|
PureBytes Links
Trading Reference Links
|
>My trouble right now is first defining position sizing clearly in a day
>trading context, and then applying some sort of position sizing discipline
to
>my trading. My concept of it in the simplest terms is basically number of
>contracts purchased: you want to have more contracts on winning trades and
>less on losing trades in the end.
Kent,
If I may add to what Ira has said:
It may help if you start thinking of position size in terms of dollars
risked on a certain trade instead of contracts traded.
For example, the dollar risk for a 1 lot on an Aussie Dollar contract is
quite different from a one lot on a Swissie or BP or CD (within your
specific market). This is because initial and maintenance margin
requirements are different, dollars won/lost per tick are different, each
has a different ATR and a different trending characteristic.
You are trading money, individual contract sizes skew the amount you are
committing to a trade disproportionately to your account equity.
>So entering each trade with 2 contracts,
>you could add to those when the trade trends in your direction, or just
lose
>the 2 if it goes against you in a short time frame.
If you think in terms of money risked per trade and then set yourself a "I
lose this much money I'm out" kind of stop, you will be better off executing
your strategies. Increased volatility (and your lack of understanding why
that is happening) will therefore force you to trade smaller dollar amounts
per trade, and reduced volatility will allow you to trade normally.
Increased volatility are signs of trend change, of dislocation in the
assumed psychology of the prevailing balance between demand and supply. If
you are on the right side of what unfolds, you will make a larger than
normal profit upon trade exit. On the flip side, you will also lose a larger
than normal amount which will in turn reduce your total account equity and
your ability to play longer term will be impaired.
>I am struggling with
>whether to make it mechanical, like if the market moves 10 points to the
good
>from my entry and indicators still look good, then purchase 2 more
contracts
>and adjust a new 4 contract protection stop down close to the new entry
point
I don't know what Van Tharp teaches and I mean no disrespect to his school
of thought. Greg recently posted another position size email to which I have
lingering questions - but bottomline is,
- in the heat of the battle why confuse yourself with complex calculations?
- why not have a simpler methodology in place before a trade and on
auto-pilot active once in the trade? This then frees up your brain to look
only for chart patterns and execute your plans once your chart patterns give
you the triggers.
- if you add to a winning position, from that point on it is a new trade and
you are risking a larger-than-normal proportion of your account to a single
bet. Either the odds are really skewed in your favor, or you are taking one
heck of an "all eggs in one basket" risk. Is the latter justified?
- Even if the odds are really skewed in your favor at the present moment,
how much time will it take for the market to realize that the odds are one
sided and those mis-aligned odds get neutralized?
- Assuming the odds get leveled and you are still in your original position
size, albeit with larger than normal profit, is there a plan in place that
you execute at this point to defend those profits from eroding - while still
leaving you in the trade? Options are a good instrument to study, for they
really do open up a whole new bag of tricks, the simplest of which starts
with Ira's characterisation of "limited loss unlimited profits".
>Does anybody have any thoughts on approaches to position sizing in a day
trading
>context?
I do not "daytrade" in the conventional sense of the word in that my net
take home every day has open positions and not just cash. However, I do
believe the concepts of defending money would remain the same.
I suppose if going home with no open positions is your objective, you would
need to increase the amount you risk on a trade much more than people like
me, since you have to extract from intraday noise without regard to the big
picture trend.
Since you make use of the word trend earlier in your emails, I wonder how
many days in a month actually have intraday trends compared to the days when
there is more range-bound trading? My gutfeel is that trend days number
about 35% of the total trading days in a month - if my experience in
bonds/snp/energies is any indication.
If 2 out of 3 days are non-trending in nature, you are really in no position
to add to your bet once in a trade, since you are betting against the odds
of a non-trending day turning into a trend day - probably towards the end of
the shelf life of the existing trade.
Once again, a dollar amount risk will help you here, since even your added
position will still get you out of the trade once you lose x amount.
Do this enough times and you will find that you are probably better off just
trading 1 capital unit out of 4 or 5 at a time, leaving the rest in cash or
in other trades that have low correlations with your current trade - even on
a daytrader's horizon.
Thoughts welcome.
Gitanshu
|