|
7 registered members (fairtrader, Quad, miwok, Martin_HH, AndrewAMD, alibaba, dpn),
581
guests, and 0
spiders. |
|
Key:
Admin,
Global Mod,
Mod
|
|
|
Downside to capturing profits?
#429341
09/11/13 14:58
09/11/13 14:58
|
Joined: Jun 2013
Posts: 1,609
DdlV
OP
Serious User
|
OP
Serious User
Joined: Jun 2013
Posts: 1,609
|
Hi all. Of course there's no downside to capturing profits, is there?!  But I mean in Zorro strategy execution terms. Specifically, in my Z scripts testing there have been times when Zorro is up a good bit. However, this is not a real "up", since Zorro won't exit at those values - it will exit later at lower stops. So, the question is: Is there a downside to stopping Trading, closing all open positions, and then either resuming Trading by clicking No to exit, or clicking Yes and restarting Zorro? I would think not, but there may be yet another thing I don't yet understand... Thanks.
|
|
|
Re: Downside to capturing profits?
[Re: DdlV]
#429344
09/11/13 15:05
09/11/13 15:05
|
Joined: Nov 2012
Posts: 126
blaub4r
Member
|
Member
Joined: Nov 2012
Posts: 126
|
The question is: How do you know where the optimal exit point is? Zorro does not know that, too, but at least captures a big deal of the profit  By stopping and starting the strategy manually you usually worsen the perfomance since you are adding some random component to it.
|
|
|
Re: Downside to capturing profits?
[Re: DdlV]
#429364
09/11/13 19:14
09/11/13 19:14
|
acidburn
Unregistered
|
acidburn
Unregistered
|
In theory, if there's a better exit strategy, why not code it directly in the strategy? But, of course, we can't do that for Z strategies without the source code.
Basically, quants suggest to never tamper with the strategy. Either improve it, or just plainly stop trading it if you don't believe it.
Now, on the practical side, I feel your pain. I've also seen nice fat paper profits in Z strategies evaporate before the trade is closed. But, even though I've seen it many times, I just can't be sure it's really so common (maybe I just saw what I wanted to see?). Probably I could invest some time, analyzing trade by trade (myfxbook could be handy for that) and try to prove it. But, then again, what would it be good for, when I can't modify/optimize the strategy?! That's the inherent problem with black box strategies, you can't do much to improve it. Either you trust/trade it, or not.
In short: if you think adding discretionary exits benefits you, by all means do it. Just be aware that it's not really recommended.
|
|
|
Re: Downside to capturing profits?
[Re: ]
#429374
09/11/13 19:55
09/11/13 19:55
|
Joined: Jun 2013
Posts: 1,609
DdlV
OP
Serious User
|
OP
Serious User
Joined: Jun 2013
Posts: 1,609
|
Thanks acidburn. Yes, I know it's not really recommended, but I'm struggling to grapple with why? It isn't a question of not believing in a strategy - otherwise, as you say, you shouldn't trade it. And I don't think it's that common. It's more a question of taking advantage of a situation where the strategy performs better than expected, rather than letting it revert the gains to the expected result... Which would put $$$ in my pocket that otherwise wouldn't make it there - a good thing!  Thanks.
|
|
|
Re: Downside to capturing profits?
[Re: DdlV]
#429379
09/11/13 20:10
09/11/13 20:10
|
acidburn
Unregistered
|
acidburn
Unregistered
|
Of course. The question that remains is how accurate you would be? You certainly can't expect to win every time, so some times you would react properly, and some times you wouldn't (you would cut profits).
Now when you think more about it, in a way you would manually implement TakeProfit, would you? But then, if that TakeProfit is really more profitable it could be built right into the strategy, so that it is triggered automatically. If it was that simple, don't you think that jcl would built it right into the Z's?
The more I learn, more I understand that lots of ideas look really good on paper, but fail miserably when put to test. The only way you could really measure the effect of your tampering with the strategy is to forward test untampered strategy together with the one where you apply discretion. So after some months have passed, you could see which one performed better.
|
|
|
Re: Downside to capturing profits?
[Re: ]
#429380
09/11/13 20:18
09/11/13 20:18
|
Joined: Jun 2013
Posts: 1,609
DdlV
OP
Serious User
|
OP
Serious User
Joined: Jun 2013
Posts: 1,609
|
Thanks acidburn. I think I'm not being understood correctly. I wasn't thinking of mucking with a strategy. I was thinking more of Zorro's display showing account equity (when FXCM's fixed their bug and it works  ). If that's (significantly) more than what I was expecting from the strategy, why not click Stop, close all trades, make that paper equity real, and then restart Zorro. I.e., would doing this in some way "take away" from the subsequent strategy execution and guarantee that it'd be unprofitable? I don't see why; hence don't see the downside to taking advantage of a surprisingly good equity result... Thanks.
|
|
|
Re: Downside to capturing profits?
[Re: DdlV]
#429383
09/11/13 20:51
09/11/13 20:51
|
acidburn
Unregistered
|
acidburn
Unregistered
|
Thanks acidburn. I think I'm not being understood correctly. I wasn't thinking of mucking with a strategy. In my book, not allowing the strategy to close the trade but instead closing it manually IS "mucking with a strategy". You're certainly entitled to a different opinion.  I was thinking more of Zorro's display showing account equity (when FXCM's fixed their bug and it works  ). If that's (significantly) more than what I was expecting from the strategy, why not click Stop, close all trades, make that paper equity real, and then restart Zorro. I.e., would doing this in some way "take away" from the subsequent strategy execution and guarantee that it'd be unprofitable? I don't see why; hence don't see the downside to taking advantage of a surprisingly good equity result... Only jcl can answer this one. And it probably depends on a specific strategy (some wouldn't even allow you to close the trade, because they would just reopen it right away and basically neutralize your action and also add some cost due to added commission).
Last edited by acidburn; 09/11/13 20:52.
|
|
|
Re: Downside to capturing profits?
[Re: ]
#429403
09/12/13 03:36
09/12/13 03:36
|
Joined: Aug 2013
Posts: 124
DMB
Member
|
Member
Joined: Aug 2013
Posts: 124
|
Here is my 2 cents worth. I too have the human emotions of greed and regret. So I decided to look into this question a little bit. I took the trade history for a Z back test, split out the currencies and individual CFDs and looked at it statistically. The question I wanted to answer is, is there a benefit to manually getting out of individual trades that 'spike' to two standard deviations above the average? Will this bring up the overall average trade, or will cutting short some of the big winners result in a lower overall average trade? This second question, though possibly over simplifying, in IMO is the underlying question.
I don't have a definitive answer to this question, mainly because the results file does not report the maximum positive excursion. Also, I found that the distribution of winners is heavily skewed to the lower end, less than average region, and has a long tail. So calculating a basic average and standard deviation was not suitable for this data set. So I looked at the percentiles to help get a feel for the distribution. That is the extent of my statistical ability. It's been many years since I did statistics for engineers.
The numbers herein are numbers that I have calculated. This is by no means a recommendation, nor has it been peer reviewed for accuracy. I found for the Z12 currencies, the P80 winners (80% of the winners are below this value) was 142 pips. P95 is 283 pips. 15% further along the winners ranking doubles the profit. This indicates the presence of a long tail. Average winner is 67 pips. For reference, average loser is -30 pips.
I calculated other numbers around these values, and other statistical measures, but this is it in a nutshell. So how I am I using this information? Well firstly, I only pay attention to the pips, not the dollars. Sometimes Zorro has me in 12k, sometimes only 1k or 2k. So I operate on a normalised basis. Hence pips, not dollars.
Secondly, if I notice that a trade has moved beyond the P80, I consider manually taking profits. CONSIDER. I like to look at my charts and see what is going on, based on my former discretionary trading methods. Of the 10 or so trades that I have taken action to close, I would say maybe half continued a little, and the rest reversed or simply went into consolidation / sideways trade for a while.
For the rest of the trades, particularly the losers, I let Zorro manage it. I want to be sure that I don't make the loser average worse. I am already using a little too much emotion without meddling with the 60% of trades that are losers.
Taking the profits off the table after a market 'spike' helps me feel positive about trading. It helps me feel comfortable trading what is a black box from my perspective. These are the WRONG reasons to trade! But I will be the first to admit that my trader psychology is sub par. After a month, my account balance is up 36%, however that is not relevant since it is really based on the margin settings of the Zs and the proportion of Capital Required to my total account size. This idea is suitable for a system with a trailing stop, where there is always a positive excursion that is much better than the actual exit. Also the markets' current volatility is not taken into account. There will be times when 142 pips will be closer to the average than the P80. I suppose, if the ATR of the recent market is higher than normal, than P90 would be a better target. Many questions, assumptions, what ifs.
The fact is, you can not program an exit at P80 of the history without actually knowing the history. It's a Catch 22! This is why it becomes a discretionary action.
Ideally I would like to run a simultaneous demo account to compare results over time. But I don't think that is practical (need another VPS). I am running Zorro with MT4 since the FXCM API seems to have some issues (FXCM related.)
Maybe one day, when I have a few of my own systems running, and my account size and income is much higher, I will not feel such an urge to interfere with the systems. But for now, this helps me feel comfortable and I am happy to risk larger profits to get locked in profits.
I have one question....I am not stopping Zorro and restarting. I am closing trades externally and Zorro is detecting this from MT4. But does closing trades externally disturb Zorro? Sometimes it seems like a few days pass before a new trade is made. On the other hand, markets ebb and flow, and after a flow period that results in these profits, the ebb period may not have many signals.
|
|
|
|