how to improve my system

[mateusz]

hi everyone, I use multiple screen system (5 screen) to trade mostly forex. I get about 50 points a week on one lot and I know it is not bad but... There is so much to do with my system as I look at it but I stuck. I dont know where to start the reasearch further. Some words of explanation on my system:
I have five sreens with different time periods, mostly 5min, 15min, 30min and 60min. Also I use 1min chart as a trigger. On each of the sreen I have the same Stochastic (8,6,3) and moving averages like 40, 20, 9 and on 5min chart also 200 lenght. I have also my home made channels on 5 min chart. Well, the general rule is that I do open position when all of the screens (stochastics) are in the same direction I mean the have the same signals. I watch moving averages to determine some important levels and I try to watch volume close enough and watch if the volume is rising in the direction of my stochastics.
How could I improve that kind of system? I also would like to know what you think about this kind of strategy. Its simple but effective but I have feeling that it could be much better.
Thanks.

PS: If anyone need more details then I would be glad to answer questions.

[DuaneG]

Hi Mat,

I know many traders that have been an advocates of multiple time frame charting. Borrowing from Jack Schwager, I like to think of multiple time frame charting as using a telescope. By being able to focus on both the long term direction as well as the short term entry, this method has allowed me to keep all pertinent indicators in plain view to prevent being blindsided by ignoring a specific time frame. I think you are on the right track in that regard.

What I have found interesting, is slightly altering the moving averages on shorter term time frames as I have found them more accurate in the sense of timing. I don't use the same moving averages on a short term time frame that I do on a longer term time frame. By assigning weights or priorities to these indicators, I can get a clearer sense of when to pull the trigger and when a given move has proven it's validity.

[mateusz]

thanks for your replay Duane,
I will try to "play" with different moving averages as triggers on different timeframes. As I undarstood your massage there should be found most effective MA lenght for triggering, wright? There will be always a kind of compromise, the thing is to find the best one.
I know that this is the wright way to go, I feel comfortable with these rules and the "only" thing that left is to get it as good as it can be - and avoid crushing the system by wanting to make it perfect.
Word about multiple timeframe: it is amazing how clear is the market when you look from the wide and zoomed perspective in the same time. No one can afford to ignore the trend in long term as well as the very present details which when you put together make that trend.

[mateusz]

I would be very pleased if you could share some more details about moving averages on multi timeframe charting. For example do you think about some special proportions of changing lenghts of MA on different timeframes?

[soylent]

Mat --

If you are reliably making 50pts a week w/o big drawdowns, I wouldn't change a thing.

10pts per day = more than enough to live on.

You've found a way to keep food on the table! Maybe add a second contract. Or, third.

My only suggestion to optimize:

If you're not already, you might try:
TARGETS: Backtest your system if you were to set a target on each trade. Does it work any better? Maybe set 1 contract for +10ticks, 1 for +15, 1 for 'running' as high as it can? On choppy days, I'll set to +3 ticks, +5, +7, +10 on my contracts....

-c

[mateusz]

thanks for advice, the problem is that I dont know how to backtest multiple screen system in esignal or metastock. I could back test each screen separatly and put the scores to the text file then combine all the numbers in excel for example but when I tried this something was going wrong and I had some stupid numbers after all.
This is I think the true disadvantege of multi chart system (well not itself but maybe the software) - backtesting. Thats why it took me few months to get it work properly - the automatic optimizations of every timeframe separatly doesnt work when you want to combine them later. Mostly I have done this manualy.
The biggest moves I catch per trade are about 50 points. Mostly the signals give me about 20 points. When I am wrong it costs me from 10 to max 30 points (20 pts of loss is the level when I usually get the out of market signal).
I have never added contracts. I would have to put it somehow into another rule. Your idea could be ok but I have to check it - manually.
Thanks very much.

[mateusz]

one more thing which I didnt mantion and which is very important in this: I trade with this system only if the market has a proper volume (for example if volume is greater than its MA 100 lenght on 5 min chart). I wait until the market gets to its normal rythm. It also showed me some special trading hours. I dont trade before important data and just after they are released - even if I have a signal.

[soylent]

Originally posted by mateusz
When I am wrong it costs me from 10 to max 30 points (20 pts of loss is the level when I usually get the out of market signal).

Woah! Those are big drawdowns partner!

The ONLY time something like that has ever happened to me... Well, never... I've had a few where the market took my stop and I lose 4-5 ticks before my fill. (I really dislike getting taken out on a stop -- ugh!)

Anyhow, I'm not sure how you're doing things.

An example:
I do trade options. On those I -will- take bigger drawdowns because I'm playing on the 13minute chart and I will take a 10% 'drawdown' of the amount invested before I call uncle. The upside is usually 20-40% target. So, to me, I can be 'right' 1 in 3 times. Usually I'm right 3 out of 5 times. So, it works pretty well... But, still, I'm looking for 30% ROI in one 'intraday' option trade. And, I don't even bother w/ the 1/3/5min timeframes for those trades.

For FUTURES:
I won't take more than a 3-4-tick drawdown. Why? Because my plan says that if I'm not right RIGHT NOW then I should get out quickly. (I trade the 3-min timeframe only)

So, I don't know if you are playing options but I would concentrate on cutting your losses more quickly... Futures, the 'get out now' plug is pretty painless (small spreads). Whereas, options you have a bigger spread to deal w/. So, for futures, I jump out fast.

As for backtesting: Yes, some of it is hard to automate....

[soylent]

Originally posted by mateusz
one more thing which I didnt mantion and which is very important in this: I trade with this system only if the market has a proper volume (for example if volume is greater than its MA 100 lenght on 5 min chart). I wait until the market gets to its normal rythm. It also showed me some special trading hours. I dont trade before important data and just after they are released - even if I have a signal.

I agree w/ you on that!

I won't be in a trade infront of news (for an intraday trade) and I don't day-trade from 11:50-1:30. I don't even look @ the screen because I get seduced into stupid trades.

Though, lately, some buggers have been tanking the market @ 12:15-12:45. so, I do pay a little attention to see if they're going to sell the market off over lunch...

-c

[mateusz]

yes, I agree that this kind of drawdown is big but I am wright about 7 times on 10 trades and usually I lose about 10 points in the beginning of my trade. This is the problem I need to optimize better. Ofcourse I like it when after I open position the market moves imediatly in my direction but this doesnt happen to often.
Well the calculation is very simple.
For one contract(it is called "lot" on forex) of EUR/USD (eur a0-fx) I need 2000$, for two contracts 4000$ and so on (this is part of my system of money managment). One point is 5$ and I let to lose maximum 100 points, then I close contract and wait until I rebuild my position with the other. 100 points is 500$, when my average loss is about 20 points then I would have to catch 5 drawdowns in a row. This never happened to my system - once I had 2 drawdowns in a row and then 70 points profit. The stops in this case are in my opinion ok because the market looks to be a wide one.
Well like I said these are the things that I need to optimize.
Thanks very much.

[soylent]

Here's an idea:

For EURO, try trading @ the tick-chart level to time your entry.

Try the 133T -- see if that gets your stop loss/entry timing a little better?

-c

[mateusz]

ok, thanks, Ill try to check it out and see if it works better. The thing is I dont get many signals a day (usually 2 - max 3 signals) and I think that it is the power of that system - it filters many wrong enters.
As I count it (your way): you have avr. 5-6 losses on 10 trades. But because of very low losses you get what you want after all. I have 6-7 profit signals but larger losses - as I count, it gets very similiar to your numbers (summ your losses and my losses - I think it will be quaite the same). Its all about the strategy efficiency. If I would trade with your strategy and my stops I would probably get wiped out very soon.
I know how important is money managment but still the interpretation of market is most interesting and profitable thing to me. Maybe that is the comfort we are looking for in creating our own style of trading.

[mateusz]

ok, there it is... sell signal from today`s morning on eur/usd future. There are some important issues with volume which you dont see on that picture but generally this is how it works. Simple and clear. This signal gave me 48 points. There was another good signal at about 14:00 but at 14:00 there were some data to be released and I didnt take it.

[mateusz]

sorry for the quality of that image but I couldnt make it better to fit in the limit of 100kb.

[soylent]

Mat --

Here's my style...

One indicator, One timeframe.

Today was 65 points on YM (Dow).

Trades were 3 for 3 (100%).

There are some trades I take that are lower percentage (50%ish) but they give better / faster returns. Risk is 3-4 ticks, vs reward of 50-60ticks. Today being a nice trend day on YM, all the TLB/ZLR's (trend line break / zero line rejects) worked out really well. I also had an excellent counter trend @ 3:15 on AB (ER2).

I've been using CCI for about 2 months. Learned (and, STILL learning) @ 'woodies' and my success rate has gone from abysmal to rarely a losing day...

-c