need some trading advice

Discussion in 'Trading' started by jedwards, Oct 6, 2010.

  1. I have a system that I've been using since January. In January, I backtested it on a few different markets going back a few years, and it was extremely profitable. In 2009, it was over 100% profitable. I went all-in in January 2010.

    To make a long story short (I posted on this many months ago), I traded it in January and suffered stunning losses, which basically set me up for abject failure the next several months because I lost faith in the system, so I would trade discretionarily from the system, cut my profits early because I was afraid of taking further losses , etc.

    In June, I did a re-analysis of the system, including all data from 2010. Had I stuck religiously with it, I would have made about 40% YTD trading 1 contract. Instead I lost 40%.

    But looking over the numbers, it appears that it is still following the same percentages of success. So after having renewed faith, October 2010 I decided to go all-in once again, with no intervention on my part, fully automated. I put in literally all my money and wrote up a trading plan, where I would trade slightly more aggressively (1 contract per $7500 instead of 1 contract per $10,000). The expected target and stoploss were roughly $150-200/contract per day (the average for 2010), and I would trade once per day. If I hit 20% loss I would quit entirely.

    So, as you can imagine, in the first 4 trading days of October, I've suffered humongous losses: 1 breakeven trade and 3 full losses. To put this in perspective, on average, my system has 1 loss per week, 1 for every 5 trading days. It usually has 2 wins per week, and 2 break evens.

    So to have 3 full losses in the first 4 trading days puts me in a significant hole. The last time 3 losses in a row occurred was January 2009.

    As well, 2 of the losses were double the expected size, $300/contract and $400/contract. So in fact, I've suffered the equivalent of 5 trading losses. My account is down 10% in the first 4 days of trading.

    My question is: What should I do?

    Should I continue with the plan and keep going until I hit 20% losses, or should I scale back my trade to 1 or 2 contracts and bide my time until next month?
  2. JPope


    Trade 1 lots or figure out why it's losing...
  3. Whatever it is, it sounds like you are too undercapitalized to trade it.

    Limit your risk to 1-3% of your TRADING capital.

    If you can't afford a full contract, use Minis.

    If that isn't feasible, then don't trade it.
  4. bstay


    you seems to be explaining your system's percentage win/loss each day each week. have you considered the ratio of your trade loser compared to your trade winner? if the loser is average -1 pt and the winner is average +3 pt, then the odds are changed (in your favour).

    suppose your win/loss percentage or batting average is 60%. you win 60% of time but lose 40% of time. suppose your average winner is 1.5 times your average loser.

    [ 0.6 x (1.5 x R) ] + [ 0.4 x (-1 x R) ] = +0.5R.

    meaning if you took 10 trades, you are likely to make positive +5R where R is your average losing amount. or risk per trade.

    there are two variables to tweak. one is obviously the batting average (60% winners). the other is the win/loss ratio (1.5). if you can carry your winners to several times the average loser, it might change the odds in your system.
  5. Friendly speaking, you should reconsider whether trading is for you, because whoever calls a $700 loss humongous should not trade. This is friendly advice to you. Trading is an expensive sport. I have lost in the past 22K in a single day. I made it all back but it took me 3 months. Your system is probably trading noise. Good luck.
  6. 1) To look at porn doesn't make somebody good at sex.
    2) Your "system" probably has a lot of hindsight bias and small sample bias that doesn't properly simulate an actual market experience.
    3) Do you properly understand how your method should earn money?
    4) With real money at stake, there can be a compulsion to trade every wiggle and squiggle that appears on the chart.
    5) Be willing to come up with another idea that has a "longer" time horizon and "larger" profit objectives so that you can minimize the number of trade decisions you are making.
    6) For day trading, you have to be able to make time-of-day-specific trades without waiting for other types of random set-ups that may not occur. Keep at it. :cool:
  7. Why would you suddenly need to go "all-in" even more aggressively?
    Why not just trade one contract?
    Or, if you are trading regular hours, find an ETF proxy for the contract you are trading (SPY, IWM, etc, etc). Trade 50, 10, or 5 shares. It won't be a perfect correlation, but it's good enough for government work. Commission will be a much larger percentage of your fees, but that's an R&D cost. If you're with somebody like IB, it'll only be $1.
    You can find short and long ETFs for the major markets, so you don't even need margin.
    This way you can:
    1) Live test your system over a much longer period of time (i.e: not a bull market like it has been much of the last year).
    2) Get stopped 50 or 100 times and still lose a tiny fraction of your account. And not need to fret every time you get stopped, even if it's 10 times in a row.
  8. blox87

    blox87 Guest

    You need to backtest your system thoroughly and understand it inside and out in order to trust it through the DD's. You should know your uncle points by running monte carlo simulations on your DD's , factoring in slippage/comms, testing out of sample data ect.

    Until you are fully confident your system works, you will stop using it at the worst time and start using it at the worst time. If you are any good as a discretionary trader then maybe diversify and use 1 contract with your system and 1 contract discretionary. Only if your risk limits would alow this of course.

    Best wishes to you.
  9. Thanks everyone for the replies!

    psytrade and jimbojim, I was using an approximate 2% risk per trade given the average size of a trade (normally between $150-200), so I don't feel like I was too undercapitalized. jimbojim, the value was per contract, but I was trading 10 contracts, so my loss in the last 3 days was about 10%, but around $7500 in notional value, which I think is pretty humongous compared to the statistical expected results (maybe 1% up).

    nazzdack, I probably wasn't clear. I backtested my results for each day using tick data from Sept 2008 to Dec 2009 (I trade futures). In addition I tested on 2 other foreign markets and got very similar results, which gave me further confidence. From Jan 2010 I traded real money, so essentially it has been forward tested for 8 months. The forward tested results and the backtested results pretty much matched (except it was more profitable in 2009), which gave me confidence that the system worked. As well, I only take the first trade of the setup per day, so I feel like it's a pretty disciplined approach. Definitely had I stuck with just the system it would have been profitable, however, I messed around with it which caused the significant losses.

    TheGoonior, the reason for me going all-in was because after forward testing the system since January 2010, I realized that it still worked. When I first went tried trading it live in Jan 2010, I suffered significant losses, which made me gun shy, but after going through the numbers again, I realized that over the long haul, the system does seem to have an edge, which is why I decided to trade it with more money. Just my luck that I suffer the worst 3 day loss in 1.5 years...

    blox87, unfortunately my discretionary trading is absolutely horrendous. It's beyond bad, it's the exact opposite of the markets, which is why I have abandoned it completely. I agree with you though, that I should stick with it because I would likely stop at the worst time, ie. right now.

    bstay, I actually have those numbers. Essentially, the trade over time has an expected value of 0.2X, where X is the value calculated by the system on a per-day basis. There is a 20% chance of a full loss, 40% chance of a break-even, and 40% chance of win with a 1:1 risk:reward ratio. There is a 20% chance of a 2X win, and a 15% chance of a 3X win. I've run the studies and going for the 2X and 3X wins aren't very profitable, so I have decided to simply go for the 1X wins. I have toyed with the idea of going for the 2X to 3X win (that is, place an entry order at the 2X target price with a target of the 3X price) because that has a 3:1 win ratio, but the risk reward is 2:1, meaning that the expected value is again 0.2X, which makes it just as profitable as the original system.

    Thanks again for the responses. After formulating my response to this, as well as rereading my original forum post, I think it's better to just stick with the system. Getting my frustrations out I guess helped. As blox87 says, I don't want to leave the system as the least opportune time, so I will just stick with the system. The worst drawdown on a per-month basis the system has had since Sept 2008 was about 7%, so hopefully the conditions change for me!
  10. There is your answer, keep trading and don't worry so much about the outliers right now. If they happen, and your system still returns a handsome % overall, you are ok. Just don't regress into discretionary trading.

    #10     Oct 6, 2010