 |
Schaefer
Registered: Oct 2004
Posts: 511 |
02-05-12 11:58 PM
One of the best threads ever, regarding entry/stops. Amazing generosity from some of the profitable traders...must have been the happy hour...
Schaefer
|
| |
|
Edit/Delete • Quote • Complain |
Wide Tailz
Registered: Sep 2011
Posts: 1515 |
02-06-12 12:21 AM
Quote from FreakofNature:
when this kind of event happens, this is when we should add to winners right ?
Yes, or if you see your buy signal on the next higher time frame. If trading daily, look to the hourly for another buy signal for adding. Works for scaling in, too.
Entry setup is all about minimizing risk. Low risk opportunity is the edge.........
|
| |
|
Edit/Delete • Quote • Complain |
Wide Tailz
Registered: Sep 2011
Posts: 1515 |
02-06-12 12:25 AM
Quote from tenthousandmen:
IMO -
as far as 100% automated systems go (not discretionary or support/resistance traditional trading)..........
Get away from the notion of a stop being your safety net, and use a filter that signals when a system should make a trade based on market conditions (such as volatility). Then use the stop as your ultimate safety net for Bad scenarios.
Also, for systems that incorporate a stop by having algo's dedicated to stop calculation, can permit much smaller stops. Just keep the same system building skills of processing data in your head and avoiding curve fits.
One way I judge a system is how close I can set a stop before it impedes profits. One of my best systemz can run a 3.5% stop, another one required 15%.
Funny thing was, the one that could run the tight stop just blew away everything else I have ever seen for total profits.
Signal to noise.
|
| |
|
Edit/Delete • Quote • Complain |
dom993
Registered: Jul 2008
Posts: 537 |
02-06-12 12:40 AM
Quote from jnbadger:
Ok, I'll bite. How the hell do you know what your worst losing streak will be ahead of time?
Backtesting (even better your own trading history for that setup) is 1/2 of the answer ... with a large enough sample size (say 500+), you start getting an idea re. what kind of losing streak can be thrown at you. My coach routinely says "take the max drawdown from backtesting and multiply it by 2".
Another 1/4 of the answer is MonteCarlo simulations (using backtesting or historical trade distribution) ... for these to be useful though, you'll have to decide for yourself how far (number of standard deviations) from the mean you want to set you limit for max drawdown.
And I am still looking for the last 1/4 of the answer 
|
| |
|
Edit/Delete • Quote • Complain |
dom993
Registered: Jul 2008
Posts: 537 |
02-06-12 12:51 AM
Quote from N54_Fan:
I think you are either not choosing your words correctly to get your point across or you are taking inappropriate risks. You should ALWAYS adjust your trading size (position size) based on your account and stop placement. This is Risk Management 101.
For example if I have a $100K acct and I'm buying a stock at $100 with $2 stop loss (@$98) then assuming I want to risk 1% of my capital I can buy 500 shares ($1000 capital to risk/$2 per share = #shares to purchase = 500 shares). The further the stop loss is away from my purchase price the less shares I buy but keep my risk the same...if the stop was at $97 then I could buy 1000/3 = 333 shares MAX.
Yes that was a little provocative - and I appreciate you jumping on it.
My point is that "assuming I want to risk 1% of my capital (on a single trade)" is a pretty poor position sizing decision, which doesn't take into account the win% of your system nor its payout ratio (average win / average loss) nor any other consideration on its statistical performance. Its a kind of one size fits all risk management, may-be a reasonable start in some cases, but far from allowing one to maximize the return of a given edge & account size.
|
| |
|
Edit/Delete • Quote • Complain |
N54_Fan
Registered: Sep 2011
Posts: 388 |
02-06-12 01:11 AM
Quote from dom993:
Backtesting (even better your own trading history for that setup) is 1/2 of the answer ... with a large enough sample size (say 500+), you start getting an idea re. what kind of losing streak can be thrown at you. My coach routinely says "take the max drawdown from backtesting and multiply it by 2".
Another 1/4 of the answer is MonteCarlo simulations (using backtesting or historical trade distribution) ... for these to be useful though, you'll have to decide for yourself how far (number of standard deviations) from the mean you want to set you limit for max drawdown.
And I am still looking for the last 1/4 of the answer
Actually, MC sim should give you almost everything you want to know. If you have a win rate, profit factor, expectancy, and payoff ratio you can run a MC sim to give yiu the ideal % to risk per trade for this system. This is done by deciding what is an acceptable max drawdown and what would be considered "ruin" by the trader trading the system. For me its a 15% max draw down and 35% loss I would consider "ruin". You then run a MC sim and determine the amount of capital per trade to risk that would give a 1% chance of having this kind of DD or ruin. (some may choose a 95% confidence ...ie 5% chance of this happening). So by playing with various amounts of risk you can determine the "best" percent of capital to risk per trade to achieve your goals with minimal risk.
There is one caveat with this method...you are assuming that your system will perform the same in the future as it did when you collected the data. So you must be carefull to trade the system in the same type of markets.
Good luck
N54_Fan
|
| |
|
Edit/Delete • Quote • Complain |
| Receive
an email whenever a new post is added to this thread by subscribing
to it. |
|
|
|
|
