8 day stock prediction test


So I wanted to try a little experiment… David was kind enough to pick 10 stocks at random from different sectors in the FTSE with values >300 points.

What I was attempting to do was predict the daily results (up/down) on the previous day and also assign a ‘confidence score’ to rate what I felt the chances were of my prediction coming true. This is only up/down on the previous days close rather than an overall direction over a period of time. What I’ve not yet done is looked at aggregate values against the stocks final resting place. That’s what your P&L is for – hahaha.

My tools were only the previous days candles (pricing) and looking at the FTSE 100 index. I also recorded my results for predicting the movement of the 100 index too.

Just a note on this chart – cells not coloured green should be red i.e my prediction was wrong. I didn’t colour them in red though because the whole thing would descend into an unreadable mess – see the attachment 🙂

So this generated some interesting results which I’ll get to in just a minute.

Now, one thing which someone could argue may screw the results is that the Japanese Tsunami hit on the first Thursday/Friday as I begun this experiment. Now it’s likely that this did indeed screw the results but in a positive way i.e. the movement/psychology of the market became more pronounced and therefore easier to read. Even with this factor I do believe the overall conclusions from the data would be the same.

My reason for this is that what I’m actually looking at is my perception of the market and so unless my perception radically changes I’d suggest that overall the results would be the same if I did this exercise again. Of course I’m going to do this another time =)

So what did I find?

I’m correct 60% of the time, purely looking at the candles in predicting whether a stock will go up or down the next day

Aligning the results with the direction of the trend (down) and re-calculating gives 80% accuracy without any thinking – this is a 25% improvement on a daily basis over my own results

Confidence in a given result seems to be irrelevant. My confidence in making predictions actually decreased greatly as this exercise progressed. Looking in the last column you see some very low marks (20%) that turned out to be correct.


There’s clearly no point at all taking a trade in the opposite direction to the trend. Seems like money flows in and out of the stock market like the tide so in a down trend people only buy back in when they see real value to be had. In other words when greed outweighs fear! Picking stocks moving in the same direction as the trend gets you a +25% chance that the trade will be a success – obviously worth having 🙂

Confidence or maybe ‘attachment to outcome’ is not a good thing to be thinking about when looking to analyse the movement of a particular stock. I think everyone usually starts this business with an apprehension about going short. This might be more difficult to overcome if you’re a naturally positive person! Attachment to the outcome of a particular decision needs to be removed in order to understand or react to what is going on as early as possible.

Mind like water grasshopper!


One thought on “8 day stock prediction test

  1. Pingback: Emotional roller-coaster survival « robertsweetman

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s