100% mechanical trading systems | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Top 100 best Trading Systems: post good Trading Systems here 62 replies

Mechanical Systems 1 reply

Systems, Systems and Systems... 1 reply

Profitable trading systems, which are 100% mechanical? 2 replies

100% mechanical systems 12 replies

100% mechanical trading systems

  • #21
  • Mar 7, 2007 3:33pm Mar 7, 2007 3:33pm
Well yeah, I guess this is mission impossible then?

  • #22
  • Edited 4:45pm Mar 7, 2007 4:31pm | Edited 4:45pm
If you want a good system that makes you money, all it takes is some hard work, dedication, and discipline. Suppose you have a system that has a 50% drawdown, but appears to generate gains, too. You should always be questioning what caused this drawdown? Is there a filter that would eliminate this drawdown? It's not an easy task, but totally feasible. In fact, I managed to come up with something that works for me:

In backtests, it generates 10,000 pips on GBP/USD, GBP/JPY, and USD/JPY combined in one year. It has a winning % of ~70% and a Reward to Risk ratio of 5:3 with a 15% max drawdown. It is about 90% mechanical.

Sounds good right? Well, this is not my attempt at bragging, despite what you may think, but rather an example of how one CAN make something work. I'll have you know, that when I first started, this method had a 65% win ratio with a 1:1 Reward vs Risk ratio, and a 25% drawdown; generating about 60% of the pips.

All it takes is some experimentation. If you think you can work with something, then always be trying to improve it. Not everything will come easy...

*Step off of soap box*

Best of luck,

Kevin
  • #23
  • Mar 7, 2007 8:42pm Mar 7, 2007 8:42pm
Quoting Traex
I want a system (a single system, not a mix of 2 systems) that's profitable in trending as well as ranging market in all timeframes larger than m30, backtest proven in last 10 years at least.
Let's search till the end of the world. Good luck to me lol.
May I join you in your magical quest?
  • #24
  • Mar 7, 2007 9:18pm Mar 7, 2007 9:18pm
Quoting kharvell
In fact, I managed to come up with something that works for me:

In backtests, it generates 10,000 pips on GBP/USD, GBP/JPY, and USD/JPY combined in one year. It has a winning % of ~70% and a Reward to Risk ratio of 5:3 with a 15% max drawdown. It is about 90% mechanical.
Wow, 10,000 pips in a year, 15% drawdown, quite impressive indeed!!

However, are you absolutely sure about these results Kharvell? And what about the five preceding years, do you still get 10,000 pips/year average profit in those 3 currency pairs ?

I have personnally never seen such numbers, most mechanical systems almost never exceed 100 pips profit per month per currency pair (1200 pips a year). Please test your system with the 5 preceding years and let us know if your results still hold.

Because if they do, you are not a trader but a psychic my friend, and Soros would love to have a chat with you! ;-)
  • #25
  • Mar 7, 2007 11:18pm Mar 7, 2007 11:18pm
Quoting FXTerminator
Wow, 10,000 pips in a year, 15% drawdown, quite impressive indeed!!

However, are you absolutely sure about these results Kharvell? And what about the five preceding years, do you still get 10,000 pips/year average profit in those 3 currency pairs ?

I have personnally never seen such numbers, most mechanical systems almost never exceed 100 pips profit per month per currency pair (1200 pips a year). Please test your system with the 5 preceding years and let us know if your results still hold.

Because if they do, you are not a trader but a psychic my friend, and Soros would love to have a chat with you! ;-)
I am testing my method already. It can be found under my thread called kharvell's trades; currently waiting for my first trade of testing my new and improved version. Feel free to follow along So far my equity is up 18% since beginning of February, and I have added more funds to increased the compounding. With the new method, I expect this to be much more profitable with less risk.
  • #26
  • Mar 7, 2007 11:23pm Mar 7, 2007 11:23pm
Quoting kharvell
Suppose you have a system that has a 50% drawdown, but appears to generate gains, too. You should always be questioning what caused this drawdown? Is there a filter that would eliminate this drawdown? It's not an easy task, but totally feasible. In fact, I managed to come up with something that works for me:
If I see crazy drawdowns and crazy profits, I first suspect the system is "curve fitting" the back test data. Good to have previously untested data to confirm your system. When you start adding filters to the mix you are even more in danger of curve fitting.
  • #27
  • Mar 7, 2007 11:30pm Mar 7, 2007 11:30pm
Quoting mrmikal
Ny,

Ummm...are you sure? let's do the math.

If you go down 1000 pips with 1 mini lot on a $5K account brings you to $4K.

Then you go up 1000 pips with 1 mini lot on your $4K account...that brings you too...ummm...let's see here...$5K...back even.

Unless you're utilizing a % theory...I'm not sure where your math came from.

As far as a drawdown goes, 20% drawdown does seem a bit excessive, especially if it were quick...but I'm not sure 20% is out of the realm of possibility in any system...My 401Ks had a 20% drawdown at one point over the course of 15 years, but it didn't take long to recover from.

a 50% drawdown would have my alarm bells ringing...but you will inevitably have drawdown of some magnitude. Of course, if a single TRADE had you down 20% (unless it were long term), I'd be a little cautious, but through a series of trades, accounts could be hit easily with 10-20% drawdown...I'm not sure that's a horrible thing if the average upswing move is more than size of the 20% downswing, right?

I would have to believe that straight upwards movement (if even gradual) would be an near impossibility...manual OR automated.
if using a % based money management, (which is normally) then if u lose say 70%.. u need 340% to break even. and even if not using %MM, your leverage might not allow u to make it back as quick as u lost it.
  • #28
  • Mar 8, 2007 12:31am Mar 8, 2007 12:31am
Don,

I know what you're saying...but what you're saying isn't exactly relevant.

IF you're using a %-account position size...then YES, what you're saying is correct. I've already acknowledged that. But what is "normal" for you, isn't normal for everyone.

Personally, I don't use a %-account position size...I use a %-maximum loss, and base my position size on the confidence of my trade. The higher the confidence of success, usually, the more daring I can be. The less confident in the trade, then smaller the position size. More often than not, my position sizes are static regardless of a win or loss.

Like I said, it depends on the situation, but more often than not, my RISK is not relevant to my position sizing requirements.

And what does leverage have to do with anything? Are we not agreed, that the per-pip gain/loss on a 1-standard lot trade of the GBPUSD will always be $10 USD regardless of whether I'm leveraged at 1:1 or 1000:1??? The only effect leverage has on a trade is the margin requirement necessary...but again, that has nothing to do with %-risk OR position sizing.

Quoting don perry
if using a % based money management, (which is normally) then if u lose say 70%.. u need 340% to break even. and even if not using %MM, your leverage might not allow u to make it back as quick as u lost it.
  • #29
  • Mar 8, 2007 12:41am Mar 8, 2007 12:41am
That system has insane profits as well, such as 1426 pips profit on GBPJPY in 8 days, between Feb 26 - Mar. 6, 2007.

We believe, just as the above sentence does not represent the correct method to evaluate a system, neither does your drawdown approach. It would be misleading if you pick up the numbers between certain dates. Depending on investor's/trader's risk profile, there are different ways to evaluate a system's performance. As we explained in our thread, we prefer monthly figures and keep our statistics based on these figures. Even the month-end figures of those drawdowns showed positive pips and 11 out of 12 months in 2006 were positive.

Therefore, it is very important to define your parameters (risk profile, time frame, discipline, faith etc.) before evaluating 100% mechanical systems.

We agree with mrmikal's views on drawdowns (his/her last paragraph on post #15).



Quoting Nybegynner
That system has insane drawdowns. I think the thread starter asked for something with a bit smother equity curve.

You see, you cant just count the pips without taking into consideration what kind of draw downs you get. If you have a 50% drawdown, you have to make a 100% to just break even. This is why all these "My system makes x amount of pips a month" systems wont get you rich.
  • #30
  • Mar 8, 2007 12:53am Mar 8, 2007 12:53am
Those numbers are quite possible, in our opinion. In February, we gained 3625 pips by trading 6 currencies with a 100% mechanical trading system and this system yields 2000 pips on average per month for about 2 years. Therefore, 10,000 pips/year average with 3 currency is quite achievable.


Quoting FXTerminator
Wow, 10,000 pips in a year, 15% drawdown, quite impressive indeed!!

However, are you absolutely sure about these results Kharvell? And what about the five preceding years, do you still get 10,000 pips/year average profit in those 3 currency pairs ?

I have personnally never seen such numbers, most mechanical systems almost never exceed 100 pips profit per month per currency pair (1200 pips a year). Please test your system with the 5 preceding years and let us know if your results still hold.

Because if they do, you are not a trader but a psychic my friend, and Soros would love to have a chat with you! ;-)
  • #31
  • Edited 4:56pm Mar 8, 2007 4:42pm | Edited 4:56pm
Quoting TeamAphid
Those numbers are quite possible, in our opinion. In February, we gained 3625 pips by trading 6 currencies with a 100% mechanical trading system and this system yields 2000 pips on average per month for about 2 years. Therefore, 10,000 pips/year average with 3 currency is quite achievable.
That means your system is making 3300 pips a year average per currency per lot or 275 pips a month.

Quite frankly it would take a miracle to duplicate these results in the future !!

Again, and sorry if I am repeating myself, you need AT LEAST 5 years of data for backtesting purposes, you cannot simply test a system with 2 years of data and say "Hey look here, it's working!", been there done that, there are TONS of systems that perform extremely well in a year or two but fail miserably if you test them with more data.

My own experience backtesting systems has convinced me that pure mechanical FX trading systems that exceed 100 pips a month per currency pair per lot are EXTREMELY rare.

And I am not talking about systems that generate 20 lousy trades in 5 years, I am talking about systems that generate at least 500 to 1000 day-trades, while still keeping the drawdown below 35%

A pure mechanical system that exceeds 270 pips a month (like in your case) profit in a 5 year backtest (without any pyramiding and per lot per currency pair) WITHOUT EXPOSING THE TRADER TO HUGE RISK (like the Firebird system) simply does NOT exist, and I challenge every one in this forum to prove otherwise.
  • #32
  • Mar 8, 2007 4:59pm Mar 8, 2007 4:59pm
Quoting FXTerminator
And I am not talking about systems that generate 20 lousy trades in 5 years, I am talking about systems that generate at least 500 to 1000 day-trades, while still keeping the drawdown below 35%
why would i care about the number of trades if that system is profitable over 5 years?
  • #33
  • Mar 8, 2007 5:03pm Mar 8, 2007 5:03pm
FXT,

Normally, I would agree with you, but I feel like playing devil's advocate.

At SOME point, Carry Trades worked...however, recent events elude to the fact that now, they don't.

I think 5 years in the forex market is actually too large a subset of time. I think results would be skewed just because of the fluctuations of the world. I mean, think, 5 years ago, a lot has changed since, and the dynamics of the market are much different than they were even 5 years ago.

I think it's entirely possible, that a great system that works today could very well be obsolete in a year or 2.

So for the statement that says consistent gains of that magnitude for a single unaltered system would be nary impossible...I agree...putting blind faith without re-evaluating a system one weakeness start to expose it is not a good system. Whether or not a single person or team of traders could consistently make that much profit, I'm not sure that's our place to judge. I think it's highly improbable, but not unlikely.

I don't want anyone to feel scared away from a system that's doing well now because you HAVE to be aware of what's working and why it works. What I do think is important, and the point I think you're trying to make, FXT, is that it's important not to believe these systems will work forever. It's simply improbable.

It's like that Freedom trading thingy...or whatever, it exploited the carry trade on the GBP/CHF...well, guess what, as the carry trade unwinds, that system's gonna be crap, and returns will weaken...does it mean the system didn't work? absolutely not! it DID work, and quite well...but it isn't guaranteed to keep working, and in fact there is the expectation that results will now be paired down due to the GBP weakness against the CHF. It's the same with ANY system, correlations, EMAs, Oscillators, etc...they work during specific periods, but they don't work ALL of the time...at some point they stop working, only to cycle back in after a while.

I think it was a pretty awesome example that occured a few months back. EVERYONE was starting to talk about Carry trades, etc etc. Some people bought at the top, some people shorted, but I tell you what...there was a point where Carry trades were still pretty decent even AFTER you could hear people jumping off the bandwagon of the carry trade. For a short while, there was still tremendous growth in the USD/JPY and GBP/JPY right before the bottom fell out...my point here being that even if system starts exposing weakness, if the system is still "trendy", well, you can STILL make money off of it. The key here is to NOT be the lasted dude to the party...that's when all the beer is gone and the women have hooked up...know what I mean?

Just be flexible...if a system is working, that's great, ride it until you see weakness, and then jump off when you're not comfortable anymore.
  • #34
  • Edited 5:37pm Mar 8, 2007 5:04pm | Edited 5:37pm
Quoting FXTerminator

A pure mechanical system that exceeds 270 pips a month (like in your case) profit in a 5 year backtest (without any pyramiding and per lot per currency pair) WITHOUT EXPOSING THE TRADER TO HUGE RISK (like the Firebird system) simply does NOT exist, and I challenge every one in this forum to prove otherwise.
you say this as if the market doesn't change in those 5 years. But since the market does change, its permissible in the world of 100% mechanical systems to tweak the system now and to stay in sync with the market. So now repeat your challenge allowing for that change:

"A pure mechanical system that exceeds 270 pips a month (like in your case) profit in a 5 year backtest (with occasional tweaking)-(without any pyramiding and per lot per currency pair) WITHOUT EXPOSING THE TRADER TO HUGE RISK (like the Firebird system) simply does NOT exist, and I challenge every one in this forum to prove otherwise."

Now what is occasional? I believe if you take 8 months of market data, and compare the first 4 months with the last 4, the responsiveness and size of the market waves will have changed completely. So I believe in optimizing any system over the last 4 months.
  • #35
  • Mar 8, 2007 5:12pm Mar 8, 2007 5:12pm
Quoting richy
why would i care about the number of trades if that system is profitable over 5 years?
That should be obvious Richy.

If a system is STILL performing well after 1000 trades for example, it only means one thing : your system is definitively giving you a statistical edge.

If system A gives me 10,000 pips after 20 trades and System B "only" 3000 pips after 1000 trades, both in a 5 year period, I would always choose to trade System B.

Why?

Because how do I know the results from system A (20 trades) are not due to chance alone ?
  • #36
  • Mar 8, 2007 5:25pm Mar 8, 2007 5:25pm
Quoting FXTerminator
A pure mechanical system that exceeds 270 pips a month (like in your case) profit in a 5 year backtest (without any pyramiding and per lot per currency pair) WITHOUT EXPOSING THE TRADER TO HUGE RISK (like the Firebird system) simply does NOT exist, and I challenge every one in this forum to prove otherwise.
Terminator,

You may call this realism, but I call this pessimism. Instead of arguing with you that it is not only possible, but also I have done it, I will let you watch my trades in my thread in Journal Factory Enjoy being proved wrong, my friend. Oh, and just so you know, February was a 1000 pip month for me; kind of making your above statement false.

Regards,

Kevin

P.S. My apologies if this post seems a bit rude. Just tired of people doubting things are possible WHEN THEY HAVE BEEN DONE BEFORE; i.e. George Soros, Bruce Kovner, Richard Dennis, Ed Seykota (who simply has a program make all his trades everyday) and many more.. Read any of the Market Wizards Series by Schwager.
  • #37
  • Edited 5:59pm Mar 8, 2007 5:48pm | Edited 5:59pm
Quoting mrmikal
I think 5 years in the forex market is actually too large a subset of time.
Au contraire, if you can find 10 years of historical FX tick data, it's even better. And if your system is still performing well after 10 years, like it normally should, then you have found a goldmine my friend.


Quoting mrmikal
I think results would be skewed just because of the fluctuations of the world.
Absolutely not, the minute you do that the system becomes just another highly optimized garbage system.

Quoting mrmikal
I mean, think, 5 years ago, a lot has changed since, and the dynamics of the market are much different than they were even 5 years ago.
I talked about this "markets change over time" myth before in this forum. Markets do NOT change over time, they just go up, down or sideway. That's ALL they do year after year, take any chart from any time frame and see for yourself. But hey, FX fund managers and traders alike love to blame "changing market conditions" (Yeah right!) for their failure.

In fact the FX market is becoming more and more volatile every year (it is normal since more and more traders are switching to FX), and that means even MORE profits for the good traders.

Quoting mrmikal
I think it's entirely possible, that a great system that works today could very well be obsolete in a year or 2.
If a system stops working after two years then your system was not robust in the first place, period, there is nothing wrong with the market. Just dump that losing system fast and back to the drawing board. Robust systems do not stop working, they go into drawdown from time to time but that's about it.

If the mathematical edge was there from the start it will always be there in the futrure, just like in a casino, the owner has the edge and will always have the edge.


Open the February issue of Futures Magazine, there you will find the top 10 trading systems of all time. These systems are STILL making money today, 10 years after their release! This is how robust they are.

Ok, I am going to eat something now, I will try to answer some more questions/comments later.
  • #38
  • Mar 8, 2007 6:02pm Mar 8, 2007 6:02pm
FXT,

My friend, I respect your opinion, but I'm going to have to respectfully disagree.

The world does change, and while the market moves up and down and sideways, they do so for different reasons. Understanding the true dynamics of those systems implies that systems that do work during one set of events may not for another. For example...oil...it can go up and down and sideways based on geopolitical factors...those factors may or may not be there the next day (you get the point).

While we'd love to believe that there is a mathematical formula to the markets, there just isn't...any system you use whether you get it to work for 10 years or 2 years will ALWAYS be curve fitting...you'll just be fitting a different curve.

What you're saying has logic behind it, so I know you're not talking out of your butt...but you obviously view the market a little differently than I do...and that's fine...however, it means that philosophically, we'll not be seeing eye-to-eye on this matter. I see your points, and I'm sure you see mine, but alas, it's not meant to be, my brother :-)

From my perspective, the FX market at this point is quite compressed. cycles that normally take years to play out in other more traditional markets tend to take much less time when currencies are involved. Trends developer faster, and they don't last nearly as long (in terms of time). It is for this reason that I think that trading systems can be found that exploit the market for a short period of time to a varying degree of success, but may falter when new fundamentals develop. That's the reason why I can't reach accord with what you're saying.

It's true that if a system could last 10 years that it may be a better system, but I don't think any system that shows a degree of success for a lengthy period of time (and this is probably where we disagree most), is a losing effort. I'd be more than happy to participate in a system that allowed me to gain 20,000 pips the first 2 years and lose 5000 pips the next. That's 7,500 pips a year in my book...and with the amounts that I trade, that'd be a damned great year!

Still, what you say has some validity so I'm not discounting everything you're saying...but let's just say that I'd rather be wrong 49% if I'm making stellar returns than not being in the market while I search for the 1 system that can last 10 years. It doesn't mean that I'll just go in with blinders and never keep looking for the next big thing...absolutely not...like I said, you need to be flexible in this market.
  • #39
  • Mar 8, 2007 6:12pm Mar 8, 2007 6:12pm
Quoting kharvell
Terminator,

You may call this realism, but I call this pessimism. Instead of arguing with you that it is not only possible, but also I have done it, I will let you watch my trades in my thread in Journal Factory Enjoy being proved wrong, my friend. Oh, and just so you know, February was a 1000 pip month for me; kind of making your above statement false.

Regards,

Kevin

P.S. My apologies if this post seems a bit rude. Just tired of people doubting things are possible WHEN THEY HAVE BEEN DONE BEFORE; i.e. George Soros, Bruce Kovner, Richard Dennis, Ed Seykota (who simply has a program make all his trades everyday) and many more.. Read any of the Market Wizards Series by Schwager.
Sure, I will keep watching, if your system can still produce 275 pips a month per lot per currency pair, with no pyramiding of any kind and no mega drawdowns in the process, 5 years from now, then you are set for life.

But remember, it must be a totally mechanical system with clear-cut trading rules and pre-determined stop/limit. In other words, 100 traders trading your system should be able to open/close positions at exactly the same price.

So keep up posted and may the pips be with you.
  • #40
  • Mar 8, 2007 6:30pm Mar 8, 2007 6:30pm
Quoting FXTerminator
But remember, it must be a totally mechanical system with clear-cut trading rules and pre-determined stop/limit. In other words, 100 traders trading your system should be able to open/close positions at exactly the same price.
It doesn't HAVE to be anything. It can be whatever I want it to be. I understand what you are saying if I plan to distribute it so others can trade it, but those are not my intentions. I have said before that my system is 90% mechanical, and I like it just like that.

Anyway, I've enjoyed the debate with you, sir, and best of luck to you!

Regards,

Kevin
0 traders viewing now