| Index | Recent Threads | Who's Online | User List | Search |
|
|
 |
Portfolio123 » List all forums » Forum: Simulations and Portfolios » Thread: Denny's Stop Loss Challenge |
|
Total posts in this thread: 66
|
[Request a Feature]
[Post new Thread] [Add To My Favorites] [Watch this Thread] |
| Author |
|
|
DennyHalwes
Advanced Member UNITED STATES Joined: Apr 28, 2004 Posts: 1532 Status: Offline |
In an earlier thread on Sam’s excellent 2 years results here , we got away from the subject with a discussion of stops. I would like to move the discussion on stops to a new thread. There were a number of earlier good discussions on stops: here : And here : And here : And here : And about a dozen more. In one thread I said that I was almost always able to improve a Sim with stops. And in Sam’s thread I asked “Question, if a stop loss will improve the annual return and the max drawdown of your Ports would you still not use it?” So the real question is: Are there good Sims that do not use stops in the sell rules that can’t be improved with stops? I don’t want to address very short average hold time Sims of 1 week or less. I already know that I can’t improve many of them with stops, but what about the other Sims that most P123 members trade? So, my challenge is to see if you can post a good Sim that can’t be improved with stops. The rules are: The Sim and Ranking System must be public, and trade 5 or more stocks. The Sim must be set up to trade a minimum or greater liquidity of: AvgDailyTot(20) > 100000, and Close(0) > 1 The Sim must be run over at least 5 years of data. This is an opportunity for you to let myself and other members see if we can improve your Sims. Denny  ---------------------------------------- "The significant problems we face cannot be solved at the same level of thinking that we were at when we created them". Albert Einstein ---------------------------------------- [Edit 1 times, last edit by DennyHalwes at Jul 6, 2007 12:24:39 AM] |
||
|
|
olikea
Advanced Member
|
The objection I have about stops is that they are entirely arbitrary. It is having a sell rule for a (say) 10% decline, but from what - your entry point. From this point of view, PctFromHi makes more sense, but if you have such a rule of say PctFromHi < -25, it follows that you should only buy a stock that is within 25% of its all time high, and if not why not? Also, I think that you end up market timing, if the market goes through a 10% correction then you will be stopped out of all your positions, but what is your signal for getting back in? People talk about stop losses limiting your risk, but they really don't, not unless once you are stopped out you just go off and live in the Carribean never to invest again. There is plenty of evidence floating around that short term weakness should be bought, ironic that a signal that triggers a stop loss, actually triggers a buy signal for systems such as those developed in Olikea's Challenge. However! That doesn't mean I don't believe in cutting your losses, but only when the trend has changed. Try adding: close(0) < sma(200) to your sell rules, its amazing how well it works. In my view this rule is vastly superior to a simple stop loss, it seeks a long term change of trend and it doesn't depend on your (entirely arbitrary) entry point. Additionally, it provides a clear barrier between short term weakness (which should be bought) and something more serious. I feel that stop losses are very blunt tools, maybe ok for discretionary traders to help enforce discipline, but p123 is such a refined tool that surely we can come up with something more sophisticated? |
||
|
|
dwpeters
Advanced Member
|
Hi, Ok, I modified one of my sims to meet your criteria. When I first tested a variation in February, stops hurt the returns without improving the drawdown, and that seems to be the case now as well. In fact in my limited testing I've found that stops generally seem to hurt returns, but I don't doubt that you can find a way to improve the returns with stops. http://www.portfolio123.com/port_summary.jsp?portid=286404 Don |
||
|
|
DennyHalwes
Advanced Member UNITED STATES Joined: Apr 28, 2004 Posts: 1532 Status: Offline |
Olikea, That is just the thinking that I am interested in challenging! Is a Rank < 95 sell rule entirely arbitrary? Or did you test it to see what value results in the best return. The same thing applies to stops. You asked, “PctFromHi makes more sense, but if you have such a rule of say PctFromHi < -25, it follows that you should only buy a stock that is within 25% of its all time high, and if not why not?” The reason why not is because when you test it the return sucks. If it doesn’t work when you test it, why would you ever follow it? You asked, “Also, I think that you end up market timing, if the market goes through a 10% correction then you will be stopped out of all your positions, but what is your signal for getting back in?” The Sim buy rules get you back in. you don’t need a signal. You said, “There is plenty of evidence floating around that short term weakness should be bought, ironic that a signal that triggers a stop loss, actually triggers a buy signal for systems such as those developed in Olikea's Challenge.” That is true, but the key is weakness. -2% is weakness. -5% is weakness. -25% is a big loss. That is why PctFromHi < -5 doesn’t improve Sims, but frequently PctFromHi < -25 does. In the Sim of your first post in Olikea’s Challenge” you used the sell rule, gainpct < -5. Isn’t that a stop loss? Why do you choose any other buy or sell rule? Besides liquidity rules, isn’t it because they improve the performance, both annual return and drawdown? For that purpose, what is the real difference between a stop loss rule and any other rule? If it improves the performance why would you reject it? If the P123 tools can show that, on the average trade, stop losses will work why not use them? I have found in my studies of Sims with and without stop losses that the performance improvement has much more to do with the return of the replacement stock than what the stock that was stopped out would have done if it had not been sold. Remember, the Sims pick winning stocks 60% to 70% of the time. So you have a 60% to 70% chance that the replacement stock will be a winner. The stopped out stock has already proven its initial performance. If we can improve both the annual return and the max drawdown of one of your Sims that does not have a stop loss, will you trade it? If so, post it here and we (that belive in stops) will see what we can do to improve it. If not, leave the thread for someone else that is interested in improving their Sims. Don, You probably have a good challenge for us, but you Sim is private! Denny ---------------------------------------- "The significant problems we face cannot be solved at the same level of thinking that we were at when we created them". Albert Einstein ---------------------------------------- [Edit 1 times, last edit by DennyHalwes at Jul 6, 2007 12:28:39 AM] |
||
|
|
probtrader
Advanced Member ITALY Joined: Oct 11, 2004 Posts: 109 Status: Offline |
I second your thought. During my research I also found out that stops, both to the downside (stop loss) and upside (profit taking), work better for higher volatility screens. That sounds logical to me: the higher the volatility, the higher the chance to hit stops. In this case stocks will float within the range defined by the downside and upside stop. If the screen has better than 50% winners, then most stocks will hit the profit taking area. Leaving the stocks free to float until rebalance time will produce more random results. One more thing: why would you make a difference between a price-based stop and selling the stock when it doesn't meet the screen requirements? Isn't the latter just another form of stop, based on fundamental data? ---------------------------------------- [Edit 2 times, last edit by probtrader at Jul 6, 2007 3:28:36 AM] |
|||
|
|
Stittsville123
Advanced Member
|
"I have found in my studies of Sims with and without stop losses that the performance improvement has much more to do with the return of the replacement stock than what the stock that was stopped out would have done if it had not been sold. Remember, the Sims pick winning stocks 60% to 70% of the time. So you have a 60% to 70% chance that the replacement stock will be a winner. The stopped out stock has already proven its initial performance." Denny - Last October I produced several ports based on your rules for eliminating stocks that aren't performing. The concept is to replace them with new stocks .... Although the sims produced excellent results, I can't say the same thing for the ports. All of the ports are automatically rebalanced. Here are the results to date: http://www.portfolio123.com/port_summary.jsp?portid=192347 http://www.portfolio123.com/port_summary.jsp?portid=192319 http://www.portfolio123.com/port_summary.jsp?portid=192326 http://www.portfolio123.com/port_summary.jsp?portid=192221 Fifth port is based upon a proprietary ranking system. I have attached a pdf of the equity curve. The fourth port in this list is actually starting to show some promise after some initial volatility. The problem I have with these ports/sims is that they are another form of sell when down. (I am a firm believer in selling when a stock is up not down.) What I really get out of this exercise is that the ports will certainly be extremely volatile but they won't necessarily produce the expected superior returns. Now in all fairness I created these ports not Denny. A little while ago I asked Denny to lock down some ports that he was displaying which had exceptional performance. I requested to see the equity curves of those ports after a period of time (six months I believe). I am prepared to wait to see the results of those specific ports before determining the worth of the concept of dumping non-performing stocks. (Denny - you haven't deleted those ports I hope??) In my opinion this is a more valuable exercise than creating more sims. After all, it is the ports that are important, not the sims. Steve Steve ---------------------------------------- 
Stitts 18 Factor.pdf
(46100 bytes)
(Download Count: 34)
|
||
|
|
bl82
Advanced Member UNITED STATES Joined: Apr 18, 2004 Posts: 97 Status: Offline |
Excellent challenge, Denny. Don't forget that the addition recent addition of some new TA functions should dramatically expand the universe of possible stops. For example: ATR -- for volatility-based stops SAR -- combines a time stop component with a price stop (a personal favorite) ADX -- can generate a stop when a stock stops trending Looking forward to seeing what comes out of this one. Cheers, -Bill ---------------------------------------- http://vixandmore.blogspot.com/ "Your one stop VIX-centric view of the universe..." |
||
|
|
DennyHalwes
Advanced Member UNITED STATES Joined: Apr 28, 2004 Posts: 1532 Status: Offline |
Steve, Since the buy and sell rules are almost the same in all 4 of your Sims, and the big difference is the change in the Ranking System, I will address the 1st one which has the worst performance. In your first Sim there were only 3 stocks that were sold due to the stop loss rule out of 30 stocks sold, or 10%. It is hard to make any conclusions from only 3 data points. However some info is available. MWRK was the first stock that was sold due to the stop loss on 12/04/06. It was replaced by GLIT on the same date. GLIT was sold 22 days later for a loss of 1.8%. MWRK was down 5.1% 22 days later. Chalk up 1 for stop losses. ROCM was the 2nd stock to be sold due to the stop loss on 04/30/07. However it was sold with a 38% gain after falling back from a 75.4% gain. ROCM was replaced by ISH on the same date. ISH was sold 21 days later on 05/21/07 with a 16.4% loss. However, ROCM would have lost an additional 31% after 21 days for a net loss of 4% instead of a gain of 38% if it had been kept. As of today ROCM is still below its price on 05/21/07. Chalk up a 2nd one for stop losses. POP was the 3rd stock sold due to the stop loss on 05/14/07 for a 37.2% loss. It was replaced by HDNG. HDNG was sold 28 days later for a 6.9% loss. Pop, if held for the next 28 days, it would be down an additional 12.9% and it is down even more today. Chalk up a 3rd one for stop losses. 3 out of 3 for stop losses. Out of the 30 stocks sold, even an improvement on 10% of the stocks is worth considering. And yes, I have kept all 6 Ports and I will be reporting back on their performance on 10/19/07. Denny ---------------------------------------- "The significant problems we face cannot be solved at the same level of thinking that we were at when we created them". Albert Einstein |
||
|
|
hyper
Advanced Member
|
The markets have been going very well for the past few years, and this can easily delude us into thinking that we are good traders. If you haven't been using stops, you may think that you don't need to because you have made good money and continue to do so. I'm quite happy myself with my performance in the past year. I've made over 100% return with 11% maximum drawdown. I try not to let the success inflate my ego and cloud my judgement. The real test will come when the market gets tough for a longer period than just a month or so, and/or when the next major crash comes. We don't know when these will come, but by using stops and disciplined money management, I will be ready. A lot of people made and then lost fortunes during the tech boom and then crash. I think a large proportion of them would have kept a lot of their gains if they used stops, and used them with discipline. I only lost a few thousand or so, luckily. Back then I did not even know what stops were, and my broker (Commsec, the most popular online broker in Australia then and still is) did not even offer stop functionality. I wish I knew back then what I know now. Did any of you actively trade during that period, and did you trade without stops? It is not possible to properly test stops in portfolio123. In reality you would have pending orders in the market and they will get executed as soon as the price reaches that level within the day. With portfolio123, the position is closed at the next open. This can be quite late and is not the way a technical trader would use stops. Portfolio123 needs to implement intra-day order execution at specified (pre-calculated) prices. Software like Metastock (with TradeSim plugin), Amibroker, Wealth-Lab, and Tradestation are much more sophisticated platforms for precise backtesting. Portfolio123 is excellent for fundamental testing, but has a long way to go to match the strength and precision of other technical backtesting platforms. |
||
|
|
hyper
Advanced Member
|
I just happened to be reading "Come into my trading room" by Alexander Elder before I came across this thread. Elder talks a lot about stops in the book. Here's a section about protective stops.
|
|||
|
|
[Show Thread Printable Version] [Post new Thread] |
Denny's Stop Loss Challenge
Re: Denny's Stop Loss Challenge
