Making money from weather forecasting?

In_For_A_Penny_2

MoneyBob wrote: »

Not heard of that one before. I'm a sailor also btw Sam.

My system can predict, with greater than 50:50 success, spells of clear weather, months in advance.

I'm still not sure if the OP is a wind up but if you genuinely can do what you claim to there is a fortune to be made – event planning - outdoor concerts, weddings, sports events etc as well as farming, construction, aerospace and tourist industries etc etc.

The only problem is you can’t.

Even “better than 50:50” is likely down to chance and the knowledge that in summer the weather is better than average and in winter it is worse than average.

You need to subject your methods to some statistical analysis.

Damage

MoneyBob wrote: »

Genuine question - how would that help me?

It would provide a verifiable account of the accuracy of your method/system.

As you can tell from the nature of some of the replies so far, people aren't taking you seriously at all. If you put in place a way of proving that your system works, then people will take notice. How silly would they be to mock you if you can prove to any sceptics that your system works?

Biggles

MoneyBob wrote: »

Genuine question - how would that help me?

Genuine answer - then people will believe that you really can do it and be more inclined to help you. That and downright nosiness, obviously.

MoneyBob

In_For_A_Penny wrote: »

I'm still not sure if the OP is a wind up but if you genuinely can do what you claim to there is a fortune to be made – event planning - outdoor concerts, weddings, sports events etc as well as farming, construction, aerospace and tourist industries etc etc.

The only problem is you can’t.

Even “better than 50:50” is likely down to chance and the knowledge that in summer the weather is better than average and in winter it is worse than average.

You need to subject your methods to some statistical analysis.

I have subjected my predictions to statistical analysis, certainly enough to punt some of my own hard earned cash on.

If you hypothetically assume I can do what I say I can, how exactly would I go about making money from it from the industry sectors you've suggested?

I'd very much like to avoid giving weather advice out to anyone. There is the potential for litigation, and there's the potential for my algorithm to reveal itself.

My best case scenario is to find a share or sector I can invest in during cloudy rainy times, which will see modest to significant price rises during subsequent periods of clear skies... perhaps no such sector or business exists?

I'm then risking only my own money, without a heck of a lot of hassle.

You mention that it's easy to do what I'm saying in summer... it's actually not. And remember, I'm saying I can make these predictions at all times of the year, not just summer.

Damage

MoneyBob wrote: »

I'd very much like to avoid giving weather advice out to anyone. There is the potential for litigation, and there's the potential for my algorithm to reveal itself.

How could you possibly come to any harm or risk giving away your algorithm simply by making some weather predictions to help prove to people that your system works?

Do you not need to prove that it works? If you don't need to prove anything then I can understand why you might want to keep it to yourself, but how can you expect anyone to pay you for your predictions without first providing some sort of convincing evidence that you're not just making it all up and keeping your fingers crossed as to the outcome?

MoneyBob

Damage wrote: »

How could you possibly come to any harm or risk giving away your algorithm simply by making some weather predictions to help prove to people that your system works?

Do you not need to prove that it works? If you don't need to prove anything then I can understand why you might want to keep it to yourself, but how can you expect anyone to pay you for your predictions without first providing some sort of convincing evidence that you're not just making it all up and keeping your fingers crossed as to the outcome?

I don't need to prove that it works if I can find a way to profit from it by buying and selling something which has a price which changes in response to the weather.

This would, for reasons I've stated above, be my preferred way to profit. It's the simplest, safest and least hassle.

Damage

MoneyBob wrote: »

I don't need to prove that it works if I can find a way to profit from it by buying and selling something which has a price which changes in response to the weather.

This would, for reasons I've stated above, be my preferred way to profit. It's the simplest, safest and least hassle.

I see what you mean now. I thought you meant that you wanted to do something along the lines of selling your forecasts to others who might benefit from them.

Still, it would be fun to make some predictions in here, then be able to bask in all the glory when you're right. Surely no one would be able to deduce your algorithm simply from a set of straight predictions.

Linton

I doubt anything that you can invest in changes price according to short term weather fluctuations. Why not set up a website funded with advertising? If people find your predictions accurate millions could visit it to plan their holidays.

How do you handle the great local variability in UK weather? Does your method just work for the UK or globally?

Damage

Have you looked into behavioural finance? Perhaps there is a correlation between the weather and stock market activity that you could profit from.

EDIT: I don't mean single company shares in something tangible that is weather-dependent, rather a general behaviour pattern over the market as a whole that is reflected in a general mood, so perhaps a range of equities spread over many categories.

StumpyPumpy

MoneyBob wrote: »

I have subjected my predictions to statistical analysis, certainly enough to punt some of my own hard earned cash on.

I have to say, that your responses to this thread don't give much indication of a statistical analysis:

(PS. I predicted the current clear spell we're having, over a month ago, which marks about 10 correct predictions in the last 18 months, which is what's prompted me to try a little harder to find some way to cash in on it)

About 10 correct predictions in 18 months? Out of how many? Let's say that shows a 75% accuracy; that would mean you got 3 wrong, but a sample of only 13 predictions in 18 months is too small to be statistically significant. Less so when you take into account the fact that predicting a clear spell is not a 50:50 bet in the first place.

For instance, I have no methods for weather prediction other than memory, but I predict, right now that Saturday will be sunny. My prediction is based on history and the fact that next Saturday is our annual village fair. I've lived in the village for 25 years and the fair is always held on the same weekend in June, and in those 25 years the fair has never been cancelled due to rain (It is in a field - if it rained heavily it would be cancelled) I have attended the vast majority of them and have no recollection of rain at all, plus photos to back my memory up.

I wouldn't say I'm 100% sure as I haven't checked with the Met Office, but 70% chance of sun would be a safe bet. That doesn't make me a weather forecaster though.

You also have to take into account the way that the newspapers' daily horoscopes work - generalisations and a reliance on people only remembering when they get things right in the loosest possible sense. e.g. "My horoscope said I might come into money... and I found a fiver down the back of the sofa! Horoscopes work!!!"

I'm not saying you are doing it deliberately, but how are you selecting your predictions for analysis? Are you discounting predictions that "weren't done properly" for instance. And how did you measure success in this statistical analysis? If you predict a 70% chance of a clear day and it turns out cloudy, that is still, technically a success as you allowed for a 30% chance of "not sunny". How did you measure "clear"? Is a sky with 30% cloud cover, clear? Or is it variable depending on your prediction? If you forecast clear and a single, small cumulus floats by, is that a fail?

It is amazing how often people can subconsciously subvert their own statistics even if they are gathered with the best of intentions, but if you want to try and make some money out of it, I'd suggest you try to displace Piers Corbyn's position of bizarre weather forecaster with a cult-like fan base who don't let the facts get in the way of their belief in his powers of prediction. (Proper analysis of his predictions show they are quite laughable, but people still pay hefty amounts to subscribe to his site)

If you think you can do it, then go for it, but there are very few industries that are both reliant on the weather and don't already have this factored into their stock market valuations which is what you would need.


SP