Reply

What’s your measure of a good strategy?

9 replies

RJL

Customer, bbp_participant, community, 70 replies.

Visit profile

5 years ago #114288

Obviously a profitable strategy is a priority, but what else do you look for when measuring a good strategy (one you’d be willing to trade live with), and do you focus analysis on IS or OOS data? 

 

A few measures I’m focusing on in my analysis and filtering out through ranking options are:

 

Stagnation % – Because I simply don’t want a strategy that doesn’t increase in profitability for months on end.

 

Stability – Similar reason to above in that I want consistent growth. So far it’s been very difficult to get a rating higher than 0.50 on OOS data. On a current test I’m running now, I’ve only 10 strategies that have met this criteria although I only dismiss those 0.4 or lower from the databank.

 

Max DD%

 

Return/DD Ratio

 

Number of trades – As I want an active strategy.

 

I might take a closer look at SQN and Expectancy in the future too, but I’d be interested to hear what others use as their primary measures of a good strategy.

 

Thanks.

0

tomas262

Administrator, sq-ultimate, 674 replies.

Visit profile

5 years ago #133282

I basically just create a strategy from scratch – idea that I know could do well and then use SQ to tweak all parameters using genetic programming IS / OOS data to the statistics in my favor – min DD %, max Win %, PF, Avg Trade, # Trades. Expectancy is a good measure too I think

The idea has to somehow work and be robust no matter what parameters are set or with wide range at least

0

stearno

Customer, bbp_participant, community, 382 replies.

Visit profile

5 years ago #133305

RJL,
I am basically looking at same things right now. But I am too looking to always find better ways to evaluate stats.

I think a few points need to be said:
1. You need to have goals set that you want when building a strategy. (maybe you have some good strats, but now looking to fill out your portfolio with an off hours range strategy. Then you would need to change your evaluation numbers because you having different needs)
2. Your strategies have to fit you. So a few pieces of your preferences showed in your post : you want it to be active and you want consistent winning. Therefore, you need to match your evaluation numbers to match yourself.

I met one guy and he wants max percent return. Bottom line, most profit. So he is willing to go with 50% drawdown.

Another guy said his strategy’s profit really boils down to 3-4 trades in a year. So his win percentage will be 25% and mostly small wins with small losses.

Another guy might want a large stop loss so that his win percentage is high.

Another guy thinks it only makes sense to have a 2:1 reward to risk. And he does not want to see a week’s worth of profit lost in one trade.

So I think that we should keep in mind that another man’s evaluation numbers might not be a right fit for our own personality or our own goals. But I do believe we can learn from others so it’s good to find out but use only that to which is in line with our goals and personality.

-Stearno

Sent from my HUAWEI MT7-TL10 using Tapatalk

0

seaton

Customer, bbp_participant, community, 161 replies.

Visit profile

5 years ago #133330

These are what I use for initial select. 

 

Net Profit per year > Min 5K  pref 10K     (trading 1 Lot on 10K account) 

R Expectancy > 0.1 
Profit Factor > 1.0
Average profit/Trade > $50
 
then some visuals filtering on the results such that Equity curve is fairly linear and smoothish, i.e. not sharp bounces up or down.
 
I have also recently started experimenting with addition of the following:
 
win rate > 55%  pref 65%
Holding time < 5 bars on D1
 
I don’t look at things like drawdown in my initial selection criteria as this will depend on 1) personal risk tolerance 2) money management and 3) account size.  All I’m worried about initially is potentially profitable strategies.
 
If the strategy is showing is profitable then I adjust to my personal risk level by using SQ analyser MC with different lot sizing or risk so that WC is within my Risk tolerence for my account size.
 
Stephen.

0

seaton

Customer, bbp_participant, community, 161 replies.

Visit profile

5 years ago #133395

What I’m now trying to look for is strategies that have high win rates and short holding times.  Ive seen in real life through my trades that longer holding periods can turn against you and effectively wasted the trade after a long period with either little return or a loss, where effectively I could have a higher % of winning short term trades that trade more frequently which will get better compounding effect over the long term.

 

A good read are the books by Howard Bandy, while not forex, they deal with automated trading and Quant.  

0

munchie

Customer, bbp_participant, community, 63 replies.

Visit profile

5 years ago #133396

Hi seaton, good input, I’m interested in getting one of Howard bandy’s books, specifically which one would you recommend as good for the beginner?

Sent from my iPhone using Tapatalk

0

seaton

Customer, bbp_participant, community, 161 replies.

Visit profile

5 years ago #133398

@munchie  They are all good and cover different areas with a bit of crossover between them especially his core principals.  

 

Hmmm for the beginner, not sure as although he explains things well, there is maths, stats and programming.  They are very Amibroker focused and more on indexes throughout the book, but can be applied to currencies.   I ended up buying all of them at the same time and did not regret it. 

 

Probably the two I find myself continually referring to presently are

 

Quantitive Technical Analysis

Quantitive Trading System

His Mean Reversion book is good as well.

 

I wish he would bring an ebook version out as I carry at least one around where ever I go to read at lunch etc.  would be nice to have them all on my iPad.

 

Stephen….

0

clonex

Customer, bbp_participant, community, contributor, sq-ultimate, 263 replies.

Visit profile

5 years ago #133399

I have this one and it is valuable Modeling Trading System Performance

0

munchie

Customer, bbp_participant, community, 63 replies.

Visit profile

5 years ago #133400

Thanks to both of you guys! Will order them today, as I need to get my head around this

Sent from my iPhone using Tapatalk

0

seaton

Customer, bbp_participant, community, 161 replies.

Visit profile

5 years ago #133578

…now we just need a export option to Amibroker 😀

0

Viewing 10 posts - 1 through 10 (of 10 total)