- Not what we wanted but not as bad as the Germans predicted
- Arsenal back at the top and have a clear plan for the future
By Tony Attwood
Amidst the general football chatter of the day CNN have moved away from their normal position of being rather a serious news organisation and have dived into the depths with the headline, “A supercomputer is tracking one of the most exciting English Premier League title races in years and with three clubs involved, separated by a matter of points, it keeps spitting out different predictions as it analyzes thousands of simulations.”
Now there are a couple of problems there. First I don’t believe anyone would have a) the money nor b) the nerve to start getting a supercomputer to analyse football matters. Such super computers as there are, are generally rather occupied with tracking what Russia and China are doing with their nuclear missiles, helping protect fishermen with minute by minute weather reports, and monitoring nuclear power stations so we might be given a bit of notice before they blow up.
The simple fact that no supercomputer is involved is revealed by the secondary fact that the article doesn’t tell us which super computer is involved. The Wiki article on supercomputers in the UK has under 30 such machines in its list.
Now the story has been put out by Opta, yet a search of their site again doesn’t tell us which supercomputer is involved but does suggest that, “With all three teams now having played 30 matches, the Opta supercomputer is giving Liverpool a 45% chance of winning the league, Manchester City 33.6% and Arsenal 21.4%.”
I would give the chance of an actual supercomputer doing that modelling as around 0.0%.
So why do CNN produce such an article which does not contain the key details (which supercomputer, and who owns it) when actually telling us which supercomputer and where it is, would give a spot of credibility to their story?
Probably because it is just a simple story that can be knocked out without any cost to anyone, by anyone. In short, lazy journalism.
CNN get a free article, and Opta gets their name bandied around in association with a bit of kit that everyone thinks must be pretty nifty. And the whole process becomes rather undermined when we get to the point that says, “In truth, after recent performances, it’s hard to see any of the top three dropping points until now and the end of the season.”
Apart from the grammatical slip (“until now” presumably should be “between now”) this notion that no one is going to drop points allows us to write an end ofseason table now, and it reads…
| Pos | Team | Pld | W | D | L | Pts |
|---|---|---|---|---|---|---|
| 1 | Arsenal | 38 | 29 | 5 | 4 | 92 |
| 2 | Liverpool | 38 | 28 | 8 | 2 | 92 |
| 3 | Manchester City | 38 | 28 | 7 | 3 | 91 |
But then, turning what they allege is a session with one of the most powerful computers on the planet, on its head they say, “But, if history tells us anything, they’ll be multiple twists and turns during the run-in.”
So the computer will be wrong. Which is worrying when it is normally used to decide if to let the nuclear bombs go, and if so where they should go.
The article is on more interesting ground when it says, “The North London club seemingly threw away the league title last season, letting slip a healthy lead at the top of the table as Manchester City cruised past them.
They are presumably talking about the eight game run from 9 April to 20 May in which Arsenal won just two games, drew three and lost two. Those two losses are indeed of interest for they were a 4-1 defeat to Manchester City and a 3-0 defeat to Brighton..
But to return to the table above, what we see is that the top two would each have 92 points if all remaining games are won, and so everything depends on goal difference.
At the moment Arsenal have a nine-goal lead when it comes to goal difference, but of course across the seven remaining games this could be knocked back.
If current trends continue Arsenal will end up with 92 goals scored and 29 against. giving a goal difference of +63. But of course Arsenal could themselves slip up and if Arsenal won each of their remaining games by one goal they would then have a final goal difference of +60. That would mean outdoing Arsenal’s goal difference by two or three goals in each and every game remaining. Arsenal win 1-0, Liverpool must win by 3-0 or 4-0. Game after game.
Obviously it is more likely that one of the teams will drop points, but it does show that although the alleged supercomputer puts Arsenal with only a tiny chance of getting the title, a more reasoned approach without an unidentified fantasy computer, still puts Arsenal in the driving seat.
But meanwhile I wonder. Why can’t Opta either tell us which supercomputer they managed to nick some time on, or that they simply used a calculator and some guesswork?
Let them tell we stand no chance. In Dutch we have an idiom saying: when two dogs are fighting for a bone, a third one runs away with it. So let Arsenal just be the third dog in this….
They flip flop every bloody week.Absolute tosh.
Firstly on whether they actually use a Super Computer or not.
This is from OPTA’s web site
“With the help of our AI-powered supercomputer, we provide our Premier League match predictions for every match in each game week”.
So, effectively they are claiming they actually OWN a ‘super computer’. Which is perfectly possible from my research.
There is plenty of information regarding buying, owning and running a Super Computer out there. This is what I found at Quora which you can access here:
https://www.quora.com/Would-it-be-possible-to-buy-a-supercomputer-Why-or-why-not
To the question “Is it possible to buy a supercomputer”? a techy replied as follows:
Yes. Absolutely.
Is it possible for YOU to buy a supercomputer?
I have no idea.
Do you have the space, and power? (most importantly power)
Do you understand how supercomputers work? (that’s as true for classic vector machines like Cray’s as it for modern HPC multi-node systems)
Do you know people who can actually sell you one?
Do you have cash?
If the answers are yes… then you can do it.
Then we had this:
“The Supercomputer of 20 or 15 years ago is the desktop of today. In my 50+ years of computing I have seen 6 orders of magnitude in CPU performance increase. From my experience your personal computer of today out performances all the supercomputer that were 15 to 20 years ago, or less.”
So it seems to me it is all a matter of classification relating to age and state of the art technology.
So perhaps the question is not, are they using a Super Computer, but rather, what is a ‘super computer’?
As such, I get the impression they are using computers that under some definition or other could be defined as a ‘super computer’. As to whether it is a ‘state of the art’ version similar to those used to save us from nuclear Armageddon is a different matter all together.
But in any case that is all utterly irrelevant because what a computer, super or otherwise, tells us is totally dependent on what is programmed in.
Just as a for example, how does the programmer define what is an easy fixture and what is a difficult fixture? It is utterly subjective and computers hate subjectivity.
It is not a fact that Aston Villa at home is an easy game. It is not a fact it is a difficult game. So how on earth does a programmer input the objective difficulty of the match? He simply cannot. All he can do is ‘subjectively’ apply a rating of difficulty between 1 and 10, because that’s what computers deal with, numbers. Turn his subjectivity in to an objective number.
So that matches difficulty, along with every matches difficulty, is as difficult or easy as the programmer decides it is.
And he does this with every other subjective call.
In other words it’s exactly what Tony suggests. Utter bollocks.
Super computer has perhaps listened to Danny Murphy, Roy Keane, Peter Schmeichel, Gary Neville, Tim Sherwood, Jason Cundy, Jamie O’Hara and other assorted Talk-sh..e “experts”
Well, considering the fact AI is now involved, I’d forget all about it.
AI reads all the crap on the internet to ‘learn’.
It learns to be sexist, racist, etc… because a majority of people developping AI are rather of the ‘white-under-30-incapable-of-getting-a-date’ nerds. One just as to look at the many failures that have appeared.
So, AI went learning all about the PL. And what kind of crap does AI read ? Well all the crap of those so-called pundits who kept on going after Arsenal for decades, and at least since the internet existed.
What did it learn ? Arsenal are bottling it, Arsenal are crap, Arsenal can’t win etc etc etc… it ain’t the better reports of the past 3 years that can compensate the decades of anti-Arsenal BS.
So, no surprise….
Perhaps they meant SuperB Computer and they keep leaving out the “B” at the end? Just saying.
Chris
Talking of this “Arsenal are bottling it” nonsense that gets endlessly recycled year after year we had it yet again yesterday from football365.
‘Arsenal feature twice in eight Premier League leads lost with seven to go…’
They just cant help themselves, but the thing is even within their own piss taking article they prove just how stupid it is because as they demonstrate Man Utd have a far bigger history of ‘bottling’ it, and that’s just under Sir Alex.
“It’s happened eight times since football was invented in 1992, with Sir Alex Ferguson throwing it away more often than he would like you to remember.”
https://www.football365.com/news/arsenal-man-utd-premier-league-leads-lost-seven-games
Which directly contradicts their own previous Arsenal piss take article from May Last year that lead with:
‘Top ten biggest Premier League title bottlejobs: Arsenal’s 2022/23 vintage cracks all-time rankings’
“It is time to place Arsenal 2022/23 among the greatest Premier League title bottle jobs ever. The Gunners make up almost half of the all-time top 10.”
Which again contradicts their previous article from April that year titled:
‘Anatomy of a Bottle Job: Eight factors to separate the true choke from a mere collapse’
The point is if you go and read all 3 articles it is obvious every team blows a lead at some point, but when it’s Arsenal it’s headline news and called a bottle job, and be sure nobody is ever allowed to forget it, as demonstrated by the fact that Arsenal are the team mentioned in the first 2 headlines and in the 3rd article we are the first photo.
I cant put all the links up as it wont post but if you google football365 bottle jobs they pop up like spring daffodils.
But, back to ‘super computers’, or rather not so ‘super’ computers. I watched a program yesterday where we visited the Met Offices shiny modern building and all it’s very sophisticated and expensive weather predicting computers. Brilliant. Except as everyone who’s ever planned a Barbeque knows, they don’t bloody work.
Much to my own annoyance, I actually have the Met Offices ap on my phone. Why I don’t know.
At the weekend I looked at it to see what was in store for the coming week. I swear it must of been the blandest forecast in the history of forecasting. Every day the same. Cloud cloud and more cloud. No sun. No rain. Just endless cloud.
Well, on Tuesday I woke with baited breath awaiting to be blanded to death by endless cloud, and blow me down, no pun intended, it was pissing down. I thought, that cant be right, the Super Computer never mentioned that, so I checked my Met Office ap to see if I was imagining things, and indeed I was. Apparently it wasn’t raining. I was imagining it. Which was a bit worrying as I must of also imagined getting soaked when I put the bins out.
Super Computers aye. They cant even help you plan when to put the bins out.
Re our penalty claim at the end:
This is all over the internet. Apparently:
Gabriel handball: Ref ADMITTED ‘kid’s mistake’ as raging Thomas Tuchel sensationally claims official didn’t give penalty against Arsenal because Champions League game was too big” “Thomas Tuchel has claimed the Arsenal vs Bayern referee admitted to knowing he’d made a “mistake” by not awarding a penalty for a Gabriel handball.”
“In the 67th minute of a pulsating Champions League quarter-final first leg clash at the Emirates Stadium, David Raya played a short goal kick to Gabriel Magalhaes, who picked up the ball and retook the kick himself, seemingly unaware referee Glenn Nyberg had already blown his whistle for the restart. Bayern players protested but no penalty was awarded.”
So if he’d already admitted, even if at the time only to himself, that he had made a ‘kids’ mistake in not giving Bayern a penalty, he’s hardly likely to give us one is he? Okay I get that, as wrong as it is. But what worries me is, is that the same reason VAR didn’t get him to look? Was they privy to the referees admittance to what he believes was his mistake? If so:
Did they not tell him to review because they knew he was levelling things up?
Did they advise a review and he just ignored them?
It was such a poor decision, and so obviously needed re evaluating that something dodgy had to be going on.
@Nitram,
read a piece, I think on the Athletic that states that in this case the referee intellegently applied Law 18.
Law 18 does not actually exist, it is however at the beginning of the laws, a paragraph that basically says good faith and spirit of the game should lead all decisions by players are referees.
Gabriel picked up the ball to pu it at the corner of the 5m box.
He was not under pressure, nor was Raya. He just had not heard or noticed the referee whistle.
Arsenal did not gain any positional advantage, did not escape a Bayern attack or threat.
Gabriel did it in good faith, and no advantage was gained.
So Law 18 applies.
End of the story.
No mistake. Just applying the rules of the spirit of the game.
Tuchel did not elaborate of H. f…g Kane deliberately trying to maim Gabriel, did he ?
Anyway, expect Bayern to keep screaming murder to make sure the referee for next week is influenced before the game even starts…. this is what you get when people or teams feel they have some divine rights.
On he other side, the english press should be well aware that Arsenal are key to the PL getting a 5th ECL spot, so they should be screaming in defense of Arsenal for the good of Manure or Sp*rs…but they visibly are stupid enough not to….
“They are presumably talking about the eight game run from 9 April to 20 May in which Arsenal won just two games, drew three and lost two”
Er, shum mishtake shurely? I make that a total of 7 games
Ando
Okay, so it was 3 draws. Some of us are bright enough to realise that.
Not really sure what you achieve with sarcastic posts like that? But hey, whatever floats your boat.
“AI powered supercomputer” 🤣
In my not inconsiderable experience, supercomputers are powered by electricity.
In football most supercomputers are powered by journalistic imagination
Nitram,
That wasn’t meant sarcastically; the ‘shum mishtake shurely?’ was meant as an amusing reference to Private Eye’s occasional editorial comment. Sorry you took it the bad way.
In the past, Tony has himself admitted to some dyslexia for which he may be forgiven (that’s just a figure of speech, by the way and not meant to be condescending). In an article about statistics and performance, I would have thought it made a lot of diference whether 2+3+2= 8 or 7. Otherwise, the main premise of Untold – that evidence based on figures & percentages is a more accurate indicator of form than journalistic speculation – is itself brought into question.
And your rebuttal regarding being “..it was 3 draws…some of us bright enough to realise that…” makes even less sense, as Tony had already stated that in the original article.
I am as committed to supporting Arsenal today as the first time I ever attended a gane at Highbury in the ’50s, when I remember Danny Clapton, Vic Groves, ‘Flint’ ‘Mc Cullough and others turning out for the team.
Come on you Gunners!