All Forum Posts by: Severin Sadjina

Hei Tore, and welcome!

Good to see there are a few Norwegians (or quasi-Norwegians, in my case) on here as well. I live in Ålesund, but am in Trondheim every few weeks for work. I'd love to meet up and talk to you at some point! I just got into Ethereum mining as well (though my brother is doing all the work there really), and I have been employing machine learning / AI models to study the market in Ålesund.

Here's what I do: I have made my own Zestimate (machine learning / AI models) for my local markets here in Norway. Because data on actual sales prices is harder to get a hold of and process for me, I use the listing price. My models then predict a listing price for a given property, given a few key metrics (such as build year and square footage etc.).* I use that to identify deals and get an idea for what the property is actually worth.

But now the important part: If something does look interesting, I absolutely go in and do check what similar properties in the neighborhood sold for, and check how much lower/higher those prices were compared to the listing prices of the properties. In my case, the market is at a point where there is only a small difference between listing prices and actual sales price on average. So I am good to go. (Otherwise, I'd have to adjust my listing price estimate accordingly. Or back off.)

So, to sum up: I think you CAN use the Zestimate. BUT make sure it's not the only thing you're looking at. A Zestimate can be off by more than 20% in some cases, so you always need to double check against other methods. ...and then again, people's estimates can also be off by more than 20%...

*) Technically, this is essentially a comparative market analysis (CMA), but much more objective and comprehensive. A typical CMA will compare with a few similar properties in the neighborhood. Knowing the actual way property prices vary due to location, square footage, build year, etc. (this is what a model does essentially), you can compare with many more properties automatically.

Hi Sean, and thanks!

1.) Did you manage to gather insights into the Denmark and London real estate markets? OR was that pre-REI?

2.) My models are off by 20% or more as well on some of the properties (maybe about in 10% of the cases). I would by no means say that mine are better than Zillow's! However, any given prediction is off by about 6% "on average", which is useful enough to identify deals more or less automatically. I actually doubt that human experts would manage to get anything much better than 6% error rate on average. And the cool thing here is that I have (or at least had) no idea about property values and the market, but could use these technologies instead to get there very fast. Same applies to rent prices: same technology can predict what a given property should rent for.

3.) Thanks for the tip on Airbnb. I have thought about it a bit. I think I'd like to start with long term rentals, but I'm intrigued to try the Airbnb thing at some point. Do have any experience in that regard?

 Hi Kim!

Can you elaborate on your statements on the Austrian and Norwegian markets? Just curious to hear your opinions. I live in Norway, and I am Austrian, you see ;)

From the little I have heard and seen, the US really does seem to offer opportunities that I just couldn't find. I keep hearing of deals on houses bought in the US for 50.000$ and rented out for 1.000$ per month. I can only dream of such numbers...

Hi Michael!

I wish I had started thinking about REI at your age! So good on you! Why do you want to focus on flips to start out? (No issues, just curious!)

All the best!

Hi,

I'm Severin. I'm from Austria (the one with Lederhosen, not the one with the kangaroos, sorry), but have lived in Norway for 11 years. I'm a research scientist with a PhD in theoretical physics, i.e., I'm a ginormous nerd. I love numbers, and that's heavily coloring my approach to real estate. I'm using artificial intelligence to help me find deals (more on that later).

For my first deal, I am currently trying to find condos as rentals, to hold for about 3 to 7 years. I'd prefer being able to rent out houses, but no one does that in Norway (everyone buys instead). And leasing is still unheard of. I'd like to build up cash flow and passive income, and hopefully grow to owning four units (the maximum number you can have here without bothering with registering a business).

It all started innocently enough, when I told my Dad that I would now finally get my own car. (At the age of 34, thing's are a little different here and/or I am different). I had just found my new position as a research scientist at Scandinavia's largest independent research organization, SINTEF, and I would soon be in the position to buy a car in cash. And all my Dad simply said was. "Well, that's a terrible investment." And after my initial surprise and confusion had settled, I realized that he was a 100% right.

Well, a few weeks later, I was starting to collect data on real estate in my local market in Western Norway. I started using artificial intelligence and machine learning techniques to build models that could predict a property's market value, given a few basic metrics. I was basically making my own Zestimate (Zillow's AI-backed value estimation model). And it worked pretty well. I am now down to about 6% median error rate when predicting the market value of properties from only four numbers (for example, square footage). And that's actually almost as good as Zillow's US-wide numbers.

Anyway, fast forward: I have looked at and studied 2400 properties (well, my AI has, really), I have seen about 10 properties (all of which my models said where potential deals), and made offers on 4. So far, I haven't managed to get any of my offers accepted. But I will soon. And I can't wait to get started!

I've also listened to several BiggerPockets podcasts, and I loved every one of them. They also got me to reading Rich Dad Poor Dad and The Cashflow Quadrant, which really helped me get into an investor mindset.

Alright, that's about it for now. I'd love to hear if there's other folks from Europe here, and who has done some more advanced mathematical modeling to help them with their real estate business.