Fully Automated Stocks Trading

[ValeryN]

I guess there are no shortcuts to developing strategies. Since I’m not a mathematician I would need to piggyback on other resources for ideas and strategies. Is there a library or community out there with strategies to play with? Examples included in real test are a great start wondering if it’s possible to find a larger library of ideas to play with or is it just read books? Trying to avoid the endless search in the wrong direction.

mhparker is absolutely correct. This is a common miss-conception that you need complex math for this. All the math involved is super basic, basically if you're able to make the right choice between a / b - you're good, where:
a. A strategy with 30% win rate and 4:1 win to loss
b. A strategy with 90% win rate and 1:20 win to loss

There are things where complex math is involved, but I don't do any of that. Even correlations you can just visually inspect in RealTest. That also gives you unique insight on when you can expect outliers.

The rest is mostly common sense. Like - will you ever put 50% of your capital on a single stock? Even if your strategy tells you, over 20 years it would have worked.

Reading RealTest manual, especially sections on backtesting engine and handling of ambiguous situations is incredibly insightful.
I think Marsten should charge 10k$ just for the folder with his strategies examples. There is a mountain of knowledge he accumulated over 20+ years in there. I wish it would be available when I started.

Val

 

[ValeryN]

Just noticed this chart today as a short position in it got closed. Jee..
I can't fathom the amount of speculation going on to create something like that.

Val

 

[HobbyTrading]

I can't fathom the amount of speculation going on to create something like that.

With a ticker name like that it is not surprising to see speculation.

 

[Nathan07]

I must have read your response several times to really let it sink in this journal is a gold mine thank you. I was wondering if you can recommend a book that focuses on systems that trade IPOs they have little backtesting that can be applied to them.

 

[easymon1]

mhparker is absolutely correct. This is a common miss-conception that you need complex math for this. All the math involved is super basic, basically if you're able to make the right choice between a / b - you're good, where:
a. A strategy with 30% win rate and 4:1 win to loss
b. A strategy with 90% win rate and 1:20 win to loss

There are things where complex math is involved, but I don't do any of that. Even correlations you can just visually inspect in RealTest. That also gives you unique insight on when you can expect outliers.

The rest is mostly common sense. Like - will you ever put 50% of your capital on a single stock? Even if your strategy tells you, over 20 years it would have worked.

Reading RealTest manual, especially sections on backtesting engine and handling of ambiguous situations is incredibly insightful.
I think Marsten should charge 10k$ just for the folder with his strategies examples. There is a mountain of knowledge he accumulated over 20+ years in there. I wish it would be available when I started.

Val

10k$.
How many hours for you to surmount RT learning curve and get traction? 10, 30, 100?
(Other tool recos?)

 

[ValeryN]

I must have read your response several times to really let it sink in this journal is a gold mine thank you. I was wondering if you can recommend a book that focuses on systems that trade IPOs they have little backtesting that can be applied to them.

Can't think of any particular book. I suggest to define a criteria of what you call an IPO and "realtest" this universe. You are probably looking for stocks that have started trading under certain number of days and have sufficient liquidity.

Val

 

[ValeryN]

10k$.
How many hours for you to surmount RT learning curve and get traction? 10, 30, 100?
(Other tool recos?)

Really depends on where you are coming from.

If you never backtested anything then most of the time with any tool you will spend on going thru basics of where get your data from, what is sufficient quality, what is sufficient test, what is realistic test etc.

After having all my strategies in AmiBroker & running live - it took me about one weekend to switch to RealTest and then couple of weeks to entirely convince myself I want to do 100% transition there. As you can imagine, when you invested years in one tool it takes a lot to move to something else. So those "2 weeks" convincing period speaks for itself. As value & simplicity I gained from the switch were much greater than holding onto AmiBroker specific knowledge.

If you're starting from no tool today, then I'd imagine you will need about a month of tinkering with data / strategy examples to get used to RealTest. Which is, IMO, way less than for other similar tools.

Couple of big questions you will need to ask yourself -

1. Will EOD data be sufficient for you?
2. Do you want to invest into high quality EOD data?

#1 - works for me. I use to use intraday but then ditched it eventually after realizing that EOD was sufficient for me. No doubt there are edges intraday, but there are also edges on EOD data.
#2 - using Norgate vs Yahoo might be an eye opener when running backtests for longer than 3-5 years intervals. Especially if you are using index constituency + impact of delisted symbols will add up. These days I would not consider any results based on Yahoo data valid.

Val

 

[easymon1]

Thanks

 

[ValeryN]

Thought I haven’t really posted individual strategies visuals in a while.

Btw, I found and fixed the problem with that made my RealTest equity differ from IB lately, it was a either a silly copy/paste error or a corrupted file as a result of multiple kernel-panics I’ve had over last month while sitting on MacOS Big Sur beta… Anyway, I found and eliminated duplicated trade records which were causing that discrepancy and now it looks very close to IB report.

All “red"ish lines are short strategies. Others - longs.

It is interesting to see that my top short strategy performed almost as well as a top performing long.

DiscretionL/S ones are how I track my discretionary trades or overrides if any, split into longs / shorts. They are pretty rare. BIGC short was the biggest one in August, JKS is a recent short I took some beating on and closed on Friday.

The bunch of other strategies hovering around 0% is a great example of “ideal” behaviour for a strategy that is waiting for its’ favourable conditions. It is unrealistic to have a single strategy making money all the time, even making money every year is hard. One way to deal with this is to

1. Develop a strategy for a specific market type or a phase or an anomaly in a sub-universe, and find some reasonable balance between how much it will be loosing while waiting for it vs how much it makes. Some people use market filters to decide when strategy shouldn’t trade - I personally don’t do it. My strategies will be trading all the time as soon as they find their setup
2. Develop as many such strategies as possible
3. Use a system of system = combined strategy, that will consist of those which have low correlation of drawdowns

Val

 

[ValeryN]

After some Saturday morning trading routine, I'm off paddle boarding!

Hopefully today will be as nice as the last time.

Val