What is HDR Analytics, and why should I subscribe?
HDR Analytics is a forecasting & data journalism site focused on politics/elections, sports, and weather. It began as HenryDRiley.com (hence the HDR) and focused on elections and sports before launching a snowday forecast in 2022, which quickly became popular. The goal of this new site is to expand upon what has already been built in a more professional, organized manner that has more actual content and reporting using statistics and my forecasts. I enjoy explaining and discussing these forecasts, as well as politics, sports, and weather in general (I do participate in competitive public speaking, after all) and hope to reach a wider audience through this website. In addition to the forecasts themselves, of which I hope to launch more this year, there will be articles with my own analysis of the odds and how things all tie together.
Subscribing is completely free, and simply means you’ll be notified when I post new articles here! Substack is a much better platform for articles, and means I can write more often and more quickly. At least for now, all forecasts themselves will remain at HDRAnalytics.com, which will always be linked in the Substack menu.
About Henry
My name is Henry Riley, and I'm a rising junior in high school who has been predicting elections since 2020, I've been forecasting snowdays for the Madison, WI area since 2022. I am an avid political junkie, and I participate in public forum (PF) debates at state/national tournaments, and also compete in Extemporaneous Speaking for Forensics. I have a fair amount of tech experience, and have had a website since age 7, so I've been doing this for a while. I currently have no official education in statistics, but will be learning Calculus next year and likely AP Statistics the year after.
Are the forecasts representative of your personal opinions?
NO. Unless otherwise stated, all forecasts are strictly data-based, and I never change/alter results or inputs in any way, whether to produce a result I desire more (like the Vikings winning the Super Bowl) or one that I personally believe will happen, despite the model having a low chance. The exception to this is within articles, I may sometimes offer predictions that are purely opinion-based, like on local elections, and that will be disclosed. Additionally, the way in which any forecast works will not change without full disclosure (and usually not at all) once it launches. Things may change between seasons or election cycles, but not within that same season or cycle.
Are your forecasts typically right?
If this is the question you're asking, then you need to understand the fundamental premise of statistics and forecasting. My models will almost always give probabilistic forecasts (with odds) as percents or ratios (like 5 of 6, etc.) and while an outcome with a 99% likelihood not happening is a pretty big error, it's still not "getting it wrong". A 1% chance is a very real chance of something happening, and (almost) nothing is impossible, whether it be an election upset (looking at you, Hillary) or an NFL comeback (Colts-Vikings 33-0 is my favorite example of this).
But of course, you can measure the accuracy of forecasts, and I fully plan to post detailed analysis of my past/future models later on, using a few different methods.
Why do some forecasts not update everyday?
The simple answer here is that I often have a lot going on, and while I'd love for this site to be a full-time paid job, I'm still in high school, and it's merely a project of mine with no income.
