Predictive Models Response

[THolt]

Shannon,

What variables did you take into account when you made your prediction that a major hurricane was going to strike Florida in 2018?

[Benjamin]

This isn't exactly a testimonial I'll move it to chatter box.

[Shannon]

Shannon,

What variables did you take into account when you made your prediction that a major hurricane was going to strike Florida in 2018?

That is not something I can discuss without giving away too much information. But keep in mind, this is a snapshot of the most probable future at the time of modeling, and is not set in stone. The reason I posted that was as a note to myself so I could see if it actually came to pass. The models seemed pretty adamant about it.

[Adrien Silva]

Should be interesting to see if this fans out

[Daredevil]

You were one year late Shannon, Hurricane Imra is a category 5 hurricane about to hit Florida. /S

[Zane]

Holy Crap,How did he know?

[Daredevil]

1 Year early Zane. Or, this is just for this year and next year is gonna be stronger.

[Shannon]

Hurricane Michael. I was hoping this would be wrong.

[Have at ye]

Any info on whether it's going to hit your area directly, Shannon?

[Oversoul]

The fact that this got predicted by the models is amazing @Shannon . Are there any other predictions that would be of interest?

[Have at ye]

The fact that this got predicted by the models is amazing @Shannon . Are there any other predictions that would be of interest?

Yeah, somebody should throw a lot of money at Shannon to show them how to optimize their weather forecasting models.

[Shannon]

I was looking for indicators in the models that were very major and repeated flawlessly. Trying to see if the models could be used to predict weather events. I haven't used them for that since. But I'm not teaching anyone how to make my models, no matter what they offer me.

Now that we know this type of prediction can be done and done accurately I'll create more experimental forecasts when I have some time.

[Shannon]

Since I can't find the original prediction thread, I'll post my latest prediction with the models here.

This experiment is to see how accurate the models are when only one run is done. Usually I do a large number of runs and follow the average value.

This prediction is made on June 21st, 2020. The question was:

What will the average value of the American stock market for each month be in the next 6 months, relative to what it is today?

Results:

June: slight increase from here to the end of the month.
July: appears to be some instabilty which challenges the market, but it seems to maintain a relatively steady position for this month.
August: A significant, but unstable, increase.
September: A very large increase.
October: A serious crash, but far from the worst possible.
November: Market is "starting over", probably has small gains, but appears to remain at about where the crash left it.

Remember the following when considering what actually happens:
1. This is an experiment to see what happens when I don't repeatedly verify the data the models give. Can they accurately predict something as unpredictable as the stock market, when it is even less predictable during these crazy times, while only being run once - which should result in the least possible accuracy? I don't know. That's why we are doing this experiment.
2. These are for the average value of the American stock market as a whole during that month (or in the case of June, what's left of it).
3. This is NOT me showing off the models by making bold predictions to prove it works. I know it works. This is me trying to see how well it works when used in the worst possible ways against the most challenging possible targets.

I'm not giving investment advice here. This is an experiment. You should not consider this data valid until after this period of time has passed, and we can consider what actually happened to compare it to.

[MD81]

Since I can't find the original prediction thread, I'll post my latest prediction with the models here.

This experiment is to see how accurate the models are when only one run is done.  Usually I do a large number of runs and follow the average value.

This prediction is made on June 21st, 2020.  The question was:

What will the average value of the American stock market for each month be in the next 6 months, relative to what it is today?

Results:

June: slight increase from here to the end of the month.
July: appears to be some instabilty which challenges the market, but it seems to maintain a relatively steady position for this month.
August: A significant, but unstable, increase.
September: A very large increase.
October: A serious crash, but far from the worst possible.
November: Market is "starting over", probably has small gains, but appears to remain at about where the crash left it.

Remember the following when considering what actually happens:
1. This is an experiment to see what happens when I don't repeatedly verify the data the models give.  Can they accurately predict something as unpredictable as the stock market, when it is even less predictable during these crazy times, while only being run once - which should result in the least possible accuracy?  I don't know.  That's why we are doing this experiment.
2. These are for the average value of the American stock market as a whole during that month (or in the case of June, what's left of it).
3. This is NOT me showing off the models by making bold predictions to prove it works.  I know it works.  This is me trying to see how well it works when used in the worst possible ways against the most challenging possible targets.

I'm not giving investment advice here.  This is an experiment.  You should not consider this data valid until after this period of time has passed, and we can consider what actually happened to compare it to.

Thank you for this. Certain astrological predictions are on the same line as well. From economics point of view it seems like that eventually ( possibly next year) we enter hyper inflation. If you don’t mind can you run your models on gold, silver, bitcoin and real estate as well?