Skip to content

Supercomputer says we will be champions this season.

edited July 2022 in Football

The supercomputer has spoken and has done the number crunching ahead of the new league one campaign.

Reach's in-house boffins have run the rule over the 24 clubs in the third tier in 2022-23 and have monitored the necessary algorithms to give a predicted final table. And in a field of heavyweights, Wycombe Wanderers have been tipped to come out on top this term.



  • edited July 2022

    I've spoken to the guy who came up with the model so can explain what went into it if anyone's interested (or we can just enjoy it out of context, which is what I'm doing tbh)

  • I don't know if this comes up for everyone, but next to the article there is an advert with a seagull pecking the ground. I imagine this to be what the Mirror's supercomputer to be like.

    47% chance of winning the League is about 11/10. Any team at this price would be ridiculous (bookies' favourites Wednesday are currently 9/2).

    Mind you seagulls are always stealing chips, so maybe there is something to it...

  • What did the supercomputer predict last season?

  • Nuclear war. Which is why this year it's been demoted to league one football.

  • This has been one of the only comments on here that has properly made me laugh!

  • Does seem mad they can come up with such a high margin of probability when we neither know who our No.1 will be, or what our frequent CB pairing will be.

  • 67% chance of automatic promotion is both massively over-estimating our chances of success and too specific for what is effectively a finger-in-the-air guess.

  • It's not a guess, but it is a flawed model as it doesn't take into account new signings (it's based on teams' Elo ratings, which fluctuate based on result, quality of opposition etc). But it's still interesting and will do decent numbers.

  • And having regard to @Wanderers82's assessment of the capabilities of our squad and the poor signings to date!

  • It might be based on a model but it’s still effectively a guess.

  • 47% (@LX1) or 67% (@Chris) ? Not that it matters a jot.

  • It’s probably akin to the prediction models they use in cricket these days. Some sort of logic behind it but little in the way of common sense

  • Ah, but can it measure our ambition?

  • I'll take it, and I suppose Wednesday and Franchise have a bit of momentum from last year and Derby and Posh shoudl be financially competitive but I find it hard to believe Ipswich and Pompey will do as badly as predicted there...

  • For me... Wednesday champions, MK/Ipswich 2nd (can't separate them at the moment) then us, Pish and Bolton in the play-offs

  • Poxford, Pompey, Plymouth and probably Barnsley all in the mix

  • I imagine that, as we were play-off final losers (sorry to say that), the algorithm gives us some extra weighting for the new season positions?

    Combine that with our previous 2 year positions and it makes sense.

    But guessing also works!

  • The article suggests the model predicts a 47% chance of us winning the league and a 67% chance of being in the top two.

  • Wednesday and Derby (sorry but their wage budget will be way beyond ours) automatic). MK, Posh, Forest Green and us in play offs.

  • It's kind of wild and presumably based on the last two seasons or something. I mean a team that requires a 7-0 win to reach promotion might be calculated by a dumb computer model to have a 100% chance of going back down but that hardly represents Bristol Rovers real prospects.

    For us I think the play offs are the starting target - and a challenging one at that.

  • If I have learned anything from the past ten years...big clubs with big wage bills and better players don't always necessarily get the results they require.

  • Forest Green?! I think they'll stay up, but that would be astonishing after the summer they've had.

  • edited July 2022

    This is how it works:

    "It all works off the Elo ratings of each club. If you win a match your Elo rating increases, if you lose it decreases. If it's a big upset it goes up/down by more points than if it's an expected result. 

    For each fixture, the simulator compares the Elo of the two clubs, adjusts for home advantage, and looks back at past results by clubs with a similar gap in ratings. That gives us a probability of the outcome of the match. The simulator then runs through all games assigning results and adjusting Elos according to the outcomes it predicts. 

    That's it in a nutshell. It's got several drawbacks. Not least that it doesn't take into account transfers, injuries or changes in manager."

    Like I said, flawed but interesting nonetheless.

    Fwiw, FourFourTwo (non-algorithmically) predicted us to finish 9th, which is a lot more likely!

    I'm holding out for 6th again 🤞

  • 6th is very satisfactory.

  • 6th and another tilt at the play-offs would do me. Anything above and beyond that would be gravy...

    Very excited for the new season.

  • edited July 2022

    Elo works well for chess, but less so for football where there is a lot more randomness even over a full season.

  • edited July 2022

    If it is based on Elo, how come Derby aren’t hot favourites given they achieved the results of a mid-table Championship side last season?

  • Cause relegation will send your rating plummeting

  • And because they’ll probably have another 12 points deducted.

Sign In or Register to comment.