By Tony Attwood
Last season Arsenal had no less than five bad runs during the course of the campaign.
The first I am sure you will remember because it was national headlines – losing the first three league games against Brentford, Chelse and Manchester City without scoring a goal. Ahead of that run Jonathan Liew was at his sarcastic worst suggesting “Arsenal fans need Mikel Arteta to work out what he is trying to do.” Readers of Untold knew exactly what he was doing, because we had seen the sensational turnaround in the last two thirds of the previous campaign, but that extraordinary run was unmentionable in the media and so the pundits could make no use of that data.
So the opening three league games (the league cup match in the midst of these being ignored by the media because Arsenal won 6-0) were talked about until November when Arsenal had a run of three defeats in four games and a new story was used.
But it was the same old Arsenal, as it was again in January with three defeats and two draws in five matches, seeing the club be knocked out of both cups in the space of two weeks.
The third such run was in March which saw four defeats and one win in the Premier League between 16 March and 16 April. And to round it all off consecutive defeats without scoring in either, against Tottenham Hotspur and Newcastle United to ensure that we didn’t get into the Champions League.
In between those runs the club did rather well, but of course it was the negative runs which needed to be removed from the itinerary. The runs in which Arsenal went 10 undefeated (September – November), the five consecutive wins in which Arsenal scored 19 goals in December, the five consecutive league wins in February/March, and the four consecutive wins in April / May were not seen as signs of where the club was going, because they were hardly mentioned.
But they did show that Arsenal could make things work well. All that was needed was to remove the blips, and that of course is exactly what has been done for most of this season. And now that we have had that first blip – a run of three defeats and a draw in which Arsenal scored just two goals – the club seems to have come through it again.
Indeed, once again the blip was four matches long, but the difference is that this season, it has been the only four-match blip so far. If we can carry on the rest of the campaign without having any more of these, we will be ok.
The near-universal media prediction for this season, before the campaign began, was for a top four of Manchester City, Liverpool, Chelsa and Tottenham with Arsenal lurking in fifth to make it seven consecutive years without Champions League football.
That was interesting, because the last two long runs without top four finishes were six seasons between 1973/4 and 1978/9 – and that run included consecutive seasons in which the club came 16th and 17th.
The last time we had a run worse than six seasons without a top four finish was 1959/60 for 11 seasons. We ended that run in 1970/71 with the club’s first league and cup double. During that era Arsenal finished seasons in 10th, 11th, 12th (twice – one of those being the season prior to the first double), 13th (twice) and 14th.
So for this dip Arsenal have not sunk anywhere near as far – the two eighth place finishes obviously being the low point.
But what of the four teams that the media near-universally predicted would finish above Arsenal this season?
Here is the current table as of this morning. The final column shows how many points the predicted clubs have achieved above (+) or below (-) the predicted level. No predictions were made across the mainstream for Newcastle, Fulham, Brighton and Brentford, although there was some talk of Brentford going down.
| Team | P | W | D | L | Pts | Predicted Pos | Points Difference | |
|---|---|---|---|---|---|---|---|---|
| 1 | Arsenal | 24 | 18 | 3 | 3 | 57 | +16 | |
| 2 | Manchester City | 25 | 17 | 4 | 4 | 55 | 1 | -2 |
| 3 | Manchester United | 24 | 15 | 4 | 5 | 49 | +10 | |
| 4 | Tottenham Hotspur | 24 | 13 | 3 | 8 | 42 | 4 | 0 |
| 5 | Newcastle United | 23 | 10 | 11 | 2 | 41 | ||
| 6 | Fulham | 25 | 11 | 6 | 8 | 39 | ||
| 7 | Liverpool | 23 | 10 | 6 | 7 | 36 | 2 | -19 |
| 8 | Brighton and Hove Albion | 22 | 10 | 5 | 7 | 35 | ||
| 9 | Brentford | 23 | 8 | 11 | 4 | 35 | +14 | |
| 10 | Chelsea | 23 | 8 | 7 | 8 | 31 | 3 | -18 |
When one thinks about it, this table shows some pretty awful predictions. As things stand only one club was predicted successfully – Tottenham. The national media universally thought that Chelsea and Liverpool would have 18 or 19 points more than they have got, and that Arsenal would have at least 16 points fewer than they currently have.
And the fact is, that if in any other business, a whole range of people who are considered experts, (or who indeed were using supercomputers to get their predictions) these people would be sacked, and quite possibly the organisation would be so discredited they would go out of business.
But not in football. In football it seems that these ludicrous failures mean one should keep doing the same, and keep alleging the use of supercomputers!
If these pundits had any moral fibre at all they would resign. But they don’t, so they don’t.
The thing is with computers, as we all know, is they are only as good as the information you put in. So the predictions they made for this season, as for every other season they have claimed to use a ‘supercomputer, all depend entirely on what they put in. So what do they put in, that is the question?
Well as I suggested the other day, I suspect very little. My guess would be they look at the average points per game of last season………….and that’s it.
I mean what else? I don’t know? Someone may be able to tell me because I would be interested to know.
Lets just take Arsenal in isolation and have a look at some of the details that Untold and it’s regulars found, analysed, and drew conclusions on for this coming season, that I don’t believe for one second was inputted in to this mystical ‘supercomputer’.
-The opening three defeats of 2021/22.
Rather than the rabid negativity of the media, followed by way too many Arsenal fans, Untold looked at those 3 matches with a clear head and concluded it was nothing like the catastrophe it was being made out to be. We all drew that conclusion on the basis that we had so many players out through either injury or COVID we were always going to have a tough start. We also faced a newly promoted side away from home, never easy, followed by 2 of the favourites for the title that season.
-New players, new tactics
Arteta was in the process of implementing an entirely new approach of minimal tackling in order to counteract PIGMOL’s crusade of yellow carding our players for breathing on our opponents.
-Last two thirds of the season.
Despite that pointless opening and difficult pre Christmas Arsenal went on to play at a level over the final two thirds of the season that would have them comfortably in the top 4.
-Poor finish
Again, despite the disappointment, a level headed appraisal of the run-in would of realised, that we had yet again the core of our defence out injured.
Other factors that we at Untold took in to account when doing analysis.
-Moving on older players not up to the required standards anymore.
-Moving on possibly disruptive influences.
-Recruiting very good players. (As usual dismissed by the media as second rate or rejects)
-Our very young core of players getting a year older, a year wiser a year better.
These are just a few of the nuances, unique to Arsenal, that Untold could see and take into account when predicting how we may perform this season.
Does anyone think for one second that any of those factors are inputted into a computer? And if they were how would a ‘machine’ interpret something so subjective as, for example, taking Aubamayang out of the dressing room? Or putting Zinchenco in it?
Take Liverpool.
How could a computer see basically the same 11 players performing not just 1 level, but 4 or 5 levels bellow last season? I doubt any of us saw that coming?
Multiply that by similar nuanced events at every club and you can see why this ‘supercomputer’ nonsense is just that, utter nonsense.
Basically all they do is input what happened last year and surprise surprise it suggests the same will happen again this year! To be fair I think that’s why most people take these ‘supercomputers’ predictions with an extremely large pinch of salt.
When it gets annoying is when we see people, like Liew, being sarcastic just because that’s what he’s like when it comes to Arsenal.
When it gets annoying is when reporters like Amy Lawrence, a supposed Arsenal fan, take a set of statistics simply to bash Arsenal, but failsto use any context, or make any comparisons.
In all honesty, do I expect a mere ‘supercomputer’ to match the analytical skills, and sheer determination of Untold when it comes to researching, analysing and predicting all things Arsenal? No I don’t.
But do I expect the ‘media’ to be similarly analytical? Diligent? Fair? when it comes to Arsenal? Yes I do. Will it ever happen? No.
Webb shd discipline or sack refs who have underperformed
Saka was denied a penalty.Martineeli had studs on his shi n.
If var can pick out Whitesfoul on the goalie,why can’t they pinpoint these incidents to the ref?
It seems there is one law for Arsenal to be penalised compared to other teams.
And don’t forget the studs on our Jap defender. How can’t var see this as a blatant foul designed to hurt or cause serious injury.
Over yo you,
Mr.Webb
Computers ? Guys stop repeating this myth.
They use excel for the most of them…..and the rest go to the pub, get enebriated as they discuss what to publish and then do it.
And as everyone can read, Artificial Intelligence is algorithms that are supposed to mimick how a brain ‘thinks’. So it needs data to be fed, then rules (like what is a win, a tie, a loss, etc etc.) and relationships between the 3.
And AI shows that it is nothing more then a mirror of the thinking/behaviour of the people who feed it all that. As most of the deadwood press has some grudge against Arsenal or an automatic anti-Arsenal mindset, nothing can make them expect something positive.
Chris
But it isn’t a myth.
“Top spot goes to Arsenal – just as BETSiE predicted before they landed the deadline day signing of Jorginho”.
They do use a supercomputer. I’ve had a look in to this and all these Supercomputer predictions basically come for ones used in the betting industry. They use xg and xa amongst other things, to predict out comes in football matches, and extrapolate it across the season, or rest of the season depending on when it’s done.
So it seems it is a genuine claim when they say “supercomputer predicts’ but that doesn’t of course mean it is accurate, because it depends what data you put in it, and if all you are putting in is a teams xg, xa, or points per game etc, it’s bound to predict that what has been happening will continue to happen.
New signings fitting in. Finding form. Not fitting in. Disrupting the team. Loss of form. Gained experience. Injuries. Suspensions. Luck. VAR. I would hazard a guess none of this goes in to the computer making it’s predictions, well if nothing else, predictable.
So your premise that “…..AI is nothing more then a mirror of the thinking/behaviour of the people who feed…..” is exactly right.
If you put in just raw data about what has been happening and alter nothing, then by definition all it can predict is that the same will continue to happen.
All those, ‘reasons’ for our poor results, all those positives we at Untold noticed, and I mentioned above, as much as they are obvious to us, are never the less all subjective?
Injuries at the beggining. A reason or indicative of a small squad?
Poor first third of season. Adapting to new tactics? Not good enough?
Good final two thirds or others not performing?
Injury hit run in or typical capitulation?
You know what I think, but it is still subjective and is never going to be fed in to a computer that basically works on 2things, 0 or 1. Yes or no. It simply cannot do subjectivity.
Never the less, and back to the point of this.
It does seem that supercomputers are indeed used to predict football outcomes, but they are using them to predict something that is so subjective as make any conclusions they draw utterly pointless.
I think our issue, well it’s certainly my issue, is with the people that ‘subjectively’ cant see what’s staring them in the face.
It’s called bias.
Nitram,
fully agree on all…except supercomputer… unles you consider any computer you buy off the shelf a supercomputer compared to the Apollo 11 computer who took astronauts to the moon and back.
AI is biased, predictions are biased (like horse racing), the ‘predictor’ just wanting to make headlines and go with the flow, please the crowd and feed it with what makes clicks and thus money.
And the argument ‘supercomputer’ enables them to plead ‘not guilty’ and put the blame of all the crap they write on a machine instead of standing up for what they wish for.
Ah and yes, no supercomputer CAN predict PGMOL incompetency, incoherence and lunacy. And as no one – except here on UA and starting with a few other blogs wants to address the issue….
Chris
To be honest I haven’t got a clue what designates a ‘supercomputer’ as opposed to a ‘computer’. I just googled around and it seems there is no ‘definitive’ parameter between the 2.
I suspect it will be down to how many gazillion giga mega bits or bites, but as I say nothing definitive. Wiki said there are approximately 500 of the buggers hanging around so you would of thought they could of found one that works wouldn’t you!!
Well, I really doubt it…and it would not take a supercomputer to spit out statistics…just a good old excel spreadsheet….
Then again, maybe some idiot taped a ‘Supercomputer’ label on a computer in their office and thus they are stating the truth…’supercomputer’ gave the predictions…
Anyway. trust me (i’ve been working in the field since the old TRS-80…their crap writings could not finance them a supercomputer…. their stats can be spit out by any pentium computer running an old Access database and an old excel spreadsheet…
Chris
I have absolute no doubt what you say is true, but never the less they call the BETSiE computer ‘super’.
Why I don’t know. What the parameters are for such a description I don’t know. As you have worked in the field you have a much better chance of knowing where the line between the 2 lies?
Either way, as you suggest, I suspect the amount of data that this machine is given to deal with wouldn’t stretch an Abacus.
From the Scottish Sun (I wasn’t aware the was such a thing, from my visits in another life).
“Boffins and brainiacs over at bettingexpert have compiled BETSiE, a supercomputer tasked with predicting what is most likely to happen over the rest of the season.”
One does not compile a supercomputer.
Supercomputers are built. Programs may be compiled.
seismic
These people are so stupid I doubt they could turn it on, let alone program the damn thing.
How would AI predict scores and ties properly if the programmer is biased and slanted . Or at worse, just stupid
But by AI standards , The Spuds would be filling their , at present empty Trophy room ( a misnomer , I’m sure !) to the brim !
How many years ago , and how many managers ago did they last win a proper trophy ?