Page 1 of 1

Pre-emptive money management

Posted: Wed Nov 22, 2006 3:57 am
by rabidric
ok, for the record, i am a true beliver in following the trend and sticking to my system rules(ie no cherry picking or screwing around based on hunches etc.). I also do not trade off the back of any fundamental info etc.

now that is off my chest, here comes the Heresy:

systems never take into account scheduled events ahead of the fact. this has impacts depending on the time frame you trade in relation to the scale of the scheduled event. so the following will mean more to short term traders than pure LTTF.

e.g.trading bonds, lets say you employ a volatility based sizing approach on entry. mkt is quiet, you have no position. but Fed is releasing a rate announcement and all the pundits are split 50/50 on how big and which way(i.e. it is a complete unknown and the release could be way out of line either way).

if you stick to the system you could be putting on a large sized position on the back of a triggered signal, that then tears you a new arse. yes, it could go your way, but if it doesn't...you are overleveraged perhaps in relation to standard size.

i guess it is like the old historical vs implied volatility thing.

anyway, to cut a long story short, in a large percentage of backtests on my systems, if i looked closely into the sequence of trades that made up the 2-3 largest drawdowns in the backtest history, i found that they inevitably included a "go large into massive adverse volatility trade" ..or 3.

i also looked at some DDs that were 4/5 as deep as the worst, and found that some of these contained a couple of "large size into volatility" entries, that happened to go my way- but if they didn't , and went the opposite way, then these drawdowns would have exceeded the max drawdown of the backtest.

To sum up:

by cutting size ahead of an expected volatilty surge(i.e a known event that meets a systematically defined criteria, not a just hunch that it is "quiet, too quiet"), i have been able to outperform the historical backtest of my current system. I should add that i replace the size within a short time, if still in the position. Even though it might raise the open trade risk when the market goes your way and you have to renter the size further onside, it is a small price to pay for the advantage of having a small postion on when the storm hits.

Also i have noticed anecdotal evidence that suggests, the biggest spikes occur AGAINST my existing positions- if i am aligned with the rest of the market, then we all have to react if an event goes against the grain(i.e. buy the rumour , sell the fact). So cutting trend following positions ahead of major figures may actually have a positive expectancy, and not just a risk control benefit, in the short term after the figure(though given enough time, the trend often reasserts itself, but by then i usually have the full position on again, or the small temporary position is stopped out, and i am taking new entries again).

over to you guys.

Posted: Wed Nov 22, 2006 4:35 am
by edward kim
many outstanding trend-followers i know often go flat before major econ reports and fed decisions.

Posted: Wed Nov 22, 2006 10:47 am
by Paul King
If one uses volatility-based position sizing then the most dangerous time for your trading is when your preferred estimate of current volatility turns out to be very inaccurate.

Obvious, and avoidable, examples include scheduled conference calls for public companies. In these cases you can either reduce position-size to account for increased volatility (as you mention), or simply avoid holding positions through these kind of events entirely.

This only leaves unpredictable random events to mess up your volatility estimate; but that's just part of risk taking period :-)

Paul

Posted: Thu Nov 23, 2006 1:02 pm
by droskill
My question is how you would measure these high vol events going forward. So, sometimes the fed speaks, and nothing happens, but sometimes the fed, or the CPI or some other economic event causes a huge vol move.

Also, aren't there so many of these events that it becomes very difficult to employ a long term strategy with all the events? Or maybe you're just reducing size....would love to hear more feedback on it.

Posted: Thu Nov 23, 2006 5:13 pm
by BARLI
Forum Mgmnt, did Rich teach turtles about managing risk during reports that come out, for instance WSDE that comes out once a month or some other report? It seems that turtles didn't care about it

Posted: Thu Nov 23, 2006 5:19 pm
by Forum Mgmnt
We didn't worry about news at all. I didn't even know when news was coming out.

- Forum Mgmnt

Posted: Fri Nov 24, 2006 4:48 am
by edward kim
droskill wrote:My question is how you would measure these high vol events going forward. So, sometimes the fed speaks, and nothing happens, but sometimes the fed, or the CPI or some other economic event causes a huge vol move.

Also, aren't there so many of these events that it becomes very difficult to employ a long term strategy with all the events? Or maybe you're just reducing size....would love to hear more feedback on it.
there are many ways to do that, and for obvious reasons, very few are willing to share that information with you.

implementing a long-term strategy has the opposite effect - it becomes easier with all those events you mention.

Posted: Fri Nov 24, 2006 8:22 am
by droskill
edward kim wrote:
droskill wrote:My question is how you would measure these high vol events going forward. So, sometimes the fed speaks, and nothing happens, but sometimes the fed, or the CPI or some other economic event causes a huge vol move.

Also, aren't there so many of these events that it becomes very difficult to employ a long term strategy with all the events? Or maybe you're just reducing size....would love to hear more feedback on it.
there are many ways to do that, and for obvious reasons, very few are willing to share that information with you.

implementing a long-term strategy has the opposite effect - it becomes easier with all those events you mention.
Quantifiable ways? You might be able to make some judgment by the position of the VIX going into the meeting or call, and the positioning of the public, via sentiment measures (put/call ratio, % of stocks over 50 day moving average) - but I've never seen a measure that was very consistent. But with all the rocket scientists out there, I'm sure someone has it....

Rethinking what I wrote - I see your point about the long term strategy.

I am also guessing that you have figured out a way to measure such things and are, naturally, unwilling to share it! :D

Posted: Fri Nov 24, 2006 9:06 am
by BARLI
droskill, usually before important meetings the previous day's range will be very small, traders dont want to take big bets so theres not much activity going on, take a look

Image

Posted: Fri Dec 15, 2006 3:53 am
by cramnhoj
if you stick to the system you could be putting on a large sized position on the back of a triggered signal, that then tears you a new arse. yes, it could go your way, but if it doesn't...you are overleveraged perhaps in relation to standard size.
In my opinion, if you become overleveraged when a position starts going against you, then your problem is in how much you risk per trade. To me, when opening a position, one should only risk what one is willing to lose. Then rely on the positive expectancy of your system for you to profit. If this trade happens to be a loss then so be it, other trades will come.

If you already have a criteria for determining when to reduce position size like when nearing a news announcement, and you're satisifed with the results in your backtesting then by all means include it in your system. But as with any part of a trading system, check whether what you're adding will curve fit your system to the historical data.

As for me, I don't look at the news and don't even know when there's a news release. I don't see it affecting my system.