People call up Chris Sutton every week on 606 to have a pop about the Premier League predictions he makes for the BBC website and tell him he doesn't know what he's talking about - but does he?
In a season that has been full of surprises, from Liverpool's loss of form to Chelsea's continuing inconsistency, Sutton's table (below) still looks pretty close to the real thing.
He has the top and bottom two right - albeit in the wrong order each time - and, overall, 13 of the 20 teams are within two places of their actual position, with three (Liverpool, Leeds and Leicester) absolutely spot on.
"This season has probably been the most difficult for predicting results since the Premier League started, so I'm pretty happy with that," Sutton told BBC Sport.
"In recent years you didn't have to think twice about whether Liverpool would win every week and you could always trust Chelsea to beat most teams too, but the wheels have really fallen off for both of them this time.
"They have been on the wrong end of some remarkable results, while another of the usual top-four contenders, Tottenham, are so up and down they are just impossible to predict.
"On top of that, the fact there is a relegation battle involving nine clubs who can all beat each other tells you how hard it has been to get their scores right.
"The team at the bottom of my table, Bournemouth, still have a fighting chance of staying up in real life, which has surprised me.
"I wrote them off at the start of the season, and did the same again around Christmas, but they keep on picking up points.
"They have proved me wrong a few times now, but they are not the only ones because I don't have many teams who I can rely on, to always be good or bad.
"All things considered, I think I'm doing really well but please remember that, if I really were a predictions genius, I would not be working as a pundit - I'd be a billionaire."
Sutton's prediction-making predecessor for the BBC, Mark Lawrenson, once tipped his former club Liverpool to stay undefeated for 159 successive games, between May 2016 and November 2020.
There have been no imaginary 'Invincibles' this season, though.
In the alternative universe based on Sutton's results, Manchester City went unbeaten for their first 23 games, before he (wrongly) tipped Arsenal to beat them at the Emirates in February.
Now it is Brighton who are currently on the longest unbeaten run - Sutton has not backed them to lose since they went to Etihad Stadium in October.
Meanwhile, West Ham and Chelsea are two of the teams faring much better in Sutton's table than they are in reality - in contrast to Manchester United, who are doing noticeably worse.
"I always think Brighton will score goals, that's why I go with them, while United are definitely one of the teams I have doubted," Sutton explained.
"There were spells before the season stopped for the World Cup where I thought they would lose every week, but they improved under Erik ten Hag as the season went on and my predictions have reflected that.
"With Chelsea, I'm a former fans' favourite from my prolific spell there, so I've got to keep their supporters happy, haven't I?
"Being serious, though, I kept expecting them to get better under Graham Potter, but that never happened. It's the same with the Hammers, who have never got going this season.
"I still thought I'd backed Chelsea to lose more often than I have done, but then I don't go back through my previous predictions for each team to check what I said about them in the past, because that doesn't help me get their next results right.
"I'm not worried about getting this table spot on because my aim is to get as many scores correct as I can each week and beat my guests. That's what I've been doing, most of the time."
|+/- DENOTE POSITION DIFFERENCE BETWEEN SUTTON'S TABLE AND ACTUAL POSITION|
Are you doing any better than Sutton or his guests? Yes, but not by much.
Sutton has got 132 out of 270 results correct this season - 49% - while his guests have managed 122 between them, or 45%.
Thousands of you have been making predictions too, using the voting buttons on his predictions page each week, and you have got 140, or 52%, right.
It's a good effort, but maybe it proves that making predictions is not as easy as it looks?