58 per cent of chance is a good percentage to start trading?

Jan 4, 2019
2
0
1
#1
Hey all,


I'm quite new on this - but here we goes my question:

I analyzed a sample with about 20.000 stock prices (stocks of 30 companies from US stock exchanges markets)
I developed a kind of machine learning with statistical analysis behind and after 2 days I got my algorithm which is predicting short positions with a percentage of correctness = 58% -> result came after test the algorithm against sample of 2000 stock prices entries.
My algorithm correctly predicted short positions in 58% of the time and was wrong 42% of the time.
So my question is the following: 58% of chance is a good percentage to start trading? My algorithm has any value?

Thanks for the help,
Do you have any kind related experiences?
 

tomorton

Well-known member
Feb 28, 2002
7,061
916
173
62
Exeter
#2
Win rate isn't going to be a great guide to profitability, and I suspect its derived almost entirely from an entry pattern, which is also a poor guide to profitability. Risk / money / position management rules will be absolutely vital for this approach to work.

For example, if you have 420 losers per 1000 trades, how would your account look if you got 100 straight losers from Day 1? Not likely perhaps but you might see what I'm pointing to.
 
Jan 4, 2019
2
0
1
#3
Win rate isn't going to be a great guide to profitability, and I suspect its derived almost entirely from an entry pattern, which is also a poor guide to profitability. Risk / money / position management rules will be absolutely vital for this approach to work.

For example, if you have 420 losers per 1000 trades, how would your account look if you got 100 straight losers from Day 1? Not likely perhaps but you might see what I'm pointing to.
Thanks a lot!! I think I will have to do some stress testing to see if the algorithm is consistent enough then.
Position management will not be a concern I think so. The goal is to buy and sell in precise time periods without looking to price. It's pure analytical methodology. I'm thinking about to trade always the same amount of money.... I will test it further. Hope I can bring good news.
 

NVP

Well-known member
Jun 21, 2004
35,851
1,737
223
west sussex, UK
fxcorrelator.com
#4
hey there

assuming the 58% relates to a distinct movement of the instrument enough to accommodate a 1:1 risk return ......then it’s a good start

i known many traders who do well under 50% due to good risk returns letting the trades run

so what next ?

N
 
Likes: tomorton