What’s the correct way to manually backtest?

schweiz said:
Can you do with less resources what they cannot?
More...
Hello schweiz,

Of course I can. Those people are scalping around to make money. There is more then one way to make money trading besides HFT.

It's not as hard as you making it. You manual back test a strategy (for free) to see if it's profitable, then trade it to make money. Scale up as needed.

The same for programming a strategy, program it and back test it. If it makes money in the past, trade it. Each way can be done for free. Who cares what the hedge HFT people are doing? That's them. All my backtesting is free of charge. I can careless what hedge funds HFT people are doing. Never even thought about them.

You said to many people dream of being rich trading. I certainly disagree. The goal is to be rich and wealthy as fast and as easy as possible
Period. I will always choose the easy and fast way. And if I die without getting close, atleast i know I tried and had fun while doing it.
 
Last edited:
Surprised when I saw your 10,000 hour number, because it's exactly the same number of hours +/- a small amount, it took me to develop the tools and knowledge to be profitable. Did it in two years of long hours, with about 2,000 Adderall for the ADD.
 
SimpleMeLike said:
Hello schweiz,

Of course I can. Those people are scalping around to make money. There is more then one way to make money trading besides HFT.

It's not as hard as you making it. You manual back test a strategy (for free) to see if it's profitable, then trade it to make money. Scale up as needed.

The same for programming a strategy, program it and back test it. If it makes money in the past, trade it. Each way can be done for free. Who cares what the hedge HFT people are doing? That's them. All my backtesting is free of charge. I can careless what hedge funds HFT people are doing. Never even thought about them.

You said to many people dream of being rich trading. I certainly disagree. The goal is to be rich and wealthy as fast and as easy as possible
Period. I will always choose the easy and fast way. And if I die without getting close, atleast i know I tried and had fun while doing it.
More...

Even when you been trading for good long while, I will crash and burn while developing a system, recently spent two months and 400 hours on a system to make 2-4 ticks per profitable signal, main part was to make 2 ticks. It sounds like it be easy to make 2 ticks in ES, but harder than one thinks. But there has to be a happy balance between protective stops and profit. Usually my stops are always far greater than profits because of noise and this works in all systems that I have automated for intraday trading. But there is a time when you have to pull the plug and just throw away failure number 1 and go different direction. One problem was system got too complicated, second problem I didn't take into account ranges of # of bars, and on and on. So I threw away trendlines, indicators, Gann, and we start again. I was in papertrading phase when two losing days wiped out 2 weeks of possible profits. So you take what you have learned and develop something else and have formed an idea, went over today's data, and will start back testing few months then papertrade it for 3 months and if happy it gets programmed. I been doing it this way for 25 years and works for me.

I think rich is an illusion, being able to do what you love and have enough to do whatever you like from it, enough for me.
 
Just an FYI. Multicharts is running a Black Friday sale (33% off) for lifetime licensing until Dec 1st. It's excellent retail grade software, with the ability to test/trade tick data. I have no affiliation with them. It's a great product for the lifetime price.
 
Handle123 said:
I think rich is an illusion, being able to do what you love and have enough to do whatever you like from it, enough for me.
More...

Hello Handle123,

You are correct. Freedom to do as please and work for what you want is rich for me.

Programming and back testing is the best advice you ever could have given me months ago. YES YES, it is very time and problem solving, but I can see the profitability of a system in minutes rather then guessing when manual back testing. For me, it's a fine balance.
 
The real backtesting happens once it is programmed which might take several months, then take fifteen years of tick data, so it is not just a few months, I and my buddy will have at least 1500-2000 hours into a system before it ready for sim automation, and not a weekend, LOL.
 
Handle123 said:
Even when you been trading for good long while, I will crash and burn while developing a system, recently spent two months and 400 hours on a system to make 2-4 ticks per profitable signal, main part was to make 2 ticks. It sounds like it be easy to make 2 ticks in ES, but harder than one thinks. But there has to be a happy balance between protective stops and profit. Usually my stops are always far greater than profits because of noise and this works in all systems that I have automated for intraday trading. But there is a time when you have to pull the plug and just throw away failure number 1 and go different direction. One problem was system got too complicated, second problem I didn't take into account ranges of # of bars, and on and on. So I threw away trendlines, indicators, Gann, and we start again. I was in papertrading phase when two losing days wiped out 2 weeks of possible profits. So you take what you have learned and develop something else and have formed an idea, went over today's data, and will start back testing few months then papertrade it for 3 months and if happy it gets programmed. I been doing it this way for 25 years and works for me.

I think rich is an illusion, being able to do what you love and have enough to do whatever you like from it, enough for me.
More...
What's your best system for a stock index compared to dow or sp returns?
 
For the OP, I would also recommend for each trade record the Maximum Favorable Excursion (MFE) and the MAE Maximum Adverse Excursion ( MAE) during a mannual back test.

Also, add a comment column incase you think of an idea or opinion when back testing. I would also add a column for market condition for the trade.

When I was manual backtesting I would always think of something else that may work.

I would test for about 200 trades and then calculate positve expectancy (and other profitbailty metrics) to determine if it's worth continuin to back test. You can also do a monta Carlo on those trades as well for more model testing.

You also want to know biggest loss and winner. There is a guy who sales a spreadsheet for manually backtesting I recall..very very good spreadsheet. Do a google search.
 
You must log in or register to reply here.
Back
Top