[ImXotep]

Are you interesting in next day direction (Up|Down) forecast?
Small introduction:
I'm financial analyst. Last 8 month I'm try to construct market direction forecast based on fundamental information. At this time I have one model for MSFT (Microsoft).This model have accuracy nearly 87-92%. This model don't construct for prediction price, it's not possible. I want to know,if this information is valuable for traders? Next step in this research i want to collect a list of market instruments for analyse based on your preference (near 3-5 symbols). For trial period (near month) i want to publish forecast on this forum. I will be very grateful to administration of this web site for help in organization of this experiment. You can post your ideas,wish for forecast and market symbols in this threat.
Thank you!
PS: Sorry for my bad English.

[thrunner]

Originally Posted by ImXotep »

This model have accuracy nearly 87-92%. This model don't construct for prediction price, it's not possible. ...PS: Sorry for my bad English.

How do you measure accuracy if you don't have a price? Are you predicting what the analysts are predicting? More specificity would be great because readers will be less skeptical.

[ImXotep]

I'm calculate accuracy like this:
i compare forecast with fact. if forecast on next day is up and fact is up, then this is true forecast, else error forecast.
In up direction I understand next day close will be highest current close, similar for down direction.

[MrPaul]

Originally Posted by ImXotep »

Are you interesting in next day direction (Up|Down) forecast?
Small introduction:
I'm financial analyst. Last 8 month I'm try to construct market direction forecast based on fundamental information. At this time I have one model for MSFT (Microsoft).This model have accuracy nearly 87-92%. This model don't construct for prediction price, it's not possible. I want to know,if this information is valuable for traders? Next step in this research i want to collect a list of market instruments for analyse based on your preference (near 3-5 symbols). For trial period (near month) i want to publish forecast on this forum. I will be very grateful to administration of this web site for help in organization of this experiment. You can post your ideas,wish for forecast and market symbols in this threat.
Thank you!
PS: Sorry for my bad English.

To be completely honest here. A directional bias based on fundamental information that has only been tested on a few stocks for about 20 days isn't going to get very far.

You're going to need at least 100-120 days of reliable data to impress anyone. 20 trading days could be just to close to randomness imo.

Good luck though!

[ImXotep]

OK. I'm ready. Now I need to collect most popular trading symbols list.

[waveslider]

I think you would need at least 500 data points... but what is most important is how it acts real time, going forward. The idea has to work in various market conditions also.

[Bannor]

I think it is more helpful to track the market in realtime, and predict the short term (think 15-minute bars) general market direction, then to know what will happen tomorrow.

After all, either I am holding swing trades, which means I must have the intestinal fortitude to hold through my prediction to the target or get out at my exit - or I am looking to daytrade a setup. Knowing how tomorrow is gonna be ahead of time might not be so useful.

However, I have found a method of predicting current day short term direction. Its not my idea, its the idea of a web-talk-show guy named the Shadow Trader, using the breadth concept. Basically, tracking upvolume (volume on buying behaviour) versus downvolume (volume on selling behaviour), comparing these two line to each other in realtime, monitoring the direction and/or gap. Combining this with TRIN and TICK works very well for me in gauging the markets shortterm (intra day) direction.

[ImXotep]

Originally Posted by Bannor »

However, I have found a method of predicting current day short term direction. Its not my idea, its the idea of a web-talk-show guy named the Shadow Trader, using the breadth concept. Basically, tracking upvolume (volume on buying behaviour) versus downvolume (volume on selling behaviour), comparing these two line to each other in realtime, monitoring the direction and/or gap. Combining this with TRIN and TICK works very well for me in gauging the markets shortterm (intra day) direction.

Can you give a url on this method?