Learning how to code EA

Trading Systems Expert Advisors and Automated Trading
1

Hi all,

I lately started doing manual backtesting of system ideas using metatrader. Then I realized a few things:

  1. Writing a program would make it much easier to quickly ascertain the effectiveness of indicator based strategies over a long period of time, as opposed to manual backtesting, and try multiple ideas faster. Right now I backtest everything by hand, doing random samplings of a month or two here and there.

  2. If I wrote an automated program, I could catch trades that were available at times when I’d be away from the computer. According to one such strategy, I could have had a very handsome return if I had traded at those times, versus a very poor return if I had stuck to my usual trading window.

Many years ago I fooled around with Basic and knew very little C (before the days of C++, hehe), so it’d be almost like starting all over again.

How long would you say it takes to have an operating knowledge of MT4’s language, enough to build EA’s? I found the official guide, and also one more modest tutorial, anyone have any other suggestions?

Also, are there any opinions as to the EA generators that are out there? I’ve seen a few, one was too simple and another one is a pay one. Any suggestions, and are they worth it or is it better to simply bite the bullet and learn to code?

Thanks for any advice!

2

Testing your system using Excel might be a cheap and simple way for you to learn the nuts and bolts of your system. Then you might have a better idea of what software you might like to use.

If you’re in a hurry, there are plenty of programmers who can code your EA for you.

3

Interesting, how would I go about using excel to write a program though?

4

You could import OHLC data into a spreadsheet and set up formulas in adjoining cells.

I fooled around with Basic back in the days also and didn’t pick up programming until I started trading. I learned some VBA to code my own backtesting application in Excel; it only took me about a month to feel comfortable with the language. Now, I can back-test any currency for any period of time I can get data for, saving me hundreds if not thousands of dollars.

More importantly, having coded everything in my system myself, I understand exactly how my system works and I have the confidence to stick to it.

A little effort and patience at the beginning more than pays for itself.

5

Interesting. But it seems like you wrote an entire backtesting program. That’s probably a bit above my head right now. I think that if I’m going to bother learning I might as well just bite the bullet and learn Metatrader’s language.

I heard that Tradestation has “Easy Language”, but that’s easy for people who have the money to pay for Tradestation!

I guess I’m going to have to resort to manual backtesting for a while, until I can get my head around programming. Besides, I see so many people who have all kinds of Expert Advisors that make such killer returns, which looks all fine and dandy but for some reason none of them are millionaires yet, even though according to their backtests, they should be :smiley:

6

I can’t speak for everyone, but many people get all starry-eyed when they see the “results” of commercial EAs. Later, when the EA gets into a prolonged draw-down, they give up and stop using it.

Any EA can be made to look great in a back-test by leaving out commission or applying it over a favorable period and ignoring less favorable ones.

Programming your own system and testing it yourself over a significant period of time (good AND bad) gives you a good sense of whether or not you can actually stick to it.

That said, you certainly don’t need to start with Excel. You could learn MQ4’s language and trade with a broker who accepts it. Or, you could learn nothing and trust someone else’s research.

7

Yes, I remember when I did my first backtest. Without commissions it was absolutely amazing - 33% monthly return. Unfortunately, the broker has to eat something too :smiley:

Perhaps you’d be kind enough to give me some advice here.

My target is to develop a trading system (or series of systems) that puts out at least 10-12% each month, with a bi-monthly compounding period. If I compound at 21% each two months, I will turn a 3.12 pre-tax multiple each year. I figure if I can pull that off consistently, without taking serious risks, I will be in good shape and be able to make a living doing this while compounding my retirement account.

The trick is to get a system that is consistent enough to allow compounding every two months, one that won’t eat two months’ worth of return in a flash. My thinking is that this requires a system which is capable of producing fairly frequent trades, so that each month there is a high enough statistical sampling that will permit effective compounding.

Without calculating commissions and slippage, I’ve come to the formula that a system needs to be at least 65% effective with a 1:1.5 risk reward ratio. That works out to 25 pips stop loss, 38 pips gain per trade if we decide to run a frequent intraday trading operation. Within any given sampling of 20 trades, it needs to produce no fewer than 13 winners and 7 losers. This is assuming no trailing stops.

Then comes in the problem of slippage and commissions. On Euro Dollar and Dollar Yen, the spread is usually no more than 3 pips. On something like the Looney or Oz, its usually around 5. Now the question is, how and where do I calculate commissions?

With each position I exit at a 40 pip profit, and a 25 pip loss. My profit is 40 pips, loss is 25, but where do I calculate the commissions, when I lose a trade, win a trade, or both? Say the commission is 5 pips, so would it be:

Win: 35 pips

or

Loss: 30 pips

Is it both?

Next, how do I calculate slippage into my results, and is there any good number to use?

Thanks for your advice!

8

My target is to develop a trading system (or series of systems) that puts out at least 10-12% each month

That’s a tall order. To achieve that kind of results, you would have to endure large draw-downs. It’s possible, but I don’t know how many months you could sustain this without wiping out your account.

Most professionals would be happy to pull off 20% yearly. You might do better by risking more than they do, as they are somewhat constrained by what their clients will bear.

Conservatively, to achieve 20% return yearly, I am prepared to see a 30% draw-down.

with a bi-monthly compounding period. If I compound at 21% each two months, I will turn a 3.12 pre-tax multiple each year. I figure if I can pull that off consistently, without taking serious risks, I will be in good shape and be able to make a living doing this while compounding my retirement account.

My understanding of compounding is that I leave my earnings in my trading account and bet a fixed percentage of my total equity. This way, my bet-size decreases when I’m in a draw-down, and increases as my equity does.

I might take out 20% of my earnings per year and allow exponential growth to work its magic.

My thinking is that this requires a system which is capable of producing fairly frequent trades, so that each month there is a high enough statistical sampling that will permit effective compounding.

To achieve that kind of statistical sampling, you might have to pay your broker more than what you earn. You had it right when you said that they must earn their living too. Each time you trade, they earn. Doing the math yourself is eye-opening.

Without calculating commissions and slippage, I’ve come to the formula that a system needs to be at least 65% effective with a 1:1.5 risk reward ratio. That works out to 25 pips stop loss, 38 pips gain per trade if we decide to run a frequent intraday trading operation. Within any given sampling of 20 trades, it needs to produce no fewer than 13 winners and 7 losers. This is assuming no trailing stops.

I’m a trend-follower so my thinking is different. My winning percentage is below 50% but my winning trades bring in, on average, twice as much as my losing trades. Smaller bet sizes spread across many different markets allow me to be in those large moves I hope to benefit from.

Other than that, I have no fixed targets, as I don’t attempt to predict/control the future.

Then comes in the problem of slippage and commissions. On Euro Dollar and Dollar Yen, the spread is usually no more than 3 pips. On something like the Looney or Oz, its usually around 5. Now the question is, how and where do I calculate commissions?

The amount you pay for commission also depends on your lot size. So, if your broker takes 3 pips per EUR/USD trade and you trade mini lots, you will pay $3. So if your profit is $40, they get $3.

With each position I exit at a 40 pip profit, and a 25 pip loss. My profit is 40 pips, loss is 25, but where do I calculate the commissions, when I lose a trade, win a trade, or both? Say the commission is 5 pips, so would it be:

Win: 35 pips

or

Loss: 30 pips

Is it both?

Again on EUR/USD, a 40 pip take profit target and a 25 pip stop-loss would net you $37 if you win and -$28 if you loose.

However, a 25 pip stop-loss might make it difficult to reach your $40 target without getting stopped out. Your win percentage is reduced. Here again, your own back-test can help you see the mechanics of it and set the optimal parameters. Ballpark, for this kind of trading, be ready to take at least a 40 pip stop-loss for a 40 pip target.

I would add that this kind of frequent trading enriches your broker and leaves you sweating over charts for pennies if you’re lucky.

Next, how do I calculate slippage into my results, and is there any good number to use?

Slippage depends on your broker and on market conditions when your order is filled. When using short-term systems, you will find that they are extremely sensitive to slippage and commission charges, enough to make a good system in theory non-profitable in practice.

In contrast, when back-testing my long-term system, I can double or triple my commission and slippage and find no significant difference in the results.

Thanks for your advice!

Glad to help.

9

I’ve often thought about this question, and I’ve draw several conclusions as to why. From what I’ve learned, it seems that while professionals do need to be conscious of their risk factor, they also have one more big handicap - they trade with size and they have to enter their positions at poorer execution rates. Consequently, they have a tough time adopting short term strategies (unless we’re talking floor traders) because they are inefficient for them. For example, if some big trader has a multi-million dollar position, an immediate fill is not always possible, they have to go through a dealer and those guys have to work the order. Given the fact that a currency can move 100 pips or more within a day, the entry rate can be all over the place. Also, such large orders could encourage the dealer to front-run, which is legal in FX.

I think for this reason mainly, people who trade eight and nine figure accounts are happy to get 20%-30% ROI, while small and fast traders can run their account to over 100% in a year. Either that, or the small traders are snails crawling on the edge of a razor. I have heard of some very good short term traders that average 1-2% a day, that is the goal that I’m shooting for.

From a mathematical basis, it seems to me that daytrading is the quickest means of potential compounding. Also, it will let you know if you’re right or wrong with your strategy pretty fast. If you have a strategy that’s say 50% effective with a 1:2 risk reward, you’ll find out much sooner how really profitable it is as a daytrader, rather than a long term trader. In my view, long term traders can get hurt badly if fundamentals make a sharp reverse, several weeks worth of profits can get suddenly dashed because Bernake says something at a press conference.

Also, I don’t imagine how a long term trader who gets 20% annually can live off their trading unless they have a $200,000-500,000 account, preferably seven figures. That’s more money than most people have available, including me :frowning:

But at least with the long term you can be a bit more “set it and forget it” than having to stick to the monitor, and as you said, the commissions and slippage are much smaller. I believe its also a little more intellectually stimulating to think fundamentals than technicals, but that’s already a personal preference I think.

Anyway, I’m still a new guy but this is what all my research and reading has led me to believe thusfar :o

The fixed percentage of the account equity formula does sound interesting. I will need to do some more pencil and paper on it and do a comparison.

To achieve that kind of statistical sampling, you might have to pay your broker more than what you earn. You had it right when you said that they must earn their living too. Each time you trade, they earn. Doing the math yourself is eye-opening.

That is true, even in the case of winnings you can give away most of your profit to them in essence. Makes me wonder if I’m on the right side of the ledger :rolleyes:

I’m a trend-follower so my thinking is different. My winning percentage is below 50% but my winning trades bring in, on average, twice as much as my losing trades. Smaller bet sizes spread across many different markets allow me to be in those large moves I hope to benefit from.

Okay, but what do you do when a pair begins to range, and how can you tell?

The amount you pay for commission also depends on your lot size. So, if your broker takes 3 pips per EUR/USD trade and you trade mini lots, you will pay $3. So if your profit is $40, they get $3.

Again on EUR/USD, a 40 pip take profit target and a 25 pip stop-loss would net you $37 if you win and -$28 if you loose.

I see. My problem is that I am trying to manually backtest my strategy, and I run into a confusion as to how to view the ins and outs properly. The price chart on metatrader represents the bid price, from what I understand. I use the cross to see if my entry point would have gotten in and if I would have nailed 40 pips, or would I have gotten stopped out at 25. I am doing this off the bid price (since that is what the chart shows), and I always run into a confusion on how to compensate for the spread, and where is the real in and out point. I am testing USD/CAD, so the spread there is usually 5 pips - no small potato.

Slippage depends on your broker and on market conditions when your order is filled. When using short-term systems, you will find that they are extremely sensitive to slippage and commission charges, enough to make a good system in theory non-profitable in practice.

I suppose it might be a good idea to set limit orders in such cases. If they don’t get filled, they don’t, but at least you don’t have a problem with unpredictable fills…

10

A P.S., I decided to test the variable equity risk rule to a fixed equity risk rule. We are assuming a risk of 2% per trade.

In the fixed rule, you base the 2% risk off the initial account size. For an account of $1000, that would be $20 risk. In the variable equity risk rule, you base each trade on 2% of your existing account.

[B]Fixed account equity, 2% from base of $1000 ($20 per trade).[/B]

12 wrong in a row, starting account size $1000

1.$980
2.$960
3.$920
4.$900
5.$880
6.$860
7.$840
8.$820
9.$800
10.$780
11.$760
12.$740

Total loss: $260

Bounceback to breakeven: 12 trades at a 1:1 risk reward.

[B]2% variable equity rule[/B]

12 wrong in a row, starting account size $1000

1.$980
2.$960.40
3.$941.19
4.$922.36
5.$903.92
6.$855.84
7.$838.72
8.$821.94
9.$805.50
10.$789.39
11.$773.60
12.$758.13

Total loss: $241

Saved using variable equity rule: $19 or 1.9% of account.

Bounceback to breakeven: 14 trades at 1:1 risk reward.

As illustrated above, the fixed risk method gives you just 1.9% more drawdown at 12 losers, but requires 2 fewer successful trades to get back to breakeven (assuming 1:1 risk reward).

I tested the figures for 30 losers in a row. At 30 losers in a row your maximal savings to your account size is 14.5% using the variable equity method. With fixed equity, you be down 60%, with variable you’d be at 45.5% down. To recover, you’d need 30 trades for fixed equity, and just one extra trade, at 31, for variable equity (and you’d even make an extra 6 bucks).

So if I got my numbers correct, it seems the variable equity risk method begins to pay off when you have very big drawdowns of over 50%.

11

I’ve often thought about this question, and I’ve draw several conclusions as to why. From what I’ve learned, it seems that while professionals do need to be conscious of their risk factor, they also have one more big handicap - they trade with size and they have to enter their positions at poorer execution rates. Consequently, they have a tough time adopting short term strategies (unless we’re talking floor traders) because they are inefficient for them. For example, if some big trader has a multi-million dollar position, an immediate fill is not always possible, they have to go through a dealer and those guys have to work the order. Given the fact that a currency can move 100 pips or more within a day, the entry rate can be all over the place. Also, such large orders could encourage the dealer to front-run, which is legal in FX.

I agree. I would add that they (the big guys) in turn get access to much better commission. At the size they’re trading, they get a BIG discount. They also have access to better programmers and computers and that gives them an edge short term that we don’t have.

I think for this reason mainly, people who trade eight and nine figure accounts are happy to get 20%-30% ROI, while small and fast traders can run their account to over 100% in a year. Either that, or the small traders are snails crawling on the edge of a razor. I have heard of some very good short term traders that average 1-2% a day, that is the goal that I’m shooting for.

Absolutely. Under-capitalization is our main problem. This forces us to take bigger risks. But we have no choice if we want to get to that level. This makes it even more important to understand the issue of risk. This is where I find back-testing answers a lot of questions.

From a mathematical basis, it seems to me that daytrading is the quickest means of potential compounding. Also, it will let you know if you’re right or wrong with your strategy pretty fast. If you have a strategy that’s say 50% effective with a 1:2 risk reward, you’ll find out much sooner how really profitable it is as a daytrader, rather than a long term trader.

I can only tell you that I started exactly where you are right now. Some of my mathematical assumptions where proven wrong when I really decided to take up the issue of back-testing. I have programmed trend-following strategies and applied them shorter term time frames (15min - 4h) across many (20+) currency pairs for several (20+ years) and found that, while for short periods of time (1 - 2 years) some outperformed my longer term systems, taken overall it wasn’t even a contest.

Perhaps a different system might work, it’s just that I have yet to find one. If you do, perhaps you’d be so kind to share your results :smiley:

In my view, long term traders can get hurt badly if fundamentals make a sharp reverse, several weeks worth of profits can get suddenly dashed because Bernake says something at a press conference.

My system is designed to handle the volatility you correctly describe.

Also, I don’t imagine how a long term trader who gets 20% annually can live off their trading unless they have a $200,000-500,000 account, preferably seven figures. That’s more money than most people have available, including me :frowning:

You got me there. I’m willing to take bigger risk to get to that level, but I’m not willing to blow out my account in order to do so. The most powerful tool that I’ve discovered that basically allows me to take bigger risk without wiping out my account is diversification. You might look that up and test it for yourself.

But at least with the long term you can be a bit more “set it and forget it” than having to stick to the monitor, and as you said, the commissions and slippage are much smaller.

Yes! And it allows you to keep your day job and keep increasing the size of your account until the day where you no longer need it. You can also rekindle your programming skills with all the time you have, which in turn makes you a better trader (IMHO) :slight_smile:

I believe its also a little more intellectually stimulating to think fundamentals than technicals, but that’s already a personal preference I think.

It may be, but I choose to look for mental stimulation in other venues. Trading to me is a way to make money.

Okay, but what do you do when a pair begins to range, and how can you tell?

There’s no way around it. When a pair begins to range, I loose money. Just like counter-trend systems loose money when it trends. But the key is to be able to withstand these losses to be able to benefit when the market resumes trending. I do this with stop-losses and position-sizing (basically bet small so as not to loose big)

I see. My problem is that I am trying to manually backtest my strategy, and I run into a confusion as to how to view the ins and outs properly. The price chart on metatrader represents the bid price, from what I understand. I use the cross to see if my entry point would have gotten in and if I would have nailed 40 pips, or would I have gotten stopped out at 25. I am doing this off the bid price (since that is what the chart shows), and I always run into a confusion on how to compensate for the spread, and where is the real in and out point. I am testing USD/CAD, so the spread there is usually 5 pips - no small potato.

Yea, it’s tough to back-test on intra-day data because it’s difficult to re-create the exact conditions that would have happened live. Unless you have tick data, I think your results might be off. For example, say you program your system to look at the next 15 minute bar. You have a High anf a Low for that bar, but if your take profit price and your stop loss are within that bar, how do you know which one was hit first? :confused:

12

You’re right, but consider this:

  1. (in theory) with the variable bet size, you never completely wipe out.

  2. To experience the true power of this variable bet concept, you might run a simulation over several years’ worth of data. Assuming your account size grows, you might be in a position in 2 years of having a bet-size double or triple what you would have now, and your account might experience exponential growth.

13

I agree, I was thinking of working the compounding in a slightly different way.

My goal is to compound every 21%. I call that my “compounding period”. This way I have a milestone that I can aim for every two months, and try to do my best to design my systems in order to meet (and hopefully beat) that milestone. If I manage successfully, every year I will have multiplied my account over 3 times, before taxes. Part of my money will be in a regular cash account, so I can live off those earnings. The remainder will be in a Roth IRA, where it can compound tax free over the course of years. I figure in 30 years if I’m any good, I’ll be able to live off of muni bond income and do whatever I want :slight_smile:

What I need to still decide is when to shut things down if I begin a long string of drawdown, and how to re-enter the game.

14

I assume you mean the dealers and traders at the bank. Yes, they do the pip by pip “jabbing” because their spreads are like 0.5 pips. But in the end, they are employees who work by committee and have to answer to management, including risk managers. They make just a small percent of their profits, so their motivation isn’t the same. Also, you have 10 traders and 10 different personalities/strategies. One guy is making a killing one year, the other guy is just barely making it, the third guy is under, etc. It all evens out. I don’t think banks with all their infrastructure can make such a killing as a focused retail guy. I’ve very rarely heard of those guys doing triple figure returns, when that happens it never lasts unless we’re talking Paul Tudor Jones.

Absolutely. Under-capitalization is our main problem. This forces us to take bigger risks. But we have no choice if we want to get to that level. This makes it even more important to understand the issue of risk. This is where I find back-testing answers a lot of questions.

I remember in the book “Millionaire Traders” there was an article about one guy, I think Rob Booker, who says he is a “backtester first, trader second”. He used to always manually backtest everything. His big thing was small but regular monthly profits. It got him from 2,700 to seven figures.

I can only tell you that I started exactly where you are right now. Some of my mathematical assumptions where proven wrong when I really decided to take up the issue of back-testing. I have programmed trend-following strategies and applied them shorter term time frames (15min - 4h) across many (20+) currency pairs for several (20+ years) and found that, while for short periods of time (1 - 2 years) some outperformed my longer term systems, taken overall it wasn’t even a contest.

Perhaps a different system might work, it’s just that I have yet to find one. If you do, perhaps you’d be so kind to share your results :smiley:

:smiley:

My personal feeling is that my system needs to prove its reliability within a given span of time that will permit me to institute regular compounding. I figure if I’m making at least 3-4 trades per day, I have a higher sampling to work from than if I go by 3-4 trades a week. In the latter case, with 10 losers in a row, that’s three weeks of being toasted. If you live off your trading, that’s pretty scary because your bills come in on a monthly cycle. With 3-4 a day, each week you have 15-20 samples to go from. Many short term traders from what I’ve read go by weekly targets, and to me that seems like a fairly safe milestone to shoot for.

My system is designed to handle the volatility you correctly describe.

I would imagine that includes trailing stops?

You got me there. I’m willing to take bigger risk to get to that level, but I’m not willing to blow out my account in order to do so. The most powerful tool that I’ve discovered that basically allows me to take bigger risk without wiping out my account is diversification. You might look that up and test it for yourself.

By diversification do you mean trading unrelated crosses? For example, trading EUR/USD and then GBP/JPY at the same time? One could be in a range, the other trending, so you’d apply the right system to each?

Yes! And it allows you to keep your day job and keep increasing the size of your account until the day where you no longer need it. You can also rekindle your programming skills with all the time you have, which in turn makes you a better trader (IMHO) :slight_smile:

Yes, but I’m willing to stay in front of the monitor for six hours a day. To me its the best job you can have. You can take a day off whenever, etc.

It may be, but I choose to look for mental stimulation in other venues. Trading to me is a way to make money.

I understand. I think everyone finds it interesting in the beginning, but I can see that once you’ve developed your system and its working, it can get a bit on the mundane side.

Yea, it’s tough to back-test on intra-day data because it’s difficult to re-create the exact conditions that would have happened live. Unless you have tick data, I think your results might be off. For example, say you program your system to look at the next 15 minute bar. You have a High anf a Low for that bar, but if your take profit price and your stop loss are within that bar, how do you know which one was hit first? :confused:

Yeah. I am using Metatrader and am just going around with the cross figure. It works for most of the time rather well, but I wish there was an easier way to do it. Have you tried Forextester for manual backtesting? Is it worth the money versus Metatrader?

Thanks again, your input is very interesting.

15

I would imagine that includes trailing stops?

Yes. One that accounts for recent volatility.

By diversification do you mean trading unrelated crosses? For example, trading EUR/USD and then GBP/JPY at the same time? One could be in a range, the other trending, so you’d apply the right system to each?

Yes. I realize this isn’t proper diversification. Ideally, I would love to have enough capital to take advantage of trends in the futures and equities markets as well.

I apply the same system (with slightly different parameters) to each instrument I trade. When I’m loosing money in one market another may be trending (making money). Overall, this reduced my draw-down percentage, allowing me to increase my bet-size.

Yeah. I am using Metatrader and am just going around with the cross figure. It works for most of the time rather well, but I wish there was an easier way to do it. Have you tried Forextester for manual backtesting? Is it worth the money versus Metatrader?

Sorry, I have no idea. I haven’t paid for any software yet. I use the results from my own research; I have a spreadsheet in which I enter rates each day and it spits out my targets and stop-loss info.

Thanks again, your input is very interesting.

You’re welcome.

16

My goal is to compound every 21%. I call that my “compounding period”. This way I have a milestone that I can aim for every two months, and try to do my best to design my systems in order to meet (and hopefully beat) that milestone. If I manage successfully, every year I will have multiplied my account over 3 times, before taxes. Part of my money will be in a regular cash account, so I can live off those earnings. The remainder will be in a Roth IRA, where it can compound tax free over the course of years. I figure in 30 years if I’m any good, I’ll be able to live off of muni bond income and do whatever I want :slight_smile:

Kind of a hybrid method. I like it.

What I need to still decide is when to shut things down if I begin a long string of drawdown, and how to re-enter the game.

One of the main benefits I’ve found in back-testing is that, done properly, it gives me an idea of what kind of drawdowns I’m likely to experience, given certain parameters. I can then adjust these parameters to keep the max drawdown at an acceptable level.

The way I’m trading now, I’m willing to see 50% drawdown before I shut it down. This leaves me something to get going again should I choose to later.

17

The backtesting is indeed very important as I am increasingly beginning to see. I think its amazing that you can trade in the past and practice that way, understanding all the potential possibilities, but few people seem to want to bother. Come to think of it, you can even do that on a Saturday, when the markets are closed…

The basic core of my trading philosophy seems to be coming together as follows:

1.I must always work to maintain an edge over the market, be focused, pay close attention to detail, and be nimble to any changes.
2.I must be a fanatical money manager and statistician.
3.I must constantly be developing systems that will play the market in any phase, range or trend (bullish/bearish), volatile or flat, and learn to apply the right system to the right pair at the right time.
4. I must understand the risk going in, and execute ruthlessly with the understanding that the statistics will play out. If they do not, I must discern this at a predefined point and take a corrective course of action.

I think that a good part of my time, maybe even the majority, should be spent backtesting and analyzing the markets. That will help me get to know it better than anything else.

I am working on my money management system right now. It will be interesting to see how it plays out with strategy development, I want the two to go hand in hand.

At the moment, I am trying to see if I can develop a 1:2 risk reward strategy with a 66% success rate, a decent spread seems to be 40 risk, 80 reward. I am shooting for a maximum of 15 setups a week (I’ll probably need to develop several systems for that to happen). That’s a good compromise I think, versus the 25 pip risk 40 pip reward scalp. The smaller you get, the more risk slippage and commissions are, as well as stop gunning. The bigger you get, the more risk of fundamental surprises.

Btw, I imagine its not possible to program in Fibonacci retracements, is it?

18

I would imagine that it might be difficult to describe in code what segment of data to look at.

The issue with trading retracements isn’t only when to get in, but when to get out of the trend. Anybody can look at a chart and say yea I could’ve gotten in right there (at the bottom of the retracement), but when would they have gotten out of the preceding move? In real time, not when looking at past prices.

Over the long hall, it’s difficult to get out ‘on time’ and re-enter at a better price consistently.