The Practical Impossibility Of Meeting Backtest Results

Chart Forum » Trading - Systems » The Practical Impossibility Of Meeting Backtest Results

«« Previous Next »»

Thread Start New Thread
Last Poster Posts Pages Last Post

Author

Message

ohkoolnutz

Member

Post Number: 366
Registered: 10-2005

Rating: N/A
Votes: 0


Monday, December 18, 2006 - 04:04 am:

I have taken 18 months to create a strategy that is finally showing me the positive movements one may expect from trading. I focused on perfecting the entry and using the tightest possible stop losses and at the same time try to not have the stop losses be activated. I have no other exit clauses in my trading. There are no trailing stop and no profit target. I use price action as primary trigger for entry. I only use indicators as monitoring tool but not as trigger. This causes my trades to be large in quantity when I start out and reduce to nil once all cash is used up and all positions that were established remain above my stop loss.

I have also found that 90% of all profitable positions I ever held and sold have moved higher without breaching my buy price meaning any retracements that came were corrections. Is 'buy and hold (profitable positions)' really dead?

Backtests have no financial limit. They say to take each signal. Each trader has a limited bankroll. How could backtests give one the true picture of a strategy if they don't take into account the limited size of the bank? Isn't it therefore impossible to take each signal? If it's impossible, isn't it impossible to achieve the performance of a backtest?

As I have run out of cash and unprofitable positions what should I do?

1. Do nothing.
2. Stop paying for the IC subscription. =)
3. Reverse my entry system into an exit system.
4. Exit the oldest position.
5. Exit the youngest position.
6. Exit the position with the smallest annual return out of the oldest five positions.
7. Exit the position with the smallest annual return out of the youngest five positions.
8. Exit the position with the greatest annual return out of the oldest five positions.
9. Exit the position with the greatest annual return out of the youngest five positions.
10. Exit the position with the smallest annual return out of all positions aged 5 days or more.
11. Exit the position with the greatest annual return out of all positions aged 5 days or more.
12. Exit the position that has shown the smallest annual return for the last five rolling trading days.
13. Exit the position that has shown the greatest annual return for the last five rolling trading days.
14. Exit the position that has shown the smallest annual return for the last five rolling trading days, excluding negative returns.
15. Exit any position not showing 20% annual return within 21 days.
16. Exit any position with 7 consecutive higher closes.
17. Exit any position with 7 consecutive lower closes.
18. Exit each position after X number of days.
19. Develop an exit system based around trailing stops and/or profit targets.
20. Only deploy 80% of your cash using the strategy and use the other 20% to trade around the long-term positions in their short-term swings by doubling up on down days and cutting them in half on up days.
21. Add new positions using your monthly earnings from your day job (one trade per month if lucky) until you can live of the dividends.
22. Use the time off to find a better-paying day job.
23. Try to find a paid job in a firm where you can trade your strategy using your money and other people's money.
24. Start trading stocks on margin.
25. Change to trading stocks using CFDs.
26. Find out the probability of my system for each original position (X) to reach a specific profit percentage (K) and create an exit on the resulting position (Y = X*(1+K)) as follows:
Sell (X*(1-((Y-X)/Y) (aka X aka the original stake)
when the probability for K to increase is less than 52%.
(It sounds very logical but too complicated for my limited capabilities.)
27. Wait for the next market correction to occur to recognise flaws in my strategy.

---
ohk

Lies, Damn Lies and Technical Analysis

tbreak

Member

Post Number: 84
Registered: 06-2004

Rating: N/A
Votes: 0


Monday, December 18, 2006 - 10:53 am:

Wow,

You've got a few questions, Some of them are major to the point they could be life changing.

22. Use the time off to find a better-paying day job.
24. Start trading stocks on margin.

What exactly are you wanting someone to say?

By the way there should be 28 questions.

28. Sell and give profits to Tbreak

perler59

Member

Post Number: 905
Registered: 09-2003

Rating: N/A
Votes: 0


Monday, December 18, 2006 - 10:52 pm:

Hi Kool. Any backtester worth its lines of code takes into account the total funds available for trading and how much bank interest you are losing while you hold a position.

A backtester has a difficult problem to solve when it has available funds to enter, say, two more positions, and five entry signals appear at the same time! How to chose?

I see your point about what to do when you get an entry signal and you have no funds. I think you need to use back testing to develop an algorithm for weighting each position (open and pending) and always exit the lowest weight positions to allow you to enter higher weight (by a big margin?) positions.

EDIT: The Turtles used to go into maintenance mode when they were "Fully loaded" and only exit positions when their system told them to. When they had no available funds, they stopped looking for new opportunities.

(Message edited by perler59 on December 18, 2006)

http://sttc.net.au/~stever

eric

Member

Post Number: 58
Registered: 09-2006

Rating: N/A
Votes: 0


Tuesday, December 19, 2006 - 09:15 am:

Perler, when I am fully invested, as I am at the moment, I simply don't invest any more. I don't exit just because I see a good opportunity and need more funds. I only exit when my trailing stop loss is hit or when the "artistic side of me" tells me that I am about to be hit.

However any shares which give me a buying signal certainly go on my watchlist

Eric

Time frame: as long as it takes

6_vs_7

Member

Post Number: 4
Registered: 08-2004

Rating: N/A
Votes: 0


Monday, December 18, 2006 - 07:29 pm:

The exit is extremely important. Get a good backtesting system, u obviously don't have a good 1 yet. Backtest your ideas 3-19 and see which works best.

The answer is most likely to be found here - 19. Develop an exit system based around trailing stops.

chart_rider

Member

Post Number: 201
Registered: 01-2005

Rating: N/A
Votes: 0


Tuesday, December 19, 2006 - 07:36 pm:

kool

I have to agree with all the above responses. A proper monte carlo back tester will provide realistic statistics regarding a mechanical strategy.

Not only does this provide a means of gaining confidence in a trading system, but also a target to work towards. With the application of discipline, it is indeed possible to match back test results.

CR

ittrader

Member

Post Number: 56
Registered: 01-2003

Rating: N/A
Votes: 0


Saturday, January 20, 2007 - 09:54 pm:

It is indeed possible that a backtest can (almost) replicate live trading results.

I use a fully automated system, and for the last year it backtests withing 2% of actual results.

kruupy

Member

Post Number: 38
Registered: 07-2005

Rating: N/A
Votes: 0


Sunday, January 21, 2007 - 02:48 am:

Hi ohkoolnutz,

Im not sure if I am allowed to write this in this forum but metastock has a system tester based upon historic data that actually does simulate a bankrole so that backtesting results are made a lot more reliable.

From Kruupy.

(Message edited by kruupy on January 21, 2007)

ohkoolnutz

Member

Post Number: 670
Registered: 10-2005

Rating: N/A
Votes: 0


Friday, October 12, 2007 - 08:43 am:

Nothing has changed...

Here I have this great trading system. It's profitable. I was 100% in stocks, all stocks were breakeven or in profit. I nailed the entries to perfection.

I know that long-term holding is not useful unless you want index-matching returns.

I had to sell something.
I didn't sell the break-even stock because I feel it will make its move soon.
I didn't sell the profitable stock that stopped moving up because I consider a great long-term investment with potential to double within a year. My mind is playing tricks on me. The long-termer is saying keep holding and the short-termer says sell and look for a new play.
I did sell the one that was profitable and was still moving up because I feared losing some of those profits (20% in a week). After I sold, it went up 20% the same day. I sold the one position I should not have sold.

I put the money into two others stocks. I feel the pressure again to sell something but I don't know what. So many opportunities so little money.

Whenever I feel like holding long-term the stocks move nowhere, whenever I feel like holding short-term they run away from me after I sell.

One thing more powerful than compound interest is compound frustration.

---
ohk

Lies, Damn Lies and Technical Analysis

ohkoolnutz

Member

Post Number: 671
Registered: 10-2005

Rating: N/A
Votes: 0


Friday, October 12, 2007 - 08:52 am:

and you say: "You don't have a complete system unless you have a sell signal besides the original stop loss."

I don't see them. I see buy signals. I see stocks that keep going up. I see stocks that wave-up. I see stocks that sine level. You don't know what your stock will do. How can your system sell signal provide for three varying scenarios if they are so substantially different?

and you say: "If you are unable to quantify the sell signal then be happy with what you get and move on. Maybe time and lack of cash is sell signal enough."

---
ohk

Lies, Damn Lies and Technical Analysis

sittingduck

Member

Post Number: 22
Registered: 11-2004

Rating: N/A
Votes: 0


Sunday, October 14, 2007 - 03:39 pm:

Kool,

Sounds like you are lacking an objective. You also seem to have a psychological issue with closing positions (do you cut you losers quickly or do you hold onto them in the hope they will come good)?

Think about what you are trying to achieve, and how you are comfortable achieving them. It would seem that you are very comfortable with a buy and hold strategy, which would suggest you are looking for long term capital gains, and not short term cash flow.

Here's some questions for you to ask yourself:
- Do shares you have bought show any trend in terms of acceleration (do they grow significantly faster just after you buy in comparison to say 6 months later). If this is the case you may be more profitable selling some of the older shares and reinvesting that in a faster moving share.
- Have you tested different money management strategies.
- What happens if you sell enough shares to cover your entry price when you hit say 50% profit, and then reinvest it. (eg buy$10,000, when they hit $15,000 sell $10,000 worth and buy $10,000 of something else), Use your own %'s based on what your testing shows.
- Have you considered running a second strategy? You seem to be saying you have a successful buy and hold strategy, how about trying a short term strategy with a small proportion of your current funds to see if you can make some pocket money for yourself. This would keep you interested while the long termers plod along, and give you some payback by paying for a nice meal/movie/whatever.

My experience has shown me that I need more than one strategy, that I can follow. In fact, I think I need 3. Short term (cash flow generator), Mid Term (Capital Growth), Long Term (Investment/Super/Dividend yields etc). So far I have Mid Term sorted, and am working on Short Term now. Super is locked in employers fund so apart from switching strategies (100% cash / 100% shares etc) not much I can do there yet.

Cheers
SittingDuck

hibikijoji

Member

Post Number: 366
Registered: 08-2005

Rating: N/A
Votes: 0


Thursday, November 01, 2007 - 07:26 am:

ohkoolnutz,

19. Develop an exit system based around trailing stops.

I agree with that as well.

If you know what you can secure in terms of profits, you can trade on margin lending. In that sense, borrow what you have have as secured profits.

Disclaimer :
The views expressed here are for experimental practices only. They have not been independently verified. No representation or warranty is made as to the accuracy, completeness or reliability of the information.

Please consult your broker or financial adviser before making an investment decision.