Click chart to enlarge:
That’s a chart of Ziggy v1.04 running on TradeStation’s “continuous” futures contract data. You would think that the blast-off after the TARP crash in September would be due to the VIX spike, but Ziggy is still going strong as the VIX declines.
But that chart is very understated. In real life, that $62,349 profit would be much higher. First, the continuous contract chart includes overnight data which distorts indicators at the open each morning. Using TradeStation’s “daytime session only” charts, the number is $78,853.
Second, for these tests I’ve been using ThinkOrSwim’s $3.50 commission. But Ziggy will trade with TradeStation’s $2.40 rate, so add another $1.10 per trade. As a scalper, Ziggy makes a lot of trades – 6,709 over this period – so add on several thousand more dollars.
Third, these numbers are for a single contract. As Ziggy builds its equity, I will have it trade more contracts. So, after it made $5,000 in October, it could have traded two contracts in November, and instead of making $13,000, it would have made $26,000. And if I let it scale up as soon as it had the equity, it would be more than that. After a few more months of that kind of progression, it would have a rather gigantic profit. And the more contracts it trades, the lower commissions go with TradeStation’s bulk-rates. With enough volume I would buy a seat on the Chicago Mercantile Exchange and cut commissions to the bone, which will free-up even more capital for trading…
So that’s the plan.
But there’s more! This version is the first to look at market internals to judge the strength of the market; so I have only scratched the surface there. Ziggy also does not use any of George’s methods yet. It doesn’t even look at higher timeframes. That will be another avenue for improvements. This version now has three new modes: in addition to “scalp” mode, it now has “bull”, “bear”, and “night”. So, it can play trends in “bull” and “bear” mode, and scrape out a few dollars overnight. But I have only begun to develop those parts of the program, so they will improve with further effort.
Believe it or not, I have also come up with a completely different strategy that surpasses Ziggy by 33% when testing on 4Q08 data. Ziggy does better over the compete range of test data though, so this new strategy will come to the forefront the next time the VIX spikes. I have built a modified version of it into Ziggy though, and the program will use it at certain times.
Ziggy has been profitable for 13 weeks in a row:
It has only made $2,900 in July, but I am very happy with that performance since Ziggy is optimized for range trading and July has been a very trendy sort of month, first plunging down, and then shooting straight back up.
Here is a picture of Ziggy scalping the futures this afternoon:
Profitable trades are connected by green lines; unprofitable ones by red lines. Lots of green, right? “1″ means it is long one contract. “-1″ means it is short one. As you can see, it constantly flips from long to short.
At 15:04, Ziggy was short when CNBC announced a positive development for nearly-bankrupt CIT. But it made up that loss on the next two scalps. Ziggy made $300 during the period shown on the chart. With 10 contracts, it would have been $3,000.
Note on the chart it says “Strong Buy”. That’s because Ziggy judged the price action to be strong and, as a result, it was more eager to buy dips, and more reluctant to turn short.
Ziggy does not use stops. It’s takes its losses like a man. Actually, I have run hundreds of thousands of back-tests trying every conceivable combination of stop-loss and profit-target settings, and every last one of them reduces performance. Isn’t that fascinating?
P.S. One last thing… Ziggy might do even better with the NDX futures since they are likely to be more volatile, just like the QQQQ is more volatile than the SPY. So, the scalpings should be better. I have done one small test so far, and this does seem to be the case.