ETFs

Discussions specific to trading the stock market.
Post Reply
-w.
Roundtable Fellow
Roundtable Fellow
Posts: 95
Joined: Wed Nov 19, 2003 11:02 am

ETFs

Post by -w. » Tue Dec 09, 2003 9:10 am

Considering that ETFs are a fast growing and increasingly liquid market segment it might make sense to trade them in a LT trend following system. However, some of the more interesting stocks only have a rather short history, further distorted by huge intraday spikes (illiquidity or false data?). Also, they bring new instruments all the time.

Some of the ideas I've come up with are backtesting the underlying indices or backtesting with the largest stocks in the index. The second has a serious flaw though as the largest stock today maybe hasn't even existed 10 years ago. And testing indices means negating liquidity, not a sensible thing to do.

ETFs have been mentioned in this forum before so I wonder if somebody has come up with alternative solutions/strategies?

-wojo

Forum Mgmnt
Roundtable Knight
Roundtable Knight
Posts: 1842
Joined: Tue Apr 15, 2003 11:02 am
Contact:

Post by Forum Mgmnt » Tue Dec 09, 2003 12:05 pm

wojo,

I haven't attempted to run simulations against the ETF historical data, but I have traded them as hedges when the market moves quickly for other positions.

I can tell you that like the futures market as compared with the underlying indices, they tend to be more volatile than the underlying and overshoot at price extremes by quite a bit at times of high volatility.

So for trading on a shorter-term basis, I wouldn't recommend trying to use data from the indices or from the underlying stocks as a proxy for an ETF. I would use the actual price history of the ETF itself. Anything else is very like to be significantly different on a minute by minute basis at just those times that matter most.

I can't say definitively without seeing the data you mention, but based on what I've seen I suspect those spikes you talk about are real trades, so be careful.

- Forum Mgmnt

-w.
Roundtable Fellow
Roundtable Fellow
Posts: 95
Joined: Wed Nov 19, 2003 11:02 am

Post by -w. » Wed Dec 10, 2003 9:35 am

I'm testing a breakout system for stocks similar to the Turtle System. One observation I have made is that closing prices in the long run seem to have better significance as breakouts than intraday hi/lo's. So I figured that buying/selling the open would make sense, for pyramiding positions I use the same rule (my mark is the closing price). That doesn't go for stops, of course. They are triggered anytime during the day when the price crosses.

As far as I know, you're entitled to the opening quote at the NYSE, so there should be no or very little slippage (as opposed to NASDAQ, where you might get partly filled, like 100 shares at the opening quote and the other 99,900 like 5 points away :roll: ). Anyway, the intraday spikes probably hurt you in the beginning, when the stops are triggered. But the same goes for other stocks as well, I guess.

The main reason I'm looking into ETFs are the potential diversification they provide for a portfolio (ok, you can have that with individual stocks as well), their relatively lower long term volatility but very importantly because clients seem to like the story. Of course, with good money management it shouldn't make a difference but I think what clients really like is the reduction in single company risks. Call me an opportunist, alright :wink:
So for trading on a shorter-term basis, I wouldn't recommend trying to use data from the indices or from the underlying stocks as a proxy for an ETF. I would use the actual price history of the ETF itself.
Thanks for your insight, that sounds logical to me. Maybe I should just go on testing my system long term for it's overall robustness on stocks and then compare the last couple of years with ETFs, traded with the same parameters.

-wojo

Post Reply