Here you go @traider
The way I calculated was - I've extracted equity $ per strategy for each date this year and used Google Sheet's CORREL formula for whole data series for each strategy against SPX equity (assuming it was 100% invested).
Val
Fully Automated Stocks Trading
- Thread starter ValeryN
- Start date
- Status
Here you go @traider
The way I calculated was - I've extracted equity $ per strategy for each date this year and used Google Sheet's CORREL formula for whole data series for each strategy against SPX equity (assuming it was 100% invested).
Val
Awesome! One more reason for me to keep posting
Btw, I'm pretty open to ideas to what next posts should be about.
Feel free to share your story here if you're doing something similar!
Aug 19th update over last 2 weeks:
Last week account touched a new high. Didn't check balance since. Out of individual trades shorts in GSX and KODK were biggest winners. There were plenty of smaller ones.
Surprisingly, "trend-following" shorts are not getting destroyed as market hits new highs. Which is counter intuitive. Hopefully that will continue.
Launched a new highly experimental strategy which targets penny stocks to increase long exposure and take advantage of less correlated market. Very small position size for now (relative to account size), monitoring executions and validating my assumptions.
This is a new domain for me as I barely ever touched small priced stocks and not sure what live executions will be like.
There are few ways it can go
(a) right out of the gate I'll see bad fills, missed fills, partials or commissions exceeding my model assumptions. Then it's easy - iterate on a model, refine liquidity filters and try again
(b) it will be working fine at the beginning, meaning executions will be matching model and commissions inside estimated boundaries. Then it will take 2-3 months to see if performance is degrading due to the impact on my own orders on the market
I'd say it will take at least 6 months to gain confidence even if things go well. Either way I am excited about understanding that part of the market better in the process.
First 5 trades in 2 days were looking good, matching my model perfectly. Position size was 50% of my target for this strategy. I started seeing some rebates opportunities.
Shared thoughts on this strategy with my friend, who is a veteran mechanical trader, and mentioned to him that I feel there are less sharks in this tank. His reply was - "More idiots in that tank as well".
Time will tell if that tank just got one more idiot in it
Val
Last week account touched a new high. Didn't check balance since. Out of individual trades shorts in GSX and KODK were biggest winners. There were plenty of smaller ones.
Surprisingly, "trend-following" shorts are not getting destroyed as market hits new highs. Which is counter intuitive. Hopefully that will continue.
Launched a new highly experimental strategy which targets penny stocks to increase long exposure and take advantage of less correlated market. Very small position size for now (relative to account size), monitoring executions and validating my assumptions.
This is a new domain for me as I barely ever touched small priced stocks and not sure what live executions will be like.
There are few ways it can go
(a) right out of the gate I'll see bad fills, missed fills, partials or commissions exceeding my model assumptions. Then it's easy - iterate on a model, refine liquidity filters and try again
(b) it will be working fine at the beginning, meaning executions will be matching model and commissions inside estimated boundaries. Then it will take 2-3 months to see if performance is degrading due to the impact on my own orders on the market
I'd say it will take at least 6 months to gain confidence even if things go well. Either way I am excited about understanding that part of the market better in the process.
First 5 trades in 2 days were looking good, matching my model perfectly. Position size was 50% of my target for this strategy. I started seeing some rebates opportunities.
Shared thoughts on this strategy with my friend, who is a veteran mechanical trader, and mentioned to him that I feel there are less sharks in this tank. His reply was - "More idiots in that tank as well".
Time will tell if that tank just got one more idiot in it
Val
In my opinion, when trading mechanically there are 2 most important things:
Everything else is kinda pre-requisite for those two. Like risk management, quality of backtest/data, realistic assumptions, bugs, etc. If anything is wrong there, the answer for either of those 2 most import questions will be NO, which means you have a problem.
Every 1-3 months I do a comparison and sometimes will literally compare every single trade in backtest vs live. This has to be done at shorter interval when running a new type of systems where assumptions are yet to be validated.
Here is a part of this analysis.
Most importantly - I want to make sure model’s performance looks right and deviations live vs model are within reason.
Conclusions are on the chart. Overall everything looks good.
New system for penny stocks (will call it "Long 3”) is not included as not enough data yet. I am looking at its’ executions daily for now as this is completely new territory with different liquidity. No red flags so far.
Notable changes were - my 2nd short model bothered me last month and I did ended up updating it. Most of my models have max DD 10-12% and I expect to hit up to 20% before disabling them. This one was the one exception with 18-20% historical Max DD. Normally having biggest losses when market has a big recovery rally like between 1999-2002 and 2009. So basically the same thing happened again in 2020. Meaning I had to be ready for x2 DD in worst case + it doesn’t have as nearly good recovery characteristics as my MR systems. Which I though I will be comfortable with but no longer feel this way. So far it seem to be a good call, but time will show.
This year started with 5 models. Then I merged two long ones into 1 and had 4 for couple of months. Starting this month I launched 1 more so it is better to 5. 3 long / 2 short.
PS. Re $ figures on charts - each 1k represents roughly 1% gain/loss. For example “Short 1 live” roughly gained 12% in since June 1. My software doesn’t show % equity change which would be more appropriate. So what I do is - import my live trades and “run” then against 100k$ model account.
Val
- Are your models performing well? Meaning - staying within their intended boundaries in terms of return / DD from equity high / drawdown duration
- Are our live executions matching the model?
Everything else is kinda pre-requisite for those two. Like risk management, quality of backtest/data, realistic assumptions, bugs, etc. If anything is wrong there, the answer for either of those 2 most import questions will be NO, which means you have a problem.
Every 1-3 months I do a comparison and sometimes will literally compare every single trade in backtest vs live. This has to be done at shorter interval when running a new type of systems where assumptions are yet to be validated.
Here is a part of this analysis.
Most importantly - I want to make sure model’s performance looks right and deviations live vs model are within reason.
Conclusions are on the chart. Overall everything looks good.
New system for penny stocks (will call it "Long 3”) is not included as not enough data yet. I am looking at its’ executions daily for now as this is completely new territory with different liquidity. No red flags so far.
Notable changes were - my 2nd short model bothered me last month and I did ended up updating it. Most of my models have max DD 10-12% and I expect to hit up to 20% before disabling them. This one was the one exception with 18-20% historical Max DD. Normally having biggest losses when market has a big recovery rally like between 1999-2002 and 2009. So basically the same thing happened again in 2020. Meaning I had to be ready for x2 DD in worst case + it doesn’t have as nearly good recovery characteristics as my MR systems. Which I though I will be comfortable with but no longer feel this way. So far it seem to be a good call, but time will show.
This year started with 5 models. Then I merged two long ones into 1 and had 4 for couple of months. Starting this month I launched 1 more so it is better to 5. 3 long / 2 short.
PS. Re $ figures on charts - each 1k represents roughly 1% gain/loss. For example “Short 1 live” roughly gained 12% in since June 1. My software doesn’t show % equity change which would be more appropriate. So what I do is - import my live trades and “run” then against 100k$ model account.
Val
Last edited:
Trading for me is a diversification of how I get to my ultimate wealth goals.
My plan has 3 components. Trading is one of them.
Making good money and not being obsessed with spendings I have some spare cash and don't want it do be idle. That's what I am posting here about. Wouldn't trust anyone else to manage it. In a past I had conversations with advisers, looked at funds, was approached by banks' wealth management and was shocked how little they know proportionally to how much they sell.
Few people approached me to take their money under management. But I declined so far. Emotional pressure, extra complexity, smaller potential returns with increased AUM and mostly availability of bigger upside projects are some of the reasons why. Who knows, that might change in the future.
Intellectual challenge of beating "the hardest game" there is is certainly part of the reason why I am doing this.
Val
It seems that a developer of the software I am using for backtesting (RealTest), is reading this thread! He have reached out to me after he saw my post. Wow!!! 
He immediately offered a very easy solution how to show % on those graphs I post so it is a bit more relevant. In the future that's exactly what I will be doing.
Btw, I feel incredibly lucky to know him. What I am doing is relatively insignificant comparing to his success in trading. And for a big part, is possible because of him.
He has an incredible story, which I wish he would share with a wider audience one day.
In "my book" he is one of the greatest traders out there, yet, incredibly humble. So far it has been his choice to stay in shadows so I will respect that and not mention his name.
Val
He immediately offered a very easy solution how to show % on those graphs I post so it is a bit more relevant. In the future that's exactly what I will be doing.
Btw, I feel incredibly lucky to know him. What I am doing is relatively insignificant comparing to his success in trading. And for a big part, is possible because of him.
He has an incredible story, which I wish he would share with a wider audience one day.
In "my book" he is one of the greatest traders out there, yet, incredibly humble. So far it has been his choice to stay in shadows so I will respect that and not mention his name.
Val
Last edited:
Btw, this is a great question and I appreciate that you asked.
Here is another way to put it - I believe in 2 ways of achieving ambitious financial goals:
- Using Asymmetric Leverage (AL)
- Using Compounding
--
Trading would be AL if
(a) a trader starts managing increasingly large pool of money. Then risk is limited but upside is unlimited
(b) a trader finds a method with limited risk and "huge" upside
For last few years my AL projects are outside of trading while I use trading for compounding of liquid assets.
How about you?
Val
Great answer, Val! This idea of AL and scalability also was mentioned in one of Taleb books. He gives example of writing the book which takes some finite effort but the payout is virtually unlimited and requires no further effort. He suggested to only engage yourself in this kind of endeavors.
Personally, I am trying to make enough income consistently to trade full-time, but I also need to make sure my family is not adversely affected by this quest
- Status