I watched BigMike's video last night and enjoyed it. I'm really on board with using multiple time frames to filter out trades. It makes so much sense. I just need to find the pairs of time frames that will work for me and which indicators I want to use for my longer-term time frame.
This morning I unleashed my first automated trading strategy on the live market. It's fairly picky so it doesn't trade very often, but it did pick off 22 ticks on the 6E this morning so that was nice. I'll post results as they come in. It backtests with 75-85% win rates with 2.0-3.0 profit factors on with calculate on bar close, but no slippage (FX contracts often suffer from bad slippage), but I risk 2:1 or sometimes or 1.5:1 so I'm not totally comfortable with those risk/reward ratios, though I had it at 4:1 in its earlier incarnations so I've done a pretty good job improving the risk/reward ratios. I'd like to incorporate some other filters to increase the win rates and work on decreasing my risk/reward ratios down to at least 1:1. And I'm only trading 1 contract at a time so I want to keep it small, trying to not risk more than 3% of my portfolio per trade. I'll keep working on it. I have another automated strategy I'm about 50% done with which I hope to launch later this month.
Back on Mike's blog (before the forums) there was a lot of discussion about the failure to create automated strategies and warnings to not chase it. Well, for me and my personality, sitting in front of some screens and looking at bars and lines all day is not my cup of tea, especially with a consulting career I'm not ready to give up yet. So having a program watch the price and indicators for me and trade for me is very appealing, but I'm very cautious and I have no misconception that this is a "set it and forget it" path to retirement. I will continue to do some discretionary trading and look to improve those skills as they will help make me a better strategy developer. I figure when I'm watching the market and trading I'm going through a set of "if...then" decisions in my mind anyways so why not have the computer do it for me? Plus, the strategy development process is very enjoyable to me and an interesting intellectual challenge, except for NT 6.5's weaknesses in back testing and optimization (slooooowwwww, the hassle of merging multiple contract periods, etc). I can't wait for NT 7.0.
Also, I'm looking at developing a risk-arbitrage indicator that tracks the ratio between index futures and their correlated ETF, say, SPY vs. ES, or YM vs. DIA, or NQ vs. QQQQ, and looking for extreme disparities between them and see what kind of opportunities they might provide. I know a lot of investment houses and hedge funds do this kind of stuff which helps contribute the choppiness often seen in index futures.
Had lunch with a more experienced, yet younger than me, trader the other day. He has some interesting things he's working on. One thing he introduced me to was "scale trading". This is very risky and it only works if you can withstand large, long-term drawdowns (imagine still being long the NASDAQ when it was over 10,000), but over time it can be very profitable. He's trying to pool a large sum of money together to do something like this and manage it. It sounded very similar to guys I know that regularly sell lots of strangles or iron condors as the markets scale up and down and just pile up inventories of decaying options contracts that they've sold, watching the theta decay put money in their pocket every day. Sharp market moves can be detrimental though and you have to know how to adjust and recover.
Last edited by shodson; September 1st, 2009 at 06:55 PM.
The following 2 users say Thank You to shodson for this post:
Do you think there is something to do with index ETF & Futures ?
These ETF have volumes, small spreads, seems to move as fast as the Futures do, so I don't think a "small" individual investor can make money with this...
Well, let's look at an example. Looking at the S&P at today's 4pm EST print we have
SPY = 100.15
ES = 996 (roughly)
OK, so we can compute ES/SPY = 9.94508
Let's assume this is an average ratio between the two instruments. If we track that number, which is a spread between the 2 instruments, and if we see it widen (to 10.5 for example) then we may have a thesis that the ES is either over-valued or the SPY is under-valued and act accordingly (sell ES and/or buy SPY). I guess this would be akin to spread trading in commodities, and why people look at "fair value" on the indexes before the market open. I'm not sure. You can then start to look at std deviations on the spread distributions and trade off of outlying spreads values in an indicator.
What's not clear to me is who's wagging who? Are the futures driving the ETF, or the other way around? Maybe we should compare the future with the index and not use ETFs. Since the index is just a measurement of it's components, I don't think the futures can push the index around without pushing around the underlying stocks in the index. But if the future moves too far away from the index's value it could set up for a nice reversal trade.
9/1 continued... - OK, Bot1 took 2 trades today, each making +22 on the 6E, charts attached. Off to a good start! I'm targeting 22 ticks with a 32 tick SL on the 6E. I also have it running on the 6A, 6B, 6C and 6J all day/night long, each with different PTs and SLs but no trades were taken on those today. BTW, this strategy only uses 1 very common indicator with uncommon parameters, nothing else! The rest is money management, back testing and optimization.
9/2 - Slept through the open, up late working on other projects. The opening gap filled nicely. Didn't do any discretionary trading either.
Bot1 didn't take any trades, didn't fire any entry signals. That's OK, it traded my plan for me and I didn't waste any time waiting for the entry signals, and I didn't lose any money. 'I didn't even have to use my AK...must have been a pretty good day." I'm adding some more features to it and diversifying the markets it watches (indexes, commodities, forex, treasuries). We'll see what happens tomorrow.
So I was playing with CL for the first time last night, backtesting BOT1 against it. I found an interestingly profitable variant that looked attractive so I let loose BOT1 on it with live money this morning. What I liked about the variant was that my stop losses were less than my profit targets, and with a greater than 50% win rate. Today, it had 1 winner, 3 losers. The big drop in the market in the morning threw off the indicators. It tends to do better in sideways action. Over time it may recover but I'm not comfortable with this variant of my strategy and I'd rather stick with the original design of the strategy I've tested and optimized for so long so I turned it off. I plan to stick with the strategy's original design.
Later in the day the 6E picked off another late-day winner, helping me recover from this morning's losses in the oil pits.
Also, I stopped following the other FX contracts (6A, 6B, 6C, 6J) and am just going to track the 6E FX contract. Since they are all dollar pairs I felt they were all too closely correlated to USD strength/weakness and left me overweight in dollar pair forex.
9/6 - Took a quick discretionary trade on the 6E while I was working on BOT2. +4 ticks, was targeting 1.4332 but I tightened my stop as it started to get close then it whipped back and closed me out. If I had let BOT1 manage the trade it would have reached it's target a few hours later.
It's 3:20 am now and I should probably go to bed. BOT2 is starting to look good (see attached, tested with 10 contracts from 6/8 to 9/7).
6E - 1 winner, 0 losers = +4 ticks
Last edited by shodson; September 8th, 2009 at 10:30 AM.