I trade a multi market, multi system and multiple timeframe portfolio. I trade the ES, EC, TF, and CL contracts. My question is regarding drawdown. Would you monitor the performance of each system separately and stop trading at a predefined drawdown limit? In my mind that's the reason I trade this portfolio in that when DD's occur another part of the system is carrying the performance of the portfolio. Curious as to others thoughts on handling DD's.
The following user says Thank You to dgresens for this post:
You should have a lot of forwarded tested trades before going cash, so that you can know the typical drawdown limits and other percentages of your methodology so you can make sure new trades continue to fit that model.
If you are trading multiple instruments concurrently, is it because you are purposely trading these based on a correlative style (or non-correlation), and some type of hedging? If so, I would view the portfolio as a whole. If you are trading each instrument on its own merits, and that trade on that instrument really has no bearing on another trade on another instrument, then you should measure them independently.
Due to time constraints, please do not PM me if your question can be resolved or answered on the forum.
Need help? 1) Stop changing things. No new indicators, charts, or methods. Be consistent with what is in front of you first. 2) Start a journal and post to it daily with the trades you made to show your strengths and weaknesses. 3) Set goals for yourself to reach daily. Make them about how you trade, not how much money you make. 4) Accept responsibility for your actions. Stop looking elsewhere to explain away poor performance. 5) Where to start as a trader? Watch this webinar and read this thread for hundreds of questions and answers. 6) Help using the forum? Watch this video to learn general tips on using the site.
If you want to support our community, become an Elite Member.
Broker/Data: Advantage, Trading Technologies, OptionsCity, IQ Feed
Favorite Futures: CL, NG
Posts: 1,040 since Jul 2010
Thanks: 1,713 given,
For swing trades, I trade these markets as well in addition to a few others. Drawdowns will always occur however, I look to be in markets that have low correlation to one another with a predefined asset allocation percentage per market. I run a core/satellite strategy (60/40) with my core positions usually being ES (long or short depending on the market's trend) with smaller satellite positions. There of course, are days when everything goes against you and you have a few stops hit. But I also offset these drawdowns with my intraday trades of CL, TF and 6E which will "hedge" my swing trades. As far as having a draw down limit per market, I guess that would simply be having my stops taken out but like I said, I intraday trade to offset some of the losses.
Your question is a bit vague in my opinion. When you say "how do you handle drawdown?", do you mean:
1. How do I know my system is broken?
2. How do I get best possible outcomes?
Either way, i'll answer the question to the best of my knowledge.
1. How do you know your system is broken?
There are practically unlimited ways of doing this. I only outline the most popular ones:
* X% Stop (e.g. 25% stop)
This is self-explanatory.
* R-drawdown stop
Basically, you calculate the how many R drawdown you have instead of % stop.
* Statistical test (z-score or chi-squared test)
Basically, the problem you are trying to solve is:
"Is the current R distribution statistically significant compared to the back-tested ones?"
If statistically different then stop.
The common way to do this is:
1. you take N samples from historical testing
2. While you trade live, keep your ongoing samples
3. Periodically perform statistical test
* Statistical test for synchronization - (Variant of above)
Just like above, the different is if it's statistically different, then don't live trade, but still keep paper trading. When the N sample back to the "synchronization" with historical sample, you can continue trading.
* Equity filter
Basically, you keep paper trading and live trading together.
1. If equity going up keep live trading.
2. If equity down by X% then stop live trading. When paper trading equity back up, then continue live trading.
There are many many many many ways to do this. It all returns to "your belief" and "your objectives".
Keep in mind that testing (backtest or forward-test) only raise your confident. Forward test is more useful as it shows you whether your system is overfit or not. However, both approaches don't give you any warranty about what will happens the next day. Therefore, you shouldn't rely on system only, you should use proper money management as another safety belt. When worst-case happened, money management will be there to save you. I know many people over-trade. Don't become like any of them. Remember, "While the beginner look at profit, the expert look at the risk". Trading decision is all about common sense. It's not funny if you get $100k / month for 6 months and lose $1 million in a day.
In the case of multi systems, it makes more sense to monitor them as individual. You don't want the entire portfolio bogged down by single sub-system.
2. How do you get best possible outcomes?
A vague question, but the thing is "best" here is relative. Here's some people definition of best trading system:
1. Has very low drawdown.
2. Has best reward/risk ratio.
3. Has best winning percentage.
Firstly, the main goal of trading system diversification is to "INCREASE REWARD/RISK RATIO" (not drawdown elimination).
Additionally goal 1 can be achieved through scaling down your position such that your drawdown is at tolerable level.
Several empirical studies regarding this issue suggest that reward / risk ratio increase when each sub-system has low correlation (almost 0).
Keep in mind, low means 0 (read - ZERO!) not positive not negative.
Positive correlation means you multiply your position, negative means you divide your position. In most cases, you pay more transaction cost.
To ensure your diversification increase reward/risk ratio, do this:
1. Ensure all system has low correlation.
2. Calculate the reward/risk ratio individually (as a system).
3. Treat R distribution for all system as one big system and calculate reward/risk ratio.
4. Make sure the big system has higher reward/risk ratio. In most case, it will though.
It's true the diversification will reduce your drawdown. But in my opinion that shouldn't be your goal.
If that's your goal, you don't need diversification, just trade half position.
What I mean is, Increase reward/risk ratio means higher profit and "lower drawdown". But having this only might not utilise the diversification to its full potential. In order to do so, you have to scale your position back up such that the new drawdown is equal to old drawdown (which is tolerable to you). In exchange, you got higher profit. The key is to get the drawdown just right, not too conservative, not too aggressive.
Also, if your combined system doesn't have significant increase of reward/risk ratio, there's no reason to do that. Because more system means more transaction cost (slippage, spread, commision). This transaction cost might be small for long-term system. But it's... abysmal for short-term system (e.g. day trading system). The question is whether that "reward/risk ratio" increase worths extra transaction cost or not.
I hope that help.
The following user says Thank You to felixtjung for this post:
I don't really know whether there's anything useful in assessing risk/reward ratio for single trade. To me, that kind of approach is highly misleading (prove me wrong though, perhaps you could teach me one or two things about this kind of assessment).
Back to reward/risk assessment. This assessment is typically done over period of time (typically one year). They way you do this is:
1. Convert each trade as R distribution.
2. Calculate net profit in term of R.
3. Calculate the maximum drawdown.
4. Reward / Risk ratio = (Total R / Max Drawdown R) over N period.
I'm not claiming this is my calculation method as it's not I who invent it. In fact, this calculation is a popular one when assessing the risk/reward ratio and also used often in many many trading literature.
About transaction cost.
Transaction cost = Slippage + Commision + Spread.
If you already factor Commision + Spread into back testing and the system still profitable, then it's more likely it's profitable. But, I'm under impression that you are accepting high commision + spread as one part of the "trading nature" or perhaps do some brokers convince you to trade that way. Whichever it's, it's your problem not mine.
I'm not trying to offend you in anyway, instead I encourage you to see it this way:
If you are trading inside the "neutral noise". I got my own trading strategy "law" applicable, I call this law "The law of oscillator". The law can be described as:
(Note: I'm not aware if there's somebody out there that already coined this law. If you know similar trading law, let me know the real name. or link please. thanks a heap!)
This law comes from my own observation of the market in the "neutral noise" (this is defined "loosely" as the timeframe where single swing up and single swing down market is about same) and proven by one of my experiments. Perfect neutral noise only happens under very rare circumstances such as tick timeframe on sideway market or an extremely volatile market. As a rule of thumb, the longer timeframe, the less effect this law has on the trading strategy. I don't expect you to take this law for granted and believe it, instead I encourage you to try it yourselves see if it's proven or not for you.
However, all markets (trending or not) still affected "a little" by this law. To see this effect, whatever your strategy is, increase your stop loss threshold, and you'll see more winning percentage. This law is by no means perfect, but it can be leveraged to our advantages in some circumstances as a rough guideline.
(Assuming you believe about my law) The transaction cost takes out your profit threshold. Therefore, bigger your transaction cost means the odds is against you.
Therefore, don't be misled by those guy who claim their system 90% or 95%. Let me tell you, I got about 10 of that kind of systems (I'm serious, not bragging!). I do it for fun. Do I trade them? Hell no way! If you want to create that system, just put big stop loss or even better don't put stop at all + trade with trends + filter out all of the major loses using filter (from backtesting). I hope you won't sell the system should you create such system. Otherwise, I'll be send to hell for telling you this.
In the case of short-term trading, transaction cost is taken into extreme, the odds which against you are also taken into extreme. However, that doesn't mean you can't be successful. Despite little profit, short-term traders are blessed with the highly frequent trades (as known as their biggest gift and also their biggest curse).
Don't get me wrong, I'm used to be day trader too (peace!). I love it, but as there's time issue there, I've decided to back off from that type of trading.
This is a little bit controversial post I just made here! I expect a lot of flame!