Short-term trading vs Long-term trading

Discussion in 'Strategy Building' started by es175, Sep 8, 2006.

As a systematic trader, What timeframe is preferable to operate in? Aim is net profit

  1. Short-term trading is better

    21 vote(s)
  2. Long-term trading is better

    11 vote(s)
  1. es175


    *The context is systematic trading.
    *The variable in question is holding period per trade.
    *Let's define short-term trading as holding an individual trade between 0.5 to 2 days. We can define long-term trading as holding a given position for 1 week to 3 months.

    What do people perceive as the benefits and drawbacks of operating in each timeframe?
    If you were being rational, which timeframe would you choose to invest your development time in?
  2. WinDiff


    How bout you start doing some serious work on your own using your own brain?

  3. es175


    The question is intended to promote discussion.
  4. There is a big difference between day trading and overnight position trading. Slippage and the opening gap can make a short term overnight position trading system unprofitable.
  5. ST is more profitable.

    Ocassionaly LT will pay off.

    No clear line.
  6. "Short term" and "longer term" dont really mean anything because to me, long term is a month or more, short term is at least a week, while to some short term is a few seconds to a few minutes and long term is an hour or more.

    The shorter your time frame, the smaller your profit/ loss will be, AND you'll have more expenses, like l2 fees, streaming news fees, charting software fees...and so on.

    longer term traders dont worry about all that. We make more $$$ per trade because we keep costs down.
  7. kut2k2


    Let's not.

    Short-term trading: intraday or eod time frames.

    Mid-term trading: eow or eom time frames.

    Long-term trading: eoq or eoy time frames.

    The length of any individual trade is irrelevant. What matters is the time frame you're basing it in.
  8. man


    consider you have a long term break out futures system like a turtle of bollinger breakout with about 1300 roundturns per million dollar per year. it trades 50 markets and trades on average seven times per year per market. this kind of system has a potential to trade a modified sharpe ratio of around 1.0, maybe a little higher. since the system trades infrequently and gets out of losing positions quite late, your daily returns are expected to be positive, but with a big standard deviation. you are quite certain that each two year period is positive, on average you chances that next month is postive are maybe 70%, your next week is more uncertain and so forth. your daily pnl shows an already quite a lot of uncertainty.

    now consider you can have the exact same statistics, just on a shorter scale. you make the same number of trades, but in a tenth of time. your odds now shift upwards for months, weeks, days. the flip side is that your trading cost increases as well as the amount of money you can manage is reduced.

    so, generally speaking, for a small account the higer the frequency the better. all risk return ratios will be better the more frequent you trade.

    my threshold for a long term trading system is a sharpe ratio of 1.0 for a portfolio of about fifty markets (all futures btw), for short term systems it is about 2.0 for a portfolio of about twenty markets.

    to start system building i would recommend to check out turtle systems. surf on curtis faith. he published their strategy and with some little tweaks you have an initial trend follower on daily data. the good thing is that you build up some feeling about portfolio effect, since on single markets long term trend following is untradeable IMHO. no trendfollowing without portfolio effect.

    for intraday acrary once posted a turtle-like trading system three years ago here on the board. if i remember correctly it tested a sharpe of about 3 on twenty markets, but it had some trouble after acrary posted it. but i do not have any update and am just talking from top of my head.
  9. bolter


    Your question is very broad - "which is better"? Better in what sense? And not much context to go on.

    I strongly suggest you bone up on the conept of expectancy as popularised by Van Tharp.

    In answer to your question, from the perspective of returns generated by a system normalised by some measure of risk (ie: better) ..... invariably a "good" short term system will exponentially outperform a "good" longer term system over time. No question.

    The issue of trading costs (bro & skid) can be misleading. If you have a definite edge (ie: very stable positive expectancy) then it is in your interest to trade as often as you can - providing your expectancy exceeds your execution costs. Think casino - they have an edge of around only 3-5%. That is, their transaction cost (payout) is 95-97%. But then again they have a hard mathematical edge.

    Longer term system (think weeks and months) are almost invariably trend-following in nature (buy strength sell weakness). The performance of these systems (including the Turtle System) have been eroding for years and most are longer viable.

    Have a look at my post here:

    Having said all that it is much more difficult to build a good short term system - but this is as it should be. The natural order of markets is being preserved.

    Good luck,
  10. I won't say he has "the right idea" about time frames, but my thinking is most in alignment with kut2k2's concept about time-frames.

    While I (and I see that most) traders voted for the short-term time frame as being "better", the fact of the matter is that what we call the short-term time frame is actually very difficult (many say impossible) to trade efficiently.

    It requires an extraordinary amount of skill to do so successfully, (or you must be willing to pay those who have figured out how to do it successfully, to either teach you or give you some insight into their methodology, or pay for their signals, etc.) to actually succeed at it.

    While it theoritcally requires less capital, it also requires the greatest amount of skill to do it, and do it well (we're talking making a living at it, if you are trading for supplementary income, it is not so difficult). Without good professional help, I won't say it's impossible, but it is a tough one.

    On the subject of money management, because of the increase in randomness which exists on the intra-day timeframe, money management IS your edge.

    Finally, while it begs to be traded in an undiscplined fashion (because of the wide swings which can exist on an intra-day time frame), it actually requieres the most discipline of all.

    3 years with a lot of work, some good help, and a little luck.
    The middle time-frame is probably optimal for most traders who have a sufficient capital base to work with, have a good technical method entering trades, can consistently determine trend based on their method, and have good money skills which are a little bit better than basic (the ability and knowledge to scale into positions will enhance your edge).

    You could probably begin with some good books on chart reading (if you choose to use charts) - try The Encyclopedia of Chart Patterns by Charles Bukowski for starters.

    6 months to a year and you could be up-to-speed
    I don't have any experience with trading leveraged securities over the long-term time frame, so I can't speak to that, hhowever, if you are trading unleveraged (or low leverage) securities over an extended time-frame (say, 6 months to ... years), that would give you plenty of time to make your decisions, very little psychological wear-and-tear and little real risk ... for the record, I know of several very smart women who only hold long-term (unleveraged) positions and they are doing very well for themselves.

    Um, to be quite honest with you the women cited in the example haven't had an forma training in securities, unless you count being taught by their boyfriends/ex-hunbands :) .

    Best Regards,

    Jimmy Jam
    #10     Sep 11, 2006