and relegation floor reduced).

The last two years have seen a real drop off in points from the bottom three, 62 Vs 87 from 9 years prior. Doesn't feel like that'll change much next season. Also the top four have picked up a bit less, 314 vs 328. So there's about 28 points for the other 13 teams to fight over, averages two each. So to get into Europe it's getting harder, say +2 points.

It's not clear what league position gets Europe so it's a punt but sixth feels like should get you there. That's averaged 65 last two years and 63 for last eleven years.

Could argue it's more like 63-65 anything below is a bonus.

Relegation clearly isn't 40 but it's been the target for so long and doesn't feel like we need to focus on that for next years (fingers crossed)

I can see where you're coming from (and I did myself half-jokingly ask whether the Europe figure should be increased, bearing in mind that we'd passed the Europe target but still weren't in Europe) but realistically, no. The current formula (and remember it is a formula, not a prediction) is an excellent guide to our progress, and long may it remain!
And a ban, please, for the next person to ask, "Why is it predicting a draw against xxxxxxxx, we should be beating them, shirley?"

 

[Falmer Flutter ©]

Not sure why the European predictor needs changing as it was pretty much bang on wasn't it? The only reason we didn't qualify for Europe was because Palace won the cup, which is an anomaly and something that can't be factored in I suppose. Without that I'd be knee-deep in a spreadsheet with how to get to Jagiellonia Białystok.

 

[Audax]

Well exactly, it's a predictor. If it was a pure tracker it would have 1.6 points per game plugged in and it would track our progress against that, but instead it attempts to pick the easiest games and assume we win points in those, and the hardest games and predict we don't win points in those, and then tracks our progress against those predictions.

Ergo, it's a predictor. It's not a GOOD predictor, but it's still a predictor.

Ummm, it literally is a points predictor and claims to be...

It's not a results predictor at all. It's a baseline-setting formula that just happens to use hypothetical match results to set the baseline. And it's perfectly good for what the OP says it's for. It's not meant to accurately predict individual match results. It's not even meant to predict where Brighton will finish at the end of the season.

The idea is for Brighton to get *better* results than the tracker baselines across the season on average, because that will guarantee we avoid relegation / qualify for Europe (ok, on the latter, the baseline needs a tweak...).

It is completely and utterly erroneous to go back and look at all of the results and say "they only matched the 'prediction' X times." The point is, the only time we want them to 'match' is when a win is 'baselined'. Every other game (draw/loss baselined) we want a better result than the baseline asks for. That's how we get ahead of the tracker.

It's really not that hard to understand. Not a prediction tool. Just a "baseline setting" tracker tool.

 

[Kalimantan Gull]

It's not a results predictor at all. It's a baseline-setting formula that just happens to use hypothetical match results to set the baseline. And it's perfectly good for what the OP says it's for. It's not meant to accurately predict individual match results. It's not even meant to predict where Brighton will finish at the end of the season.

The idea is for Brighton to get *better* results than the tracker baselines across the season on average, because that will guarantee we avoid relegation / qualify for Europe (ok, on the latter, the baseline needs a tweak...).

It is completely and utterly erroneous to go back and look at all of the results and say "they only matched the 'prediction' X times." The point is, the only time we want them to 'match' is when a win is 'baselined'. Every other game (draw/loss baselined) we want a better result than the baseline asks for. That's how we get ahead of the tracker.

It's really not that hard to understand. Not a prediction tool. Just a "baseline setting" tracker tool.

Fine, yes, it is a tracker against a hypothetical baseline, but we can acknowledge that this baseline is in turn based on predicting the hardest and easiest games, and thus a points 'expectation' from each. It *could* be made better though, as its kind of pointless to set a baseline which isn't especially valid. Otherwise a tracker which predicts a flat 1.65 points per game would be just as useful, if admittedly less fun. If we're going to assign probalistic outcomes to each match, then we might as well try and make these as accurate as possible.

As for the 'matched the prediction' - I agree we want to do better than the baseline, yes. But overall the Europe tracker aimed for 60 points, we got 61, but only 10 of the 38 results 'matched'. I don't think its erroneous to point this out. Surely we can agree that if we can refine the algorithm to have more 'matching results', then we can better track our progress to this target? Otherwise how do we know if we are really on track

God I feel like I've gone right down a rabbit hole here. Love the tracker Mr Giraffe

 

[Gabbiano]

Fine, yes, it is a tracker against a hypothetical baseline, but we can acknowledge that this baseline is in turn based on predicting the hardest and easiest games, and thus a points 'expectation' from each. It *could* be made better though, as its kind of pointless to set a baseline which isn't especially valid. Otherwise a tracker which predicts a flat 1.65 points per game would be just as useful, if admittedly less fun. If we're going to assign probalistic outcomes to each match, then we might as well try and make these as accurate as possible.

As for the 'matched the prediction' - I agree we want to do better than the baseline, yes. But overall the Europe tracker aimed for 60 points, we got 61, but only 10 of the 38 results 'matched'. I don't think its erroneous to point this out. Surely we can agree that if we can refine the algorithm to have more 'matching results', then we can better track our progress to this target? Otherwise how do we know if we are really on track

God I feel like I've gone right down a rabbit hole here. Love the tracker Mr Giraffe

Yes, and also just because it's a "baseline" doesn't mean that the "baseline" is as representative as it could be.

The tracker is very unevenly weighted towards home fixtures. Yes it all levels out over the course of a whole season, but the point is to track progress game by game, so having the fixture weighting so wonky does have an impact.

But all in all I suppose this is just minor griping, for one of my favourite long running threads.

 

[Audax]

Fine, yes, it is a tracker against a hypothetical baseline, but we can acknowledge that this baseline is in turn based on predicting the hardest and easiest games, and thus a points 'expectation' from each. It *could* be made better though, as its kind of pointless to set a baseline which isn't especially valid. Otherwise a tracker which predicts a flat 1.65 points per game would be just as useful, if admittedly less fun. If we're going to assign probalistic outcomes to each match, then we might as well try and make these as accurate as possible.

As for the 'matched the prediction' - I agree we want to do better than the baseline, yes. But overall the Europe tracker aimed for 60 points, we got 61, but only 10 of the 38 results 'matched'. I don't think its erroneous to point this out. Surely we can agree that if we can refine the algorithm to have more 'matching results', then we can better track our progress to this target? Otherwise how do we know if we are really on track

God I feel like I've gone right down a rabbit hole here. Love the tracker Mr Giraffe

I see where you're coming from, but I disagree on the basis that it's impossible to gain 1.65 points from a match and I like using real possible results for the tracker.

The logic for how wins/losses/draws are assigned to matches is easily understandable and generally sound - ok, some teams perform better at home than away (and vice versa), and some teams perform better against top sides than bottom sides (*cough*). But it does work IMO, and if we're to keep the ability to compare across the league as a whole then I think trying to get picky about individual matches or using a ppg baseline takes us away from the value adds that @Giraffe has been able to use it for; eg actually making predictions about season-end (using tracker performance) across the entire league, and even at times picking out some teams whose league position was out of kilter with everyone else because of the run of fixtures they'd had.

 

[Bodian]

Fine, yes, it is a tracker against a hypothetical baseline, but we can acknowledge that this baseline is in turn based on predicting the hardest and easiest games, and thus a points 'expectation' from each. It *could* be made better though, as its kind of pointless to set a baseline which isn't especially valid. Otherwise a tracker which predicts a flat 1.65 points per game would be just as useful, if admittedly less fun.

I guess a flat rise of 1.65PPG wouldn't really show where we are in comparison with the actual games. For instance, we could have a far harder first half of the season (on paper) as so we'd all be up in arms that we were 'behind the PPG tracker', when actually we could be bang on 'expected' points.

Writing that, that's the key isn't it - it's a comparison again 'expected' points. Not 'what we need' points.

I like it the way it is. But maybe do need a few extra for the Europe tracker.

 

[Kalimantan Gull]

I see where you're coming from, but I disagree on the basis that it's impossible to gain 1.65 points from a match and I like using real possible results for the tracker.

The logic for how wins/losses/draws are assigned to matches is easily understandable and generally sound - ok, some teams perform better at home than away (and vice versa), and some teams perform better against top sides than bottom sides (*cough*). But it does work IMO, and if we're to keep the ability to compare across the league as a whole then I think trying to get picky about individual matches or using a ppg baseline takes us away from the value adds that @Giraffe has been able to use it for; eg actually making predictions about season-end (using tracker performance) across the entire league, and even at times picking out some teams whose league position was out of kilter with everyone else because of the run of fixtures they'd had.

You could have probablistic 2.3 points, 0.5 points, etc., against every fixture, but yeah, not real.

Or, you could generate those and then round to the nearest result.

My biggest gripes are having no away wins, and when a club has a particularly bad or good season they fall in arguably the wrong place. We will have 4 point expectations against Spurs and Man U next year, like we did against Chelsea this year - but perhaps a more historical look at their performance would be better than just the previous season?

f*** it, not my tracker

 

[Machiavelli]

Really interesting discussion from @Kalimantan Gull and @Audax, enlightening us on predictions, modelling and related issues.
I'll just add my twopenneth worth and it's that the tracker that @Giraffe has devised is simple to understand (ditto with the Europe tracker, a supplement that demonstrates how well we've done over the past seven seasons). That brings enormous benefits.
Both trackers are extremely useful guides for us to track (the clue is in the name) our progress as the season develops against clearly relayed criteria. The fact that you won't need anywhere near 40 points to survive these days, or that Europe often requires a few extra than the 60 points is less important than the simplicity and clarity these trackers provide.

 

[Kalimantan Gull]

I guess a flat rise of 1.65PPG wouldn't really show where we are in comparison with the actual games. For instance, we could have a far harder first half of the season (on paper) as so we'd all be up in arms that we were 'behind the PPG tracker', when actually we could be bang on 'expected' points.

Writing that, that's the key isn't it - it's a comparison again 'expected' points. Not 'what we need' points.

I like it the way it is. But maybe do need a few extra for the Europe tracker.

Yeah, but thats only a valid complaint if we actually matched the results for the 'easier' or 'harder' games. Maybe its just Brighton that refuses to do that

 

[A1X]

Not doing a relegation tracker? What next, orange on the "Teams That Can't Catch Us" list?

 

[Balders]

Just my opinion - it's a tracker against points gained at any given point, using a simplistic formula and just a general overview as to where we are. If we accept that 40pts sees us safe from relegation and 60pts sees us with a chance of Europe (given the Club we currently are) then I really can't see any reason to change it to some kind of more advanced predictor, which again will never be accurate.

Set Europe to 65pts and 61pts is actually enough in any given season and it's still inaccurate! Set relegation to 35pts and in any given season that might not be enough

The other salient point is this year Europe went down to 8 places, as opposed to 7 places which is the EPL's allotment of places - it only went down to 8th due to the performance of English Clubs in Europe this year and that was only known a few weeks ago - so do we assume that the same number of points will apply whether 7 places or 8 places are available.....

40pts these days is plenty to avoid relegation, 60pts to get Europe is debatable but that's the joy of the tracker!

 

[Questions]

He‘s desperate to, but doesn’t want to tempt fate.

Do both then I say

 

[The Optimist]

Do a European qualification tracker next season. Nothing else….

As the title is always ‘To avoid [manager’s name] bashing’ and getting just 40 points next season would absolutely not avoid Hurzeler bashing this actually makes sense.

 

[The Optimist]

I’ve been wondering whether a comparison to last season tracker would be interesting or not. I.e. next season the tracker would predict we get exactly the same results against each team as we did this season with the promoted teams replacing the relegated teams.

 

[Perkino]

I’ve been wondering whether a comparison to last season tracker would be interesting or not. I.e. next season the tracker would predict we get exactly the same results against each team as we did this season with the promoted teams replacing the relegated teams.

I like this idea as an adjustment or additional feature. However, the tracker is a brilliant thread and the one that brings clarity to a good/bad weekend of fixtures

 

[Giraffe]

Some interesting points of discussion here.

The thing I’d add (which was the original rationale behind the tracker) is that it is unbiased. We could all predict ourselves but we’d all have in built opinions and biases.

This is an unbiased tracker based on the previous seasons league standings. Of course it’s not perfect, but it’s a tracker, a baseline. It’s like the control in a scientific experiment. All things being normal is what we are comparing against and I think there is huge value in a consistent season on season approach.

That’s my post season thoughts for now.

 

[AZ Gull]

Some interesting points of discussion here.

The thing I’d add (which was the original rationale behind the tracker) is that it is unbiased. We could all predict ourselves but we’d all have in built opinions and biases.

This is an unbiased tracker based on the previous seasons league standings. Of course it’s not perfect, but it’s a tracker, a baseline. It’s like the control in a scientific experiment. All things being normal is what we are comparing against and I think there is huge value in a consistent season on season approach.

That’s my post season thoughts for now.

Exactly.

Let those clamouring for changes start their own trackers, and see how popular they are (spoiler: they won't be).

 

[The Optimist]

Some interesting points of discussion here.

The thing I’d add (which was the original rationale behind the tracker) is that it is unbiased. We could all predict ourselves but we’d all have in built opinions and biases.

This is an unbiased tracker based on the previous seasons league standings. Of course it’s not perfect, but it’s a tracker, a baseline. It’s like the control in a scientific experiment. All things being normal is what we are comparing against and I think there is huge value in a consistent season on season approach.

That’s my post season thoughts for now.

I hope you don’t take the discussions as a sign that your trackers aren’t appreciated. They are one of the best threads on NSC.

I always like the league tables as a helpful guide to who’s had an easy / hard run and might be in a slightly false position.