If your using EFS, enter the highest return you've had...

[BakedWafer]

Originally posted by shogun
how much have you allowed for slippage and comission per trade, as this is often overlooked.

Everything has been accounted for.

Regards

[Doji3333]

Baked....

In response to your question - your system looks good. Without knowing more, all I can say is TEST TEST TEST (speaking from experience).

The strength of any system is the ability to contain losses - called "risk management". Good entry systems are critical, but "risk management" will make or break a system.

I mentioned the system for the Hedge Fund I create - well there are actually 3 of them (for varying degrees of agressiveness). Here is the most agressive system (a test for 5 days) I will run a longer test later, but these clients are not staying in stocks for more than 2~5 days.

Now remember, these guys are buying on margin with huge buying power - so if you work out he numbers on your own, you can figure out the monthly ROI. I estimate at least 20% monthly.

This test shows 3000 shares per trade.

B

[derekg]

Here's a copy of my curve over 120 days. It's not as smooth as
fibgann's or as profitable as doji3333's, but these guys are pros.
EFS is still new to me.


Initial Capital $50,000
Commission $0.02

[BakedWafer]

DerekG, what are you trading?

[derekg]

Originally posted by BakedWafer
DerekG, what are you trading?

Just equities. It doesn't really matter which stock, but some sectors work better than others.

[BakedWafer]

Wanna swap EFS's, mine also has IB integrated.

Let me know.

[Doji3333]

New System...

Folks,

Here is a graph of a new system that I developed over the past 4 days.. This is still very much a proto-type, but it is something that I might consider marketing.

This is still a "work in progress", but some of you have asked if I had anything available for the "public". This is something I might make available....

This is a 25 day test on the ES M4 chart. The results look impressive. This system's results are based on trading 4 contracts.

Take a look...

B

[Doji3333]

Here is the stats...

Backtest results....

B

[atlas]

Perhaps I am missing something here:

$17726 / 4 contracts = $4431 / contract

$4431 / 250 trades = $17 / trade / contract

$17 / trade / contract = .35 ES points / trade / contract

After slippage (unless you are using limit orders), I don't see how this can win? Or are you scaling in trades?

Thanks,

atlas

[Doji3333]

Atlas...

Commissions are already accounted for (@4.80 per trade with IB).

I'm using limit orders for all the trades - so slippage should be minimal (if any).

Even if you allocated $5 per trade slippage (per contract) - let's say that came to $1050...

So, $4331 - 1050 = $3281 - trading a single contract...

That is about 130% per month ROI - would that work for you??

B

[atlas]

Ah, limit orders. Yes, that makes a difference. How do you account for real-life expectations of order entry? If there is one tick higher than the limit amount (sale), is that accepted as possible?

[TraderJO]

Brad,

How do you get the backtester to assign $50 to every ES point?
I have to multiply all my results by 50 to get actual dollar profit values.
___
JO

[Doji3333]

TraderJO

You simply assing the default # of shares to 50 when you run the BT.

or within your code, you create the allotments of shares to use... like this..

var DefEntryContracts = 300;
var Exit1Contracts = 200;
var Exit2Contracts = 100;

this way, the BT does the calculations for you.

B

[augustus]

Hi Brad,

i am trading since a long time an intraday future system based on three timeframe each with about three or four indicators/price/statistic relations. It is difficult to code because the last decision is still a result of my estimation of the overall sentiment, in which market phase we are, what the reaction on the last support and resist was and so on. I still believe that I localize support/resist and marketphase better and faster than I ever could programm it into an automatic system for this futures, but I’m planning to integrate some details in an longer-term system for stocks.

Without going into details of an specific system I hope you can answer me some general questions from your experience:

1. How many inputs do good systems use, what is a reasonable upper limit? Is smaller better to avoid curvefitting?
2. How many timeframes are normally used? Do most system use a larger timeframe to estimate the overall situation and a smaller to determine the entries/exits?
3. Is it better to construct a system which works only at specific conditions (trend, range trading, fast market) or are systems who are always in the market preferable? Are statistic results (day of week/time of day, hit rate of candlesticks or formations, average hit rate of reaching high probability zones/pivot zones) integrated in some systems coded with efs?
4. Generally, do price based systems perform better than indicator based?
5. Most systems I saw are still based on standard indicators. Is the secret still the right composite and good money management rules or use for example the system you programmed for the hedge fund some indicators which could not found with some modifications/ enhancements here on the webpages/books and magazines?
6. How often is there a need to adjust good/poor systems or use good systems always adaptive variables to fit different market conditions? Or have all short term systems an expiration date?
7. What is the best ratio to evaluate a systems performance? What is the best rule for position sizes (every time same size, fixed fractional, kelly f, vola based ...)?

Regards

[Doji3333]

Replies to your questions..

I answered your questions in a WORD DOC. That way they are clear and understandable

The file is attached.

B