$500 to Million in 11 months - Higher Risk vs. Higher Reward EA | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

What's your TP? Fixed reward vs dynamic reward 5 replies

Low Risk, High Reward Trading Experiment 3,318 replies

Rationale behind "Risk Per Trade" and "Risk/Reward Ratio"? 8 replies

Risk Reward Ratio and Account Risk 30 replies

Results for Low Risk,High Reward 11 replies

Attachments: $500 to Million in 11 months - Higher Risk vs. Higher Reward EA
Exit Attachments

$500 to Million in 11 months - Higher Risk vs. Higher Reward EA

  • #1
  • First Post: Edited Jun 8, 2014 8:20am Jun 6, 2014 10:02pm | Edited Jun 8, 2014 8:20am
Here's an example of a very high risk and high reward EA.

Features:

Multiple positions taken per signal with varied SL/TP.

Hidden SL and TP.

Overall Risk Percentage taken is scalable up or down.

Non Martingale, but dynamic lot sizes according to equity at entry.

Trades both Long and Short per entry signal but quantity and exits differ. (Hedge).

Takes up to over 100 positions per entry signal.

NON-Indicator based. Uses no indicators. Just price movement action.

In reality it's 4 EAs merged into one EA, it scans for for different entry parameters and executes trades independently for each.

Trades based on bar opening. Open test, Control test and Tick test produce same results.

Based on 2.5 pip spread, no commissions.

Similar version of EA engine is currently trading in Live real money account and is entering and exiting trades perfectly according to live signals and live market prices.

After over 400 downloads of the PDF EA stats in less than 24 hours, I've taken it off the server for security reasons. I had no idea that it would generate this many eyeballs and had shown a little more than I perhaps should have. I may repost the PDF with proprietary EA settings and broker data whited out at a later date. Sorry.
  • #2
  • Jun 7, 2014 3:43am Jun 7, 2014 3:43am
I need this EA dear.... Pls help me.... How it works on live trade.....
  • #3
  • Jun 7, 2014 3:49am Jun 7, 2014 3:49am
Pretty cool, backtest the strategy on 10 other date ranges and let's see how the results vary.

I think the only problem that could happen is if you have more consecutive losses then wins, could take out your account.
  • #4
  • Jun 7, 2014 4:01am Jun 7, 2014 4:01am
where can i find the EA you have not posted it. could you please post here in the forum so i can test it.


thanks,
MRMALIK1
  • #5
  • Jun 7, 2014 4:54am Jun 7, 2014 4:54am
Quoting claudfs
backtest the strategy on 10 other date ranges and let's see how the results vary. I think the only problem that could happen is if you have more consecutive losses then wins, could take out your account.
Walk forward test within same dates produce similar 10x plus results starting from each 1st of the month.

A straight up or straight down market for an extended period of time with absolutely no retracements would take out the account. Am working on ideas to implement for this scenario as well.
  • #6
  • Jun 7, 2014 8:14am Jun 7, 2014 8:14am
Quoting Pipologist
Takes up to over 100 positions per entry signal. Thoughts, comments? {image}
I wouldn't be trading live based on backtest data...

Even though it is UJ and can trade in a very limited range, it will trade the same as all other pairs.

Lock in trades with the range and you will continue to lose on trend.
Lock in trades on trend and you will continue to lose on Range...

Tough deal even with hedging and especially with a dynamic lot size implemented.

I think Lifesdream tried this same ideas, several times over and I've worked with some of that same code....all sunshine and happiness for a month, maybe 2...and then Ka-BOOM!

Be interesting to watch if you throw up an explorer or myfxbook link....

Indicator based, trigger dynamic, martingale grid trader, rotating by signal on 28 pairs in a first come first serve to 3 separate time frames, live test data.
HERE.

Good luck in your quest!

Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #7
  • Jun 7, 2014 8:49am Jun 7, 2014 8:49am
Quoting Magix
{quote} I wouldn't be trading live based on backtest data... Even though it is UJ and can trade in a very limited range, it will trade the same as all other pairs. Lock in trades with the range and you will continue to lose on trend. Lock in trades on trend and you will continue to lose on Range... Tough deal even with hedging and especially with a dynamic lot size implemented. I think Lifesdream tried this same ideas, several times over and I've worked with some of that same code....all sunshine and happiness for a month, maybe 2...and then Ka-BOOM!...
Thanks for your comments. I agree with most of what you said.

Assuming my 1 minute data is accurate, I have no reservations trading live on backtest data. I don't use indicators or strategies that alter entry or exit based on data other than price. A similar but less robust ea (2 instead of 4 EAs in one) is trading live starting recently and taking the exact same signals and exits as the backtests show for the same trading time periods between the live and the backtests. As long as the market does not goes absolutely straight up or down with zero retracement it should be profitable but with wild equity swings between sets of open positions.

The US/JP was almost straight up from end of summer to winter last year, and then in a range this year. It was massively profitable from July 2013 to November 2013, then flat, and then from January 2014 to now it was also quite profitable. So the market would really need to go vertical with no retracements whatsoever for it to wipe out. It's worth the risk with small starting capital. If I was trading starting it with more than 20,000 I would simply dial down the risk per position and thus the DDs accordingly.

I saw your MyFXbook, good job. I would say that I would be interested to know more about your EA but I would be more interested to see how it performs with a NON marty MM method. Using one or two max three consecutive bad trades wipe out all martys. If your method is sound it should be able to perform just as well or even better without a marty.

As for me using tradex or myfxbook, it's not going to happen. I'm not trying to attract investment capital and don't need to prove to the world whether my EA works or not. If it doesn't work, I remorse privately. If it does work, I'm happy privately without dozens of people contacting me to give them my EA.

Finding a balance of being connected to a community of like minded coders and traders and while keeping away folks asking for a free handout without even really knowing what they are asking for is sometimes more work and not really related to improving the bottom line.
  • #8
  • Jun 7, 2014 9:21am Jun 7, 2014 9:21am
Quoting Pipologist
{quote} Thanks for your comments. I agree with most of what you said. Assuming my 1 minute data is accurate, I have no reservations trading live on backtest data.

Your 1 minute DATA can be 100% spot on...backtests are bullshit!

You don't have to believe me...but you should take the challenge...

Trade your EA for 1 month...start at the first Sunday/Monday market bell, and allow it to trade for 5 consecutive weeks. (I'd recommend demo)

After Market close on week 5...you will have HARD, Solid data. As well as have the true value of each tick and candle (lol), in your broker history...

Now, run your backtest.

If you come up with anything that resembles accuracy, and post this data....
1. You will be the first person in the history of Meta4 to have accomplished this.
2. I would probably have a stroke!

We've all had our moments in the sun and dreams based on backtests...but are you really ready to gamble your current account against something that has been consistently proven to be inaccurate to the point of absolute falsehood?

As for my EA..

It's a counter trend trader, based on profitability on probability.

By working with both daily trends, weekly cycles and applying this to multiple time frames...it does what it does by looking for short ranges after market typical movements...

Perfection? Hell NO!
HIGH RISK? HELL YEAH!

But the link I posted wasn't based on the mythical Unicorn of back test report accuracy...that is live testing.
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #9
  • Edited 9:42am Jun 7, 2014 9:30am | Edited 9:42am
Quoting Magix
{quote} Your 1 minute DATA can be 100% spot on...backtests are bullshit! You don't have to believe me...but you should take the challenge... Trade your EA for 1 month...start at the first Sunday/Monday market bell, and allow it to trade for 5 consecutive weeks. (I'd recommend demo) After Market close on week 5...you will have HARD, Solid data. As well as have the true value of each tick and candle (lol), in your broker history... Now, run your backtest. If you come up with anything that resembles accuracy, and post this data.... 1. You will be the...
BTW, the data is from a live account server with more than 5 weeks of tick data. The EA is trading live. Like I said I have a very similar EA that's working exactly as tested. It's been up and trading live over 5 weeks and actually has better signal execution than it's historical backtests...i.e. better fills.

As I stated. My EA does not rely on tick data. if the candle contains the price that the signal executes off of, the EA sends multiple buys and sells at market. It does not rely on bid or ask prices to generate signals other than market orders. The rest of the calculations of signals are based on previous price action off of printed HLOC data. I deliberately set the spread in the tests a bit higher than actual live spreads thus live trading is getting slightly better fills.

I didn't start trading off of demo backtest data.

I also would rather run it live on real money even small money rather than demo it.

Any other challenges or challenges to post the results for the public to see, we'll I guess you really didn't read or understand my previous posts.
  • #10
  • Jun 7, 2014 9:41am Jun 7, 2014 9:41am
Quoting Pipologist
{quote} BTW, the data is from a live account server with more than 5 weeks of tick data. The EA is trading live. Like I said I have a very similar EA that's working exactly as tested. It's been up and trading live over 5 weeks and actually has better signal execution than it's historical backtests...i.e. better fills. I deliberately set the spread in the tests a bit higher than actual live spreads. As I stated. My EA does not rely on tick data. It compares
lol

Whatever dude...

Don't say I didn't try to warn ya...

Here's your last 5 weeks on USDJPY...IF I am not sadly mistaken, next week, is breakout week...the gloves come off and the market dynamics change.

Good luck!

Attached Image (click to enlarge)
Click to Enlarge Name: 5weeks.png Size: 27 KB
Money Can't Buy Happiness. Poverty Can't Buy SHIT! You Choose!
  • #11
  • Jun 7, 2014 10:00am Jun 7, 2014 10:00am
Quoting Magix
{quote}but are you really ready to gamble your current account against something that has been consistently proven to be inaccurate to the point of absolute falsehood?
There are two potential problems with backtests.

1. Accuracy of data.
2. Accuracy of the EA trading with the same data.

For example, in the beginning I used to code EAs, test them, and then trade them. But what I found is that the live trades and the concurrent backtested EA over the same dates and the same data, would produce different entries and exits. This is because either, the EA takes too many signals, or because the indicators used calculate different values over different time periods. Or, because the EA taking signals at different times.

I've learned how to code to get my live EAs to sync to backtested data and also know when to start the EA so that it correlates with the same entry signal times as the backtested EA would over the same live data. The algorithms and logic is solid. So I'm not worried about "data accuracy", and I understand the risks of different market conditions and how this EA will perform in those conditions that it is not designed for.

Fortunately, the market ranges or breaks and retraces generally more than it goes consistently vertical so I'm not concerned about risking a few hun or thou to "test it live". I just don't need to show the world live results as others do with tradex or myfxbook.

I just wanted to generate commentary. Sometimes, someone says something in these threads where I say, "yeah that's a good idea".

Anyway, keep it coming. Don't need to attack, just keep it constructive.

I hope it's a breakout week. Just not a run away breakout. Breakout week and normal retracements followed by more breakouts is how this EA rakes it. It would have to breakout over 200 pips after a signal is generated without retracing at all for this EA to bust.

I did not say it but it works off of a very simple volatility calculation without derivative calculations. I'll leave it at that.
  • #12
  • Jun 7, 2014 2:51pm Jun 7, 2014 2:51pm
You could never have a 100% accurate back-test.. at best is 90 to 95%.. the 5-10% unknown could (most likely would) make a different on the overall result. I have had a up close hands on experience with this when team of programers were working on my EA project. Back test would only give you some ideas of what a successful EA could project but never a 100% accuracy.
Passion, Persist, Patience, Protect
  • #13
  • Edited 7:19pm Jun 7, 2014 6:28pm | Edited 7:19pm
Quoting FXEmpire
the 5-10% unknown could (most likely would) make a different on the overall result. Back test would only give you some ideas of what a successful EA could project but never a 100% accuracy.
Agreed completely.

The best we can do is to imagine or to learn about as many "unknowns" as we possibly can and then to work it into the plan as much as possible.

Unknowns:

1. Will the market go up or down or sideways?
A: It doesn't matter where it goes. The method should make consistently money under any market or at best not consistently lose money.

2. How long will the market stay up, down or sideways?
A: This is the billion dollar question to me. And is more important than the 1st question.

3. Will my server be online when it needs to execute trades?
A: I hope so, so far it has been pretty reliable.

4: Will my broker allow me to withdraw large sums of profit when requested?
A: Yes.

5. Will this EA take 500 to a Million in 11 months like the backtests?
A: Probably not. Understood.
Does it have a chance of taking it to 5000 within a few months. Absolutely. It has proven itself to double itself within a few weeks.

6. Will this EA go to zero on 1 or 2 bad trades like a martingale strategy?
A: No.

7. Can this EA recover from a 90% DD and still make a huge profit?
A: Absolutely. In fact, it can recover from up to 99% DD and still earn a decent profit. The EA is designed to avoid Margin Stop Outs from the broker server and take itself out of the market stepwise during DD periods. Most people including myself would not be able to recover from a 90% DD when trading manually. Most EAs and people crash and burn after an 80% DD either by margin stop out from the broker and/or psychological defeat. My main purpose of this EA is to not let it emotionally drain me and burn me out like manual trading does. It can still function according to it's plan after an extended DD period, whereas I trading manually start making massive errors in judgement.

8. Am I naive to think the live trading will mirror the backtests?
A: Absolutely not.

9. Will news spikes ruin the EA?
A: Probably not. Most of the time the EA will have traded in and out before any massive news releases or come into the market after an impacting news release.

10. Can a black swan destroy the results?
A: Umm...Yes. But if that happens, it's probably not just my EA that I'm worried about.

I don't expect it to go to a million. However I actually have more aggressive tests up to 85% gradual DD (not one one price move) that generate even more profits but I didn't post them here because those kinds of results bring out even more.....what's the word here....

I know there are EAs out there that can backtest more profits. But can they contain risk?

Is the EA still risky?
Absolutely. But much less than my extended periods of manual trading. So for me, speaking personally, it is preferable or an adjunct to my manual trades and other EAs.

Without going into my exit strategies and MM methods in detail. All I can say, is that although it seems I'm trying to maximize best EA backtest results here, what my EA is actually trying to attempt, is to contain market price risk as much as possible when it does enter trades. While at the same time, risking "gambling capital" to achieve higher than "normal" returns.

On a 100k starting capital account, I would naturally reduce the overall risk and gain to about 20% DD or less, and about 80% return.

On a 500 starting account, which is the example shown here, I would dial up the risk vs return. Turning 500 into 1000 in a year is not as challenging as 10xing it or much more.

So can we get off the backtests critiques?

Why even bother posting this kind of backtest on FF?
Why not? It's for the fun of it, and like I said to generate commentary and to see different kinds of comments, market and trading comments and psychological perspectives.

Many years ago around 1995, an "ex-Turtle" who was a member of the original "Turtles" as mentioned in Market Wizards was a client of ours at the company I worked at and used our proprietary trading analysis software. I spoke to him on the phone a few times, a really laid back guy. I won't say his name here. That year he published a 400%+ return with a 500 million dollar currency hedge fund (I remember the return, but not the amount of the hedge fund, it was massive though). It opened my eyes to the level of "normal returns" for gurus as opposed to "regular returns" for retail traders. After that I was never psychologically happy with "just" a 100% return.
  • #14
  • Jun 8, 2014 2:07am Jun 8, 2014 2:07am
Quoting Pipologist
{quote} Agreed completely. The best we can do is to imagine or to learn about as many "unknowns" as we possibly can and then to work it into the plan as much as possible. Unknowns: 1. Will the market go up or down or sideways? A: It doesn't matter where it goes. The method should make consistently money under any market or at best not consistently lose money. 2. How long will the market stay up, down or sideways? A: This is the billion dollar question to me. And is more important than the 1st question. 3. Will my server be online when it needs to...
pipo, I think you misunderstood what I meant by unknown.
unknown is the amount of signals that don't show up on your backtest, bcz it captures 90% of the whole 100%. so you won't know what is the outcome of that 10% worth of signals. Which it could for example change the outcome of the win rate accuracy that you may be depending on from lets say 50% to 60%.. or other factors. basically your in the dark for the about 10% of what your trying to achieve and in putting a EA together that is a huge sum.

Thats all brother just some facts from my own experience that I thought it could keep you aware in your journey, just ignore it if you have too.
Passion, Persist, Patience, Protect
  • #15
  • Jun 8, 2014 3:59am Jun 8, 2014 3:59am
Quoting FXEmpire
{quote} pipo, I think you misunderstood what I meant by unknown. unknown is the amount of signals that don't show up on your backtest, bcz it captures 90% of the whole 100%. so you won't know what is the outcome of that 10% worth of signals. Which it could for example change the outcome of the win rate accuracy that you may be depending on from lets say 50% to 60%.. or other factors. basically your in the dark for the about 10% of what your trying to achieve and in putting a EA together that is a huge sum. Thats all brother just some facts from...
Good point. I understand what you are saying. However with this EA, real live trading has shown to take more signals than the backtests. You may wonder how.

1. I've widened the spread in the BT a bit more than the average live spread.

2. ALL signals of entries and exits are sent as market orders.

3. ALL signals are calculated off of one minute HLOC data, not from an indicator or a coefficient of price data.

The EA is simply sending multiple market orders on entry and exit. As long as the required calculated equity is in the account the orders will execute.

Regarding this, the risk is that the server is offline for the minute or minutes that the trades would have executed, since the exit stops and profit hidden, this is also a concern but a managed risk.
  • #16
  • Jun 8, 2014 4:18am Jun 8, 2014 4:18am
there are no easy money - 500 to 1mln in 11 months seems nearly impossible. even if the plan works in strong bull/bear market and brings profits - market is not always trending and there are no methods that work in all market conditions. I could get 30-80% profit a month in the past as market had strong trends. at the moment I think 10% a month would be more than I dare to dream of.
On my way to financial independence - if others can do it - I can do it
  • #17
  • Edited 8:13am Jun 8, 2014 7:26am | Edited 8:13am
Quoting Eur Usd
there are no easy money - 500 to 1mln in 11 months seems nearly impossible. even if the plan works in strong bull/bear market and brings profits - market is not always trending and there are no methods that work in all market conditions. I could get 30-80% profit a month in the past as market had strong trends. at the moment I think 10% a month would be more than I dare to dream of.
Agreed. The interesting thing about this EA is that it will not generate an entry in a market that has not moved up or down. In a small ranging market it will just sit and wait. So it will not lose money in a tight ranging market, and it doesn't have to be a strong bull or bear to profit. In fact, if it's a market that goes straight up or straight down with no retrace then it can be a problem, fortunately most of the time, markets go up or down and then retrace some and then continue in either direction. The EA does not rely on indicators that can give false overbought or oversold indications. In ranging markets indicators work well. In ranging markets this EA sits and waits if it's not in a trade already, and if it is in a trade already, it just waits till the market moves and takes out SLs and TPs in EITHER direction.
  • #18
  • Jun 9, 2014 11:45am Jun 9, 2014 11:45am
Can I ask what the purpose of this thread was? What are you wanting to achieve or offer?
  • #19
  • Jun 9, 2014 10:00pm Jun 9, 2014 10:00pm
Quoting Mister Dave
Can I ask what the purpose of this thread was? What are you wanting to achieve or offer?
As stated in an earlier post, just wanted some commentary. Sometimes, there are other useful or interesting posts.

Thread closed.
  • #20
  • Jun 11, 2014 5:39am Jun 11, 2014 5:39am
Drawdown: 59.59%

a good reason to keep away
0 traders viewing now