kevinkdog 

It is good to see you take @jamrock 's post to heart. Whether you agree or disagree with his points, things he says should make you think. Nice post @jamrock , and nice reflection on that post @Bionan !

Thanks, Kevin. I was actually hoping that you would be one of the members who would reply to my post, because, as I stated in my reply to jamrock, I am using your book, and I have found it very useful. And for the benefit of everyone else, no, I do not receive anything for mentioning Kevin's book.

In fact, I did take the post by jamrock to heart. My wife did not actually think I should stop trading, but she believes, as he seems to be stating, that I should incorporate discretionary aspects into my trading strategies, and that I should not rely solely on automation.

This creates a problem for backtesting. The discretionary part of trading involves human pattern recognition, and I have found it to be almost impossible to code. Take today's (10/06/2020) MNQ, for example. The daily ETH chart shows that we were at the top of an almost month-long trading range. The only way I would want to take a long trade today would be on a successful breakout above 11600, and even then a reversal would be high probability. I can see this, but I cannot figure out how to have a computer recognize it. As a result, my trending strategy would have made four losing long trades today. The only thing I could do is manually disable long entries.

Not to hijack my own thread, and if you want me to ask this question in a separate post in order not to disturb the fine moderators on this forum I will, but in order to maximize position sizing on a $100,000 account I use the MNQ rather than the NQ for this strategy. If I switch to the NQ I am sure my results will be entirely different. But the MNQ has only been in existance since summer of last year, so I cannot use out-of-sample data to backtest it. If I change the account size in order to change to NQ, the backtest results will change, especially because I am using tick charts. Do you have any suggestions?
Again, let me know if I should re-post this question.

Thanks-
Rick

waylon 

Hi Rick, I have developed about 40 strategies - and the average time to develop and test each one has been around two months. However to develop one that is consistently profitable in live trading over one hundred trades has taken 10 years- so do persevere.

Thanks for your post. I am hoping to slightly speed up the process. At my age I may not have 10 years!

Rick

MY main point was not that you wasted your time. I dont believe you wasted your time learning or watching or trading the markets but based onwhat you wrote and i know
nothing really about what you did or what your future plans are with trading and automation;however, automation WILL NOT make an unprofitable strategy profitable because you have cut out emotion and no indicator is so custom that it will work if automated. if you now have these so called amazing k evn indicators which this is beginnign to sound
more and more like a plug for this book anyway if yuo ahve these new tweaked indicators why not manually trade them immediately and see if they help you and if so why even automate.

I am not sure if you realize most automations in trading are due to speed where the strategy loses money if it desnt get price and time priority or if it is taking liquidity aggressive then it must rapidly exit in a very small window of profitability or it is gone. Automation has nothing to do with peoples emotions and what i mean by that is no on ein the big leagues is automating due to emtional reasons. just like any other real job. if you are clicking a mouse or a keyboard to trade and it is short term intraday or weven wing weeks then automation will not help you at all. do you really think hft started out mouse clicking? soes bandits maybe but doubtful. my point is that those firms make money before 1 tick is drawn on your chart before you can blink physically they live in a realm of automation because humans cannot see anything that fast and react on it.

I do wish you the best but if it doesnt matter that much then why blog about it at all?

If i work in a restaurant and im tired of some guy asking me 1000 questions and constantlty bothering me i cannot just lash out at him and call him an a hole because i will lose my job. if I am working prop or for a hedge fund and i am manually making trades i cannot hold onto a loser beyond my risk point becuase i will LOSE MY JOB. maybe
you should be a little more honest about your reasons for wanting to automate because you already do not have a winning strategy and you are searching still and you can search for 20 years jumping around to the next latest and greatest and end up with a whole lot of lost time and energy.

it sounds to me like taxes are not a worry for you because you are not profitable.
sounds to me like this is your hobby so why does your wife care so much if it is just a hobby.
sounds to me like you arent being hoest with anyone onthis blog or maybe yourself
soudns to me like you are not able to figure out the market on your own and you seek advice from boosk and fancy indicators and blogs to steer you somehow
but let me tell you this. If i put you in a kitchen with all the right ingredients and tools will you know how to make me a cake without a good recipe?

NO you will not! indicators are tools like a knife or a mixer and you must figure out how to make a profit out of it which means that even after
you figure out how to make a cake you must then figure out how to sell that cake and profit because there are 10,000 people maybe a million who learned how
to make a cake on their own way before you and we are making money by taking your money. undeerstand?

ONCE you actually get consistently profitable you then have a real BELL CURVE a gaussian distribution of the winners and what they make?

where do you think you will fall into the distribution of the winners? everyone thinks oh 90 % lose at trading the other 10 % are wealthy nope. no way.
the distribution of huge winners to break even traders or small winners is enormous and the middle is just that the middle.

Go on and do what you are going to do with trading and yes you do learn a lot over the years but warp speed is what you need at this point because
as you automate and as you learn more the more mixed up you will be.

mtzimmer1 

I've written a handful of strategies that generate 'good' (subjective) results on the longer-than-intraday timeframe but I have yet to develop even ONE intraday strategy that tested well enough live to warrant scaling up. I've probably tested and thrown away well over 100 intraday strategies over the past year and a half. In stocks, I feel that it is much easier to find edges on longer timeframes, especially if trading long-only. -Zimmer

Thanks for the reply. I have heard that it is easier to trade long because the motivating factor of the trend is greed and FOMO. The motivating factor of shorts is fear of loss, and it is a much stronger motivator. Therefore, short trades tend to move much faster and are shorter in duration. Automation or strict adherence to trading rules is paramount, because any delay in entry could result in a loss.

We have been blessed with a huge secular bull trend since Obama's election and extending through Trump's term. I wonder how a long-only edge will hold up if we move into a protracted bear market or trading range market. Since most of my funds are in retirement accounts, the government, in all its wisdom, as decided that shorting stocks is too risky. So I am left with two MORE risky options- either buying puts, selling cash covered calls, or trading futures.

-Rick

no need to delete the book from the posts all good but about that book how can anyone write a book about trading systems and even have the
word BOOTSTRAP in the index?

all in all it was a pretty good book

Bionan 

Thanks for the reply. I have heard that it is easier to trade long because the motivating factor of the trend is greed and FOMO. The motivating factor of shorts is fear of loss, and it is a much stronger motivator. Therefore, short trades tend to move much faster and are shorter in duration. Automation or strict adherence to trading rules is paramount, because any delay in entry could result in a loss. We have been blessed with a huge secular bull trend since Obama's election and extending through Trump's term. I wonder how a long-only edge will hold up if we move into a protracted bear market or trading range market. Since most of my funds are in retirement accounts, the government, in all its wisdom, as decided that shorting stocks is too risky. So I am left with two MORE risky options- either buying puts, selling cash covered calls, or trading futures. -Rick

um you can short futures without margin an use them as a cash short or long . the margin levels are set by the fcm and can be lower or higher than the exchange

with nasdaq christmas contracts (dec) trading at 11,270 the micro equiv is 11,270 x $2= $ 22,540 so for every 22,540 in your account you buy or sell a micro nq. that would be a zero leverage trade. why do this well you get tax benefitsw from trading futures if it is short term trading and you free up capital to earn interest on (small now) versus buying qqq or spy or you can buy bonds and pledge those as margin also for futures

kevinkdog 

Have you thought about taking signals from NQ, and just executing on the MNQ?

For a while, when the MES was new and more thinly traded, in NinjaTrader I overlaid the ES chart over the MES, made the MES invisible, and traded MES signals based on the ES. With the arbitrageurs trading the markets, I never saw more than a tick or two difference between the two contracts. But it became a nightmare once the roll date came around, because I had to reconstruct all my charts. Fortunately now the micro volumes are high enough that there is no need.

To answer the intent of your question, unless I am missing something in your reply, the signals are not a problem. The strategy is based on trading two positions per signal. Each has a 1.5 ATR SL, one has a 2 ATR PT, and the other is a runner with a 5 ATR PT so I can make on-screen adjustments if desired. Each position risks 0.5% of total account size, so each trade risks 1%. The position size is weighted more toward a 1 ATR profit target, so as long as I get 1 ATR out of the trade a loss on the swing portion is small. If I use the NQ contract, there will be a tighter distribution of contracts in each position, and some higher ATR trades might only be able to risk one contract instead of two. This will drastically change the outcome of the strategy. Commission will be lower, slippage will probably be less, but the ability to fine-tune position size will disappear.

So I am back to my original issue- how to backtest a contract with only 15 months of historical data.

Thanks for your reply.
Rick

Bionan 

For a while, when the MES was new and more thinly traded, in NinjaTrader I overlaid the ES chart over the MES, made the MES invisible, and traded MES signals based on the ES. With the arbitrageurs trading the markets, I never saw more than a tick or two difference between the two contracts. But it became a nightmare once the roll date came around, because I had to reconstruct all my charts. Fortunately now the micro volumes are high enough that there is no need. To answer the intent of your question, unless I am missing something in your reply, the signals are not a problem. The strategy is based on trading two positions per signal. Each has a 1.5 ATR SL, one has a 2 ATR PT, and the other is a runner with a 5 ATR PT so I can make on-screen adjustments if desired. Each position risks 0.5% of total account size, so each trade risks 1%. The position size is weighted more toward a 1 ATR profit target, so as long as I get 1 ATR out of the trade a loss on the swing portion is small. If I use the NQ contract, there will be a tighter distribution of contracts in each position, and some higher ATR trades might only be able to risk one contract instead of two. This will drastically change the outcome of the strategy. Commission will be lower, slippage will probably be less, but the ability to fine-tune position size will disappear. So I am back to my original issue- how to backtest a contract with only 15 months of historical data. Thanks for your reply. Rick

you have 8000 stock symbols and etfs. im moving on from this post but if you cannot use stock and etf datat to back test a trading strategy then you probably shouldnt be trading period. also you have inverted etfs and even lgverged etfs so not sure again how one could not back test something and believe it with all of the almost free data we have and all of the different ways to stress test your strat especially after the corona virus sell off and the 2008 banking mortgage issues but again seems most people dont want to actually utilize their thinking ability.

kevinkdog 

@Bionan - Please edit your posts to remove all references to that book. Apparently some misinformed people think it is a sneaky marketing ploy.

Done. I removed the title. I have seen the fact that you wrote a book mentioned in posts by other people, but not the title. If you want me to even remove the references in my posts that you wrote a "book", let me know, and I'll remove those, as well.

saying it was a plug for the book was obviously not meant to be literal! lol