• 11 Posts
  • 562 Comments
Joined 1 year ago
cake
Cake day: February 14th, 2024

help-circle
  • This is a great answer.

    The other aspect is even if you do these, the US venture capitalists go “we’ll give you money, but you have to headquarter in Silicon Valley so you’re close to us”

    Then they funnel their other startups and companies on to your product, so you get funding and customers.

    We do need to clean up the small business process. Hell, as a hobbies I find getting data out of stat can such a pain in the ass, while the US has a ton of open data.

    We also need to build a viable investor culture here so that these startups aren’t squashed in the germination phase.












  • Looking at any individual poll isn’t a good idea.

    Polls can and do use population adjustments to account for demographics bias, but it’s very difficult to accommodate unknown response biases.

    538/338 style models that use polling error to rank pollsters and reweight them based on objective biases to election outcomes. 538 famously was one of the best forecasts of 2016 because at the time this methodology was uncommon, and they had given a forecast of 1/3 which is an entirely viable event.

    Trump 2016 was an outcome that most traditional election signals missed, which 538 and 338 models also include.

    Trump 2024 had an over 50% chance of winning by most forecasts.

    Anyways, all that to say, I have a background in stats and forecasting, I have a lot of faith in these models, but they’ll never account for everything like how Trump’s tariffs have shifted Canadian opinions and preferences.