Modelling

Applying Math To Guesses - Dilbert by Scott Adams

Source

This entry was posted in Uncategorized. Bookmark the permalink.

20 Responses to Modelling

  1. Professor Fred Lenin

    Crap fed in Gold comes out , yes right comrade ,lot of it about ,cant beat the science can you ?

  2. Jock

    This is why I dont trust models. As an old Finance guy I have seen my share of dud models. They are technically well put together. But they incorporate myriad assumptions. The inherent bias in those assumptions skews the model. Added to that is the fact that these days they overlay Monte Carlo and other methods to assess risk and probabilty etc. but they are still an assessment of the future at a moment in time. If a Board wants an asset bad enough the figures and returns will magically reflect that. If you believe CO2 will increase temperature by 2 degrees then the models will reflect that. This is nor rocket science. The more assumptions, the greater degree of model degradation.

  3. Tintarella di Luna

    But they incorporate myriad assumptions.

    ass u me

  4. HT

    When I was running a “Capability” section in defence, when starting off with a new capability we always started off asking the contractor for a “SWAG”. That is, a “Scientific Wild Arse Guess”. At least we were honest about it…and as it turned out, our SWAG was usually better than subsequent detailed modelling conducted by the Dept of Finance when were numbers got high enough to require Ministerial sign-off.

    The real crime in all this modelling was, post-project implementation, called IOC (Initial Operating Capability), nobody, no-one, ever checked that the capability we spent the money on was actually delivered, but I digress.

  5. Bruce of Newcastle

    Camelot!
    Camelot!
    Camelot!
    It’s only a model.

  6. H B Bear

    Shouldn’t you get a consultant to do the modelling so it will be right?

  7. HT

    H B Bear
    #3587293, posted on September 16, 2020 at 6:30 pm
    Shouldn’t you get a consultant to do the modelling so it will be right?

    How cute, somebody thinks consultants are to get things right LOL.

    No Humphrey. Consultants are there to blame when it all turns to custard. Nothing more, nothing less. Even if the Consultant’s recommendations were completely ignored, they are still to blame for any cockups. THATS what they are paid for, and they know it. It’s all part of the game.

  8. Mother Lode

    A lot of people will see that comic without thinking to apply its insight to the AGW or Covid scares.

  9. H B Bear

    No Humphrey. Consultants are there to blame when it all turns to custard.

    Wouldn’t know about that. Was always long gone before the final deliverable hit the printer. If it all goes wrong it was always the implementation.

  10. JohnJJJ

    Consultants are there to cut the Gordian knot of bureaucratic decision confusion. No one can guess what the boss wants. So hire a consultant and watch how the boss reacts to any of the consultant’s gobbledygook. If a bad reaction, blame the stupid consultant, if good take the praise for hiring them. I’ve made a lot of money being that consultant.

  11. Squirrel

    Yes, it’s one of the great mysteries of the universe how modelling always produces the outcomes desired by the people who commission it – and dimwit journos lap it up and trot it out like it’s come down from the mountain on tablets of stone.

  12. Lee

    Yes, it’s one of the great mysteries of the universe how modelling always produces the outcomes desired by the people who commission it –

    That’s because assumptions are included that support their bias, and anything which doesn’t fit their preconceived ideas is rejected.

  13. HT

    H B Bear
    #3587387, posted on September 16, 2020 at 7:49 pm
    No Humphrey. Consultants are there to blame when it all turns to custard.

    Wouldn’t know about that. Was always long gone before the final deliverable hit the printer. If it all goes wrong it was always the implementation.

    See, win win. Your just as much a player as the PS 😉

  14. HT

    BTW, I love Dilbert. I had a complete set of the books and referred to them regularly. Amongst the best management books published, ever.

  15. Arnost

    The more assumptions, the greater degree of model degradation.

    Indeed!

    “with 4 parameters I can fit an elephant, and with 5 I can make him wiggle his trunk”

    https://twitter.com/fermatslibrary/status/1235582404388982785?lang=en

  16. Hay Stockard

    Scott Adams. Ridicule is the best weapon against the left. He is a hero. If he was female he could have my children. There is no higher praise than that.

  17. Professor Fred Lenin

    Modelling is like a Royal Comission the retired judge is given a copy of the result before the first sitting
    Cant have the government having the truth thrust on them , its not fair and might destroy their lucrative careers ,

  18. Tim Neilson

    Modelling is like a Royal Comission the retired judge is given a copy of the result before the first sitting

    I don’t think it’s quite like that Prof.

    The terms of reference are the main way that the government ensures that no undesirable conclusions are reached.
    Like the inquiry into the Black Saturday fires. Everyone knew that the cause was catastrophically bad mismanagement of fuel loads and other forestry matters, caused by the ALP trying to protect inner Melbourne seats by pandering to Greens voters.
    So when the Labor government announced the inquiry they said it would be a no holds barred nothing off limits totally transparent investigation into everything that happened from the moment the fires started.

    See? Simples! Bernie Teague did state expressly in the report that witness after witness blamed the years of green-driven mismanagement – but none of the official findings could even refer to the real culprits and real cause. So the government could solemnly say that it accepted the findings and would act on them.

  19. vlad

    Didn’t John Stone once say that whenever the computer spat out obviously wrong answers he would substitute his own estimate for whatever it was – inflation next year or unemployment in six months’ time.

    If that story is as I remember it, the model wasn’t doing any real work at all.

Comments are closed.