Weekly results of a wanna B algo trader

Discussion in 'Journals' started by Engine99, Jun 4, 2010.

  1. I had an awesome weekend. Got back this afternoon from a family camping trip full of fun activities (and the right amount of dirt). Didn't think about the markets at all except when we left on Friday and I was still happy that I was saved by the bell.
    Overall this was a decent trading week, except for Thursday. I finished up 1k5 while the S&P 500 was down 1100 bucks. I'm happy with these results but it could have been better considering I wasn't on the road this week.
    I started messing around with weekly options, so I've a new strategy (S11) which is S10 except for weekly expiration cycles. This is mini small right now, since the fills on these weekly is somewhat bad and the delta/gamma risk is huge.
    I'm disappointed with my S70 performance. That system backtested so well, however several weeks into this it's just showing losses. I'll give it another month but I'll have to look into the execution and how it matches the backtesting.
    I'll be on the road in Texas next week, leaving my systems on auto-pilot. I'm hoping to load up on some more S2 trades but with the earnings season winding down that's not very likely. Anyway, for anyone following this, enjoy the rest of the weekend.


    Here are the results in detail
    S1: -584
    S2: 3832
    S10: 6084
    S12: -63
    S70: -356
    S11: -4
    Weekly Performance:
    1578
    S&P 500 Weekly Performance:
    -1126
    Overall Performance since launch:
    9333
    S&P 500 Performance since launch:
    -2523
    Risk Free Checking Account Performance since launch:
    308
     
    #51     Aug 22, 2010
  2. What a lucky idiot I am. Friday saved my week. Overall Up 700 bucks (really lucky). Here are the daily results: Mon +314, Tue -2883, Wed +997, Thur -681, Fri +3007. I've had to make some accounting changes after hours on Friday. I'm basically adding 20k into my trading account, so I'll use Friday's S&P 500 SPY close as the reference (even though the funds won't be available for trading until mid next week). Now the S&P 500 reference is up to 1482 shares. That's what I'll measure against. Next week's results will reflect that addition. As far as my trading goes, it was pretty quiet except Tuesday (when I flew to Texas) and Friday (when I was back home). I made some small S1 trades and they worked well. S2 added a little bit of time decay on it's positions. S10 kicked butt on Friday after that big SPY move. S12 did ok, after the system reversed the VZ position and went slightly long. S70 still sucks, I'm not sure what the issue here is. I'm debating whether or not to continue that strategy after the September expiration. S11 (S10 based on weekly options) is down but down much less than it was on Tuesday. It triggered all stops on Tuesday and started fresh this Thursday. I also did ok at the online poker tables this week, cashed in almost every 6 handed no limit game I played. Feeling good as I head into the weekend. I'll be in Seattle next week, so systems are on auto-pilot again and I hope to continue my current run. For all the followers, have a good weekend and enjoy some relaxation time.


    Here are the results in detail
    S1: -281
    S2: 3972
    S10: 9165
    S12: -36
    S70: -358
    S11: -98
    Weekly Performance:
    754
    S&P 500 Weekly Performance:
    -1618
    Overall Performance since launch:
    10087
    S&P 500 Performance since launch:
    -4285
    Risk Free Checking Account Performance since launch:
    337
     
    #52     Aug 27, 2010
  3. S70: Equity Options Strategy, hold time 1 month. Based on a single technical indicator, it's a coin flip that benefits from theta and delta changes.

    Based on the above description, I'd say it's a Las Vegas strategy-at-best. Get out Get out GET OUT.
     
    #53     Aug 27, 2010
  4. promagma

    promagma

    Curious what software you used to backtest option strategies? Home grown?
     
    #54     Aug 27, 2010
  5. I use several things.

    I use thinkorswim (it has a backtesting capability called thinkback) but that's just to eyeball something cause it's a lot of manual effort. It's similar to optionetics platinum except it's free if you have a think or swim account.

    Then I've bought 385 million option quotes about 8 month ago that I have in a mysql database so I mostly use that for any mass/bulk testing.

    I also have a ton of stock quotes in that database, so any delta strategies I backtest against stock proxies.

    hope that helps
     
    #55     Aug 27, 2010
  6. The backtests were about 57% in my favour though, so it's slightly better than vegas ;-) it was like being the casino ;-).

    But you're right, it doesn't even closely live up to the backtest results since I've been trading it. I'm not sure yet why :(
     
    #56     Aug 27, 2010
  7. Two issues come to mind immediately.

    (1) Where did you buy the data and how was IV calc'd?

    (2) Are you trading the penny spreads?
     
    #57     Aug 28, 2010
  8. add to that:
    the backtesting was probably do one or more of the following:
    1) buying at the bid and selling at the ask.
    2) no slippage was taken into account
    3) stop orders probably assumed a 7x24 market....the huge risk with options is that any stop losses can be executed MILES away from their intended stop point. The overnight risk can be huge on a percentage basis.
     
    #58     Aug 28, 2010
  9. I appreciate your input. Here's some more info on S70.

    S70 doesn't have stops, it's basically a system that trades verticals only.
    The risk is very defined, so I let it run to expiration. It uses a pretty comples money mgmt schema where if the week is a loosing week for S70 it cuts the position size for the next new position. The minimum position size is 150 bucks. If the week is a winning week it increases the position size up to five times.

    Backtesting does assume I get filled on the mid price + 5 cents on the entry. So far that has been the case for all real trades, I basically put them on @ the mid price @ 3.50 and once one trade gets filled I cancel the other ones. It also takes into account commissions (very low with IB) or possibly exercise cost (15 bucks per leg) if the vertical is between strikes upon expiration. It assumes that in that case (e.g. I'll exercise the long and the short expires worthless) I have enough capital to offset the trade with stock.

    It does not take into account possible assignment but that's not likely with these trades anyway.
     
    #59     Aug 28, 2010
  10. Good question on the IV (basically I've no idea how it was calculated). I bought my data from here
    http://www.historicaloptiondata.com/purchasing.aspx
    and so far the tests I did have been good. The trades are all verticals so IV doesn't really change the risk graph since I let it run to expiration


    I only enter trades that have an underlying average volume of > 2 million shares/day, so most of them trade in penny spreads, however I assume a fill price close to the mid price (+ some small slippage)
     
    #60     Aug 28, 2010