khalid,
Agreed Monte Carlo analysis is ONLY a simulation. It isn't reality. However, that simulation can be a good guideline for helping you predict where your brokerage statements will end up.
Additionally, fat tails do occur in the market and they can mean that your estimates based on past trades may or may not help you predict future trading.
Edward
Buy Zone on Hour Bars?
Moderator: moderators
Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.
Thank you for your support.

 rank: 50+ posts
 Posts: 64
 Joined: Mon Feb 11, 2008 2:57 am
 Reputation: 0
 Gender:
Edward,
Thanks for the monte carlo ideas. I'm going to create my own monte carlo software. It'll be simpler for me than using Excel. Excel is a little awkward for doing this kind of thing.
FYI, there must be an error in the calculations of your monte carlo software.
Why? Because it's impossible to completely lose an entire account with 5% or even 10% risked each time.
Why do I say that?
Think mathematically. If you divide a number by 2 and do it again and again, you can never get to ZERO. You can do that to infinity and never completely reach zero.
That's an exaggerated example.
But if you remove only 5% each time, 100 consecutive losses still leaves you with $6.00 in your account.
Agreed! I would consider that wiped out. But how could you randomize 100 80% profitable trades and show even 2% of 10,000 wiping out the account?
I'm guessing it is take out an 5% of the original balance each time rather than 5% loss on the current balance.
If you do it that way, it only take 20 losses to reach ZERO.
I'm going to write my own this week using C# and let you know the results.
Wayne
At 5% losses, it takes 44 consecutive loses to get to 100.
Thanks for the monte carlo ideas. I'm going to create my own monte carlo software. It'll be simpler for me than using Excel. Excel is a little awkward for doing this kind of thing.
FYI, there must be an error in the calculations of your monte carlo software.
Why? Because it's impossible to completely lose an entire account with 5% or even 10% risked each time.
Why do I say that?
Think mathematically. If you divide a number by 2 and do it again and again, you can never get to ZERO. You can do that to infinity and never completely reach zero.
That's an exaggerated example.
But if you remove only 5% each time, 100 consecutive losses still leaves you with $6.00 in your account.
Agreed! I would consider that wiped out. But how could you randomize 100 80% profitable trades and show even 2% of 10,000 wiping out the account?
I'm guessing it is take out an 5% of the original balance each time rather than 5% loss on the current balance.
If you do it that way, it only take 20 losses to reach ZERO.
I'm going to write my own this week using C# and let you know the results.
Wayne
At 5% losses, it takes 44 consecutive loses to get to 100.
 eudamonia
 rank: 500+ posts
 Posts: 536
 Joined: Thu Jun 15, 2006 9:50 pm
 Reputation: 0
 Location: Rocklin, CA
 Real name: Edward Heming
 Gender:
 Contact:
Wayne,
When you can no longer make a margin call the software considers you blown out. So if you have $6 left in your account you can't trade (with most brokers).
Edward
When you can no longer make a margin call the software considers you blown out. So if you have $6 left in your account you can't trade (with most brokers).
Edward
Eudaimonia (pron.: youdiemoanea) (Greek: εὐδαιμονία) is a classical Greek word commonly translated as 'happiness'. The less subjective "human flourishing" is often preferred as a translation.

 rank: 50+ posts
 Posts: 64
 Joined: Mon Feb 11, 2008 2:57 am
 Reputation: 0
 Gender:
With MB Trading you can trade with only $2. I know if for a fact because I accidently tried which testing an automated strategy the other day.
I thought NinjaTrader wanted the number of "lots" as 2. But it actually took it in MBT as $2 of current bought and sold.
MBT bases it's commission at 50 cents for each $10,000, per side.
So guess what, after making a few test trades this way, it shows the trades on MBT Navigator but showed commission as ZERO.
That's what got me curious and realized it was only trading $2 of USD/JPY instead of $20,000 USD/JPY.
So the round turn commission on $2 is only $0.0001 or 1 one thousandth of a cent. So it shows up as $0.00 as commission charged!
Wild! I never thought it could go that small.
While MBT allows you to trade that small, it not worth it because making 10 pips would be be 1/100th of a penny profit at that amount.
So while there's not a margin limit like options, I'm interested in using a 25% draw down when I use my monte carlo runs.
Sincerely,
Wayne
I thought NinjaTrader wanted the number of "lots" as 2. But it actually took it in MBT as $2 of current bought and sold.
MBT bases it's commission at 50 cents for each $10,000, per side.
So guess what, after making a few test trades this way, it shows the trades on MBT Navigator but showed commission as ZERO.
That's what got me curious and realized it was only trading $2 of USD/JPY instead of $20,000 USD/JPY.
So the round turn commission on $2 is only $0.0001 or 1 one thousandth of a cent. So it shows up as $0.00 as commission charged!
Wild! I never thought it could go that small.
While MBT allows you to trade that small, it not worth it because making 10 pips would be be 1/100th of a penny profit at that amount.
So while there's not a margin limit like options, I'm interested in using a 25% draw down when I use my monte carlo runs.
Sincerely,
Wayne
Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.
Thank you for your support.

 rank: 50+ posts
 Posts: 110
 Joined: Wed Mar 14, 2007 2:15 pm
 Reputation: 0
 Location: Lahore
 Gender:
I MUST apologise for being so blunt!
Wayne,
Please read Edward here
http://kreslik.com/forums/viewtopic.php?p=10362#10362
And, please note the risk the professionals take.
God speed.
Khalid
Wayne,
Please read Edward here
http://kreslik.com/forums/viewtopic.php?p=10362#10362
And, please note the risk the professionals take.
God speed.
Khalid
greaterreturn wrote:Khalid,
Thanks very much. How much do you recommend risking? Here's my thought. When I had $100,000 in my account, I would probably only risk 1% per trade since that would be $1000. And if the trade wins, it could make $1,000 to $2,000.
But with an account of $1,000 it seems I need to risk 5% which is only $50.00 to make anything significant. 1% would be only $10.00 and a potential reward of only $10 to $20. That's doesn't seem worth my time.
Does that make sense?
FYI, I did some analysis this weekend and found that if I use the TRO Dynamic SR indicator along with the Buy Zone, it works better.
Rather than 60% win/loss, it goes to almost 80%. Plus the number of pips almost triples. Why? It's because the SR shows whether the buy zone falls in the middle of a range and therefore more risky or else at the bottom or top and ripe for a breakout or retracement.
Do any of you use filters to guide you with the Buy Zone?
This week I'm going to forward test with 5 pip SL and 21 pip TP, one trade per morning of the first or second 10 minute bar at 8 am.
Looking back over March and April, it appease I can get 20 pips at least twice a week with either scratch or 5 pip losses on other days.
That means that with 4 to 1 risk to reward, less risk can still provide decent profit according to spreadsheet projections.
Wayne
Thats fine so you will be putting on $10 each bet with a 5 pip stop (risking $50) and aiming for 20 pips ($200). I have to point out your %winners will drop significantly (compared to 1:1 or 2:1). I guess you might get 40% winners with 4:1 straight off the bat, can you do that on a simulator? Its tougher for real As Edward pointed out (I think) that will make for a far more volatile equity curve, with a small account you'd be much better off going for 1:1 with a higher percentage winners.
Let us know how it goes.
 eudamonia
 rank: 500+ posts
 Posts: 536
 Joined: Thu Jun 15, 2006 9:50 pm
 Reputation: 0
 Location: Rocklin, CA
 Real name: Edward Heming
 Gender:
 Contact:
Blowfish,
You are dead on about having your winning % drop when looking for a 2:1 or higher risk to reward ratio. For myself I see about a 3:1 R to R and my win ratio is about 42%.
However, you are committing what I call the "want to be right" error in saying that your equity curve is going to be better with a 1:1 risk to reward ratio. Yes your equity curve will be smoother, however, your expectancy will be lower and your profits lower. So is your equity curve actually "better"?
Wayne,
Although it is fascinating that you can trade very small size, there is still a minimum account size with MBT ($200 or something like that). I agree with Khalid that you are living in a bit of a fantasy land. Consider this fact. If you suffer a 25% drawdown you need a 33% gain to recover. If you suffer a 50% drawdown you need a 100% gain to recover. However, if you suffer a 90% drawdown you need a 900% gain to recover.
So if you suffer a 90% drawdown or greater your odds of recovery are miniscule. With a 1:1 risk to reward ratio you would need to win 180 times in a row to recover. Does that seem likely even with an 80% win ratio? Not to mention the physic blow it would do to any trader to sustain such a loss in their account. Most traders would puke up their positions after a 50% drawdown.
I think you are suffering a bit from the delusions caused by small accountitis. Because you are only trading with what you consider a nominal account the gains do not satisfy you. So to compensate for that you have cranked up the risk. I guarantee that if you were trading a $1 mil account you would not trade at 5% risk because the dollar gains/losses would scare you. Obviously like I said you are welcome to trade any way you like. However, realize that you will likely blow your account (small that it may be). As Khalid pointed out this is not very professional  but then again most small traders are gamblers anyway.
Best of luck.
Edward
You are dead on about having your winning % drop when looking for a 2:1 or higher risk to reward ratio. For myself I see about a 3:1 R to R and my win ratio is about 42%.
However, you are committing what I call the "want to be right" error in saying that your equity curve is going to be better with a 1:1 risk to reward ratio. Yes your equity curve will be smoother, however, your expectancy will be lower and your profits lower. So is your equity curve actually "better"?
Wayne,
Although it is fascinating that you can trade very small size, there is still a minimum account size with MBT ($200 or something like that). I agree with Khalid that you are living in a bit of a fantasy land. Consider this fact. If you suffer a 25% drawdown you need a 33% gain to recover. If you suffer a 50% drawdown you need a 100% gain to recover. However, if you suffer a 90% drawdown you need a 900% gain to recover.
So if you suffer a 90% drawdown or greater your odds of recovery are miniscule. With a 1:1 risk to reward ratio you would need to win 180 times in a row to recover. Does that seem likely even with an 80% win ratio? Not to mention the physic blow it would do to any trader to sustain such a loss in their account. Most traders would puke up their positions after a 50% drawdown.
I think you are suffering a bit from the delusions caused by small accountitis. Because you are only trading with what you consider a nominal account the gains do not satisfy you. So to compensate for that you have cranked up the risk. I guarantee that if you were trading a $1 mil account you would not trade at 5% risk because the dollar gains/losses would scare you. Obviously like I said you are welcome to trade any way you like. However, realize that you will likely blow your account (small that it may be). As Khalid pointed out this is not very professional  but then again most small traders are gamblers anyway.
Best of luck.
Edward
Eudaimonia (pron.: youdiemoanea) (Greek: εὐδαιμονία) is a classical Greek word commonly translated as 'happiness'. The less subjective "human flourishing" is often preferred as a translation.

 rank: 50+ posts
 Posts: 64
 Joined: Mon Feb 11, 2008 2:57 am
 Reputation: 0
 Gender:
You guys are right. I've been running statistical probabilities using 10 minute bars. I made a chart in excel of the market going 1, 2, to 10 pips higher or lower before the end of the bar when entering at a certain prices levels like the previous bars high, low, or middle.
Amazingly, they all have the virtually the same odds in both directions.
And I've been doing some testing with NinjaTrader to verify the results.
In other words, the odds of hitting a 4 pip TP are the same as a 4 pip SL if you enter when the market hits the high or low of the previous day [edit: should say "bar"].
It doesn't matter which direction you take.
If you raise increase the TP or SL the odds decrease. So a close SL has higher odds of getting hit than a higher TP. The reverse is also true.
So the numbers work out that if you have a slightly higher TP you make more money but it's killed by commission.
An equal TP and SL lose money as well as a higher SL than the TP.
I tried the idea of eliminating the TP and exiting at the end of the 10 minute bar.
That makes slightly more than commission simply because it lets the profits run on the 10 % or less of days [edit: bars] that have large bars. But it's not enough money to be worth trading it.
After studying the charts and all the losing trades this morning as well as statistics on entry and exit efficiency it gave me an idea.
There's a significant number of profitable trades that reverse till the bar closes at a loss. Setting up a breakeven + 1 pip to cover commission at the right trigger point will increase odds of profitability.
In Excel calculations show that 20% of the bars reach 5 pips profit and reverse to a loss. Converting 20% of the losing trades to scratch trades makes a big difference. I have yet to test that on NinjaTrader.
I tried using a simple pipbypip stop loss built into NinjaTrade at 5 pips but it makes the odds worse since it increases to soon.
I'm going to write my own that only sets the break even when it hits a 5 pip trigger.
Another thing, there are 25% of trades that lose more than 9 pips (25% for 10 pips) from their highest "favorable excursion".
So it appears that after the trade reaches 9 pips, I can add an 8 pip trailing stop.
Then when it reaches 13 pips (14% of the time), I can add a 4 pip trailing stop.
For my statistics, I'm using 10 minute bars exclusively from the 8 am to 11:59 pm EST time frame on USD/JPY. I have 1000 bars so it's statistically accurate.
I have an automated trading system running this strategy (without the break even or trailing stops) against a demo account.
Guess what! This morning just after realizing this stop plan above, the auto strategy went short and immediately to over 15 pips profit but reversed and hit the stop loss before the 10 minute bar ended!!!!!
Hey, whatever the statistics say, we have to keep that kind of runup from turning into a loss just keep our sanity!!
In conclusion, it appears to be true what the experts say that it isn't the entry that matters, is the EXIT!
I love this site, it was Michal who started me thinking probabilities, looking at scatter charts, and graphs of market data. I like 10 minute bars better than hour bars. Hour bars offer too few trades for an automated system.
Once I get it profitable, I'll run it simultaneously against crossing the high of the previous bar and crossing the low. That will certainly offer plenty of trades per day.
Please if you have any other advice about this chime in!
 Wayne
Amazingly, they all have the virtually the same odds in both directions.
And I've been doing some testing with NinjaTrader to verify the results.
In other words, the odds of hitting a 4 pip TP are the same as a 4 pip SL if you enter when the market hits the high or low of the previous day [edit: should say "bar"].
It doesn't matter which direction you take.
If you raise increase the TP or SL the odds decrease. So a close SL has higher odds of getting hit than a higher TP. The reverse is also true.
So the numbers work out that if you have a slightly higher TP you make more money but it's killed by commission.
An equal TP and SL lose money as well as a higher SL than the TP.
I tried the idea of eliminating the TP and exiting at the end of the 10 minute bar.
That makes slightly more than commission simply because it lets the profits run on the 10 % or less of days [edit: bars] that have large bars. But it's not enough money to be worth trading it.
After studying the charts and all the losing trades this morning as well as statistics on entry and exit efficiency it gave me an idea.
There's a significant number of profitable trades that reverse till the bar closes at a loss. Setting up a breakeven + 1 pip to cover commission at the right trigger point will increase odds of profitability.
In Excel calculations show that 20% of the bars reach 5 pips profit and reverse to a loss. Converting 20% of the losing trades to scratch trades makes a big difference. I have yet to test that on NinjaTrader.
I tried using a simple pipbypip stop loss built into NinjaTrade at 5 pips but it makes the odds worse since it increases to soon.
I'm going to write my own that only sets the break even when it hits a 5 pip trigger.
Another thing, there are 25% of trades that lose more than 9 pips (25% for 10 pips) from their highest "favorable excursion".
So it appears that after the trade reaches 9 pips, I can add an 8 pip trailing stop.
Then when it reaches 13 pips (14% of the time), I can add a 4 pip trailing stop.
For my statistics, I'm using 10 minute bars exclusively from the 8 am to 11:59 pm EST time frame on USD/JPY. I have 1000 bars so it's statistically accurate.
I have an automated trading system running this strategy (without the break even or trailing stops) against a demo account.
Guess what! This morning just after realizing this stop plan above, the auto strategy went short and immediately to over 15 pips profit but reversed and hit the stop loss before the 10 minute bar ended!!!!!
Hey, whatever the statistics say, we have to keep that kind of runup from turning into a loss just keep our sanity!!
In conclusion, it appears to be true what the experts say that it isn't the entry that matters, is the EXIT!
I love this site, it was Michal who started me thinking probabilities, looking at scatter charts, and graphs of market data. I like 10 minute bars better than hour bars. Hour bars offer too few trades for an automated system.
Once I get it profitable, I'll run it simultaneously against crossing the high of the previous bar and crossing the low. That will certainly offer plenty of trades per day.
Please if you have any other advice about this chime in!
 Wayne
Please add www.kreslik.com to your ad blocker white list.
Thank you for your support.
Thank you for your support.