When To Start Your System?
When To Start Your System?
I'm hoping to get some food for thought here as I have a dilemma. I've done a lot of reading and research into Trend Trading and have found a strategy that seems satisfying to me. The one thing that the system can't tell me is when to get started? Here's the dilemma:
My system is a long only stock/ETF system. Most systematic traders are aware that the long term success of a system is directly correlated to the system's success out of the gate (I wonder if the Turtle's would have had a story to tell if they had immediate drawdowns past their limits, but I digress). True, I've saved money by not getting in at the recent market highs, but if I entered now, I'd still be "inheriting" signals that were generated before I started. So, I guess I'm just wondering how others have tackled this seemingly common dilemma. Here are my 3 options as I see them. If others have suggestions, I'd love to hear them:
- Jump into the system as is, using all "inherited" trades as if I was trading the system all along. (That is, after all, how I tested it and I know the saying, trade as you test...)
- Jump into the system and take all "inherited" trades with 50% of size. Add new signals 100% as they come.
- Wait to see if the market pulls back further to theoretically save more money. Yes, greedy I know since I've already saved by not jumping in two months ago, but since I'm not even close to the "average drawdown length", not sure I'm in a huge hurry.
I read an interesting article about this topic. Here's the link: http://www.automated-trading-system.com ... ew-system/
The article is not definitive, but leans toward the first option. I guess I'm curious to know what others have experienced. Please feel free to comment.
Thanks in advance!
JP
My system is a long only stock/ETF system. Most systematic traders are aware that the long term success of a system is directly correlated to the system's success out of the gate (I wonder if the Turtle's would have had a story to tell if they had immediate drawdowns past their limits, but I digress). True, I've saved money by not getting in at the recent market highs, but if I entered now, I'd still be "inheriting" signals that were generated before I started. So, I guess I'm just wondering how others have tackled this seemingly common dilemma. Here are my 3 options as I see them. If others have suggestions, I'd love to hear them:
- Jump into the system as is, using all "inherited" trades as if I was trading the system all along. (That is, after all, how I tested it and I know the saying, trade as you test...)
- Jump into the system and take all "inherited" trades with 50% of size. Add new signals 100% as they come.
- Wait to see if the market pulls back further to theoretically save more money. Yes, greedy I know since I've already saved by not jumping in two months ago, but since I'm not even close to the "average drawdown length", not sure I'm in a huge hurry.
I read an interesting article about this topic. Here's the link: http://www.automated-trading-system.com ... ew-system/
The article is not definitive, but leans toward the first option. I guess I'm curious to know what others have experienced. Please feel free to comment.
Thanks in advance!
JP
-
- Senior Member
- Posts: 28
- Joined: Tue Dec 30, 2008 5:36 pm
I think Schwager once did a study that showed something along the lines of investors jump out of CTA accounts when there are large drawdowns. The result was that they experienced subpar performance - after the drawdowns the system recovered. This would suggest starting after a drawdown to experience the reverse effect. If you are truly in no hurry, then waiting until you hit at least your average drawdown might not be a bad idea. Once you decide to trade, my vote would be to initiate trades to match the new and old positions. Your analysis is probably linked to the equity curve and you won't be jumping on the equity curve you have studied unless you initiate to match the open positions as well. In fact, it is possible that the effect of the reversal after a drawdown arises primarily from the open positions (I have no idea if that is true or false, just a possibility) and thus there would be no reason to wait for a drawdown if you were not planning to match the equity curve.
Just my two cents. Good luck.
Just my two cents. Good luck.
Why not test it and see... You could step the start date and on that date take all positions, or from that date, only take new entries. There are pros and cons to either path. Look out one week, month, quarter, x years... See what has happened in those time frames individually and as a group. Excel will be a good place to sort out the results.
-
- Roundtable Knight
- Posts: 427
- Joined: Mon Jun 01, 2009 4:12 am
- Location: once again in the UK
even with testing a lot will depend on your personal mental ability to handle the results if the timing is good or bad for starting the system.
A good system started at a bad time may cause you to discard the system as you cannot handle the initial drawdown.....if started at the right time for a bad system it might give you too much optimism.
Another conservative approach maybe to start a system and implement it 100% utilizing only a third of the money at a time over a period of time....the results maybe different to your 2nd option. ...food for thought.
A good system started at a bad time may cause you to discard the system as you cannot handle the initial drawdown.....if started at the right time for a bad system it might give you too much optimism.
Another conservative approach maybe to start a system and implement it 100% utilizing only a third of the money at a time over a period of time....the results maybe different to your 2nd option. ...food for thought.
Thanks for the responses:
@ - TechTrader - I haven't read the Schwager study but would concur that most people buy high, sell low and consequently underperform the market. It's the basic herd mentality, for lack of a better way to put it. In an ideal world (or should I say MY ideal world, I don't wish for others to lose money but this is from my perpective, after all), I'd get stopped out of all "inherited" positions on paper before I go live. Already been kicked out of a few. But still thing the general market has a ways to pull back. That said, once I start, I'm committed to using the system as designed. So I'm thinking I'm going to use some variation of average drawdown/average length of drawdown as my timing mechanism...
@Redrock - I have stepped the start date during my testing and the system proves to be robust enought to handle initial drawdowns. But certainly it performs better over time if we avoid the initial pitfall. So the dilemma is that the system works, it just works better if it gets a head start, so to speak. If I can get it, why wouldn't I take it?
@Moto - I'm thinking we're on the same page, just stating things differently. But as Techtrade and the referred to article suggest, it seems that all or nothing is more true to the testing I've done.
I'm thinking of digging deeper and seeing if there is a correlation to some technical indicators that may also give me a heads up as to when to jump in. Since the portfolio uses a large protion of Nasdaq 100 stocks, maybe look at the qqqs to see if a technical indicator on the general market may tip me off...
Thanks all for your responses. This is a very generous community. I know there is no one right answer, just wondering if I'm missing an angle...
JP
@ - TechTrader - I haven't read the Schwager study but would concur that most people buy high, sell low and consequently underperform the market. It's the basic herd mentality, for lack of a better way to put it. In an ideal world (or should I say MY ideal world, I don't wish for others to lose money but this is from my perpective, after all), I'd get stopped out of all "inherited" positions on paper before I go live. Already been kicked out of a few. But still thing the general market has a ways to pull back. That said, once I start, I'm committed to using the system as designed. So I'm thinking I'm going to use some variation of average drawdown/average length of drawdown as my timing mechanism...
@Redrock - I have stepped the start date during my testing and the system proves to be robust enought to handle initial drawdowns. But certainly it performs better over time if we avoid the initial pitfall. So the dilemma is that the system works, it just works better if it gets a head start, so to speak. If I can get it, why wouldn't I take it?
@Moto - I'm thinking we're on the same page, just stating things differently. But as Techtrade and the referred to article suggest, it seems that all or nothing is more true to the testing I've done.
I'm thinking of digging deeper and seeing if there is a correlation to some technical indicators that may also give me a heads up as to when to jump in. Since the portfolio uses a large protion of Nasdaq 100 stocks, maybe look at the qqqs to see if a technical indicator on the general market may tip me off...
Thanks all for your responses. This is a very generous community. I know there is no one right answer, just wondering if I'm missing an angle...
JP
-
- Roundtable Knight
- Posts: 138
- Joined: Wed Nov 10, 2004 4:36 pm
@babel... - Your point is well taken, but as I said, just looking for a good starting point. Not looking to filter every trade. I'm satisfied with the system as is, just want to know when to jump in. Perhaps there is a way to program this one time event, but I'm not saavy (sp?) enough to know how to program a filter for just a starting point, nor am I sure it would be worthy of a lot of time. I'm thinking of studying the characteristics of the general market as it changes from drawdown to profit to see if there is a common indicator for some guidance.
JP
JP
What Schwager said, if I remember correctly, was that CTAs often show large draw-downs (don't we know
) and that they usually perform great after such draw-downs. However, he didn't come up with a solution as how to time buy decisions based on this knowledge (which obviuosly is the key, and slightly more difficult
). What I think he meant was that when you invest in CTAs it's important to sit tight, and that what eventually ends up as gain comes with periods of severe pain.
To be honest, I don't think it's possible to test which is the best approach, since historic (and future) draw-downs just describe one of many possible paths a specific strategy could have taken. Having said that, if you believe that your system has positive expectancy, you should go all-in day one. Think about it, that's your best way of maximizing your outcome. From a personal or business perspective it may be wiser to scale (from day one) into your positions, however.
Well, that's what I think anyway...
![Smile :)](./images/smilies/icon_smile.gif)
![Wink :wink:](./images/smilies/icon_wink.gif)
To be honest, I don't think it's possible to test which is the best approach, since historic (and future) draw-downs just describe one of many possible paths a specific strategy could have taken. Having said that, if you believe that your system has positive expectancy, you should go all-in day one. Think about it, that's your best way of maximizing your outcome. From a personal or business perspective it may be wiser to scale (from day one) into your positions, however.
Well, that's what I think anyway...
I just launched a new basket of systems last week
Straight down 10 %
I launched the main fund last year and it all went in the other direction - straight up
Its luck in the short term
Do you feel lucky ?
I would recommend getting in and forgetting about it ? I would have thought most systems in a failry decent d/d right now and since most TF's flat on the year (on average) - this is about as good as it gets to launch
C
Straight down 10 %
I launched the main fund last year and it all went in the other direction - straight up
Its luck in the short term
Do you feel lucky ?
I would recommend getting in and forgetting about it ? I would have thought most systems in a failry decent d/d right now and since most TF's flat on the year (on average) - this is about as good as it gets to launch
C
How about your start trade drawdown? Was that better too? Which gave you the least chance of going broke right from the get-go, while trading real Benjamin's, fresh from your savings ?rgd wrote:I did a bit of research on this and found that putting on all the existing positions on the launch date vs. taking only new positions outperformed more than 70% of the time.
Scenario #1: you start trading immediately. If the system(s) go into a deep drawdown, and you lose large amounts of real money, then you feel Remorse#1. On the other hand if the system(s) go into a huge profitable run-up, you make large and real profits, then you feel Delight#1.
Scenario #2: you decide NOT to start trading for a while. If the system(s) go into a deep drawdown and you avoid the loss, then you feel Delight#2. On the other hand if the system(s) go into a huge profitable run-up, you miss out on the huge profit and so you feel Remorse#2.
If you choose either Scenario#1 or Scenario#2, you might get a big steaming pile of Remorse. (You might get Delight too).
What if you choose half of each? What if you invest half your money according to Scenario#1 and invest the other half of your money according to Scenario#2? Then you'll get half-Remorse and half-Delight. No matter what happens, you're guaranteed to be half Delighted. Maybe that's preferable.
What does it mean to "choose half of each?" With half your money, put on all positions right away. With the other half of your money, put on no positions at all. In other words, put on all your positions at half-size. But don't do this unless YOU like the idea.
Further reading: http://goo.gl/UBVDn
Scenario #2: you decide NOT to start trading for a while. If the system(s) go into a deep drawdown and you avoid the loss, then you feel Delight#2. On the other hand if the system(s) go into a huge profitable run-up, you miss out on the huge profit and so you feel Remorse#2.
If you choose either Scenario#1 or Scenario#2, you might get a big steaming pile of Remorse. (You might get Delight too).
What if you choose half of each? What if you invest half your money according to Scenario#1 and invest the other half of your money according to Scenario#2? Then you'll get half-Remorse and half-Delight. No matter what happens, you're guaranteed to be half Delighted. Maybe that's preferable.
What does it mean to "choose half of each?" With half your money, put on all positions right away. With the other half of your money, put on no positions at all. In other words, put on all your positions at half-size. But don't do this unless YOU like the idea.
Further reading: http://goo.gl/UBVDn
I should state that my research only looked at 50-60 data points, and evaulated the results on a rolling 1 year basis. This obviously overlooks those times that the theoretical launch immediately went into drawdown, yet managed to maintain the positions and was better off after 12 months.
Sluggo makes the most important point of all, it is a slippery slope to begin making decisions along these lines. What if you open a big, new account 6 months after launch? What do you do with them?
Sluggo makes the most important point of all, it is a slippery slope to begin making decisions along these lines. What if you open a big, new account 6 months after launch? What do you do with them?
I use the 1/2 'sage advice' tool as an option routinely now when faced with for example, a synchronization error, or unusual price/market scenario. Having that option is a relief from the quandaries traders sometimes face. Thank you Sir Sluggo, for bringing the tool to light! While it seems obvious, the obvious is oft elusive in critical decision point scenarios.
-
- Roundtable Knight
- Posts: 1436
- Joined: Mon Dec 22, 2003 12:24 pm
- Location: Des Moines, IA
- Contact:
I dont like this procedure for the fact that position sizes for day 1 of the test will not typically match theoretical position sizes for earlier entries, especially if you employ profit taking.RedRock wrote:Why not test it and see... You could step the start date and on that date take all positions, or from that date, only take new entries. There are pros and cons to either path. Look out one week, month, quarter, x years... See what has happened in those time frames individually and as a group. Excel will be a good place to sort out the results.
LeviF wrote:I dont like this procedure for the fact that position sizes for day 1 of the test will not typically match theoretical position sizes for earlier entries, especially if you employ profit taking.RedRock wrote:Why not test it and see... You could step the start date and on that date take all positions, or from that date, only take new entries. There are pros and cons to either path. Look out one week, month, quarter, x years... See what has happened in those time frames individually and as a group. Excel will be a good place to sort out the results.
When I do this, I'm asking what has been the outlook for the first week. month, 6 mo. etc. If I had by chance started trading the system with my initial investment over various start up philosophies. ( all in, half, only new positions etc) I'm most interested in what I may loose, vs what I may gain. Using the presumption that all existing positions, if taken, would be sized basis my starting investment or filtered by position size exceeding my risk tolerance. STDD Start trade drawdown is one thing I look at, though past is not future. Helps to understand the risks of a new campaign. Ever imagine the poor soul who starts a new program just before a new max dd? Or the one who waited patiently for a dd of say half the ''expected'' levels, only to have the dd continue to double the tested max. Having that info is just one more piece in the decision process I've found useful. Lest we forget, there is research, but also luck...
I started my system on July 28, 2010 and looked like a hero for a number of months. I happened to launch at a good time. That was luck, not skill.
Another option which hasn't been noted would be to invest the funds in equal amounts at given time intervals. For example, invest one third in open positions now, one third a month from now and the final third in two month's from now.
Fred
Another option which hasn't been noted would be to invest the funds in equal amounts at given time intervals. For example, invest one third in open positions now, one third a month from now and the final third in two month's from now.
Fred
MY new share class is now down 15+ - I feel large chunks of remorse
howver I only started it - to be honest on 80% of the monies - also knowing that the size of teh fund doubles in October anyway - so i feel remorse and at the same tome sort of happy - its the sluggo approach in a more aggressive way - O dont believe there are any easy answers to this conundrum - suffice to say that when you backtest your system you should have certainly tested out the fact that whenever you started your system never blew up - it may well have gone into a steep draw down but it should certainly never have blown up - this is a warning sign your system may not be robust ?
howver I only started it - to be honest on 80% of the monies - also knowing that the size of teh fund doubles in October anyway - so i feel remorse and at the same tome sort of happy - its the sluggo approach in a more aggressive way - O dont believe there are any easy answers to this conundrum - suffice to say that when you backtest your system you should have certainly tested out the fact that whenever you started your system never blew up - it may well have gone into a steep draw down but it should certainly never have blown up - this is a warning sign your system may not be robust ?