[…]first month Instead of ranking on the 6 month return of the ETFS, the ranking return is calculation by ignoring the most recent month. If we wanted the 6 month rank on the ETF and today is June 30, we then use the return from Dec 31 to May 31. The idea being that one avoids ETFs experiencing a blow off. I have tested this before on different universes but with no success. Maybe it will be different on these ETFs. Rules Rank ETFs from high to low using the N month return but ignore the most recent month. If […]
[…]is how the initial research in the article was done. The method of testing all trades is quite common but I rarely do it this way. It is not my preferred way of testing an idea like this which I will cover on why in the next blog post. Setup Rules Stock is a member or was a member of the Russell 3000 The Dollar Volume of the stock is greater than $500,000 Close is above $1 Close is above the 200-day moving average The 2-period RSI is below 5 Place limit orders 4% below close for the next day. […]
[…]post for rule details. Tested from 1/1/2004 to 6/30/2014 Idea 1 – Remove MA100 rule Comparison is on Russell 1000 only. For S&P500, Russell 3000 and NDX100 see the spreadsheet. Take the original rules and remove the rule that the Close is greater than the 100 day moving average. New Results Original Results For this time period, as expected the exposure went up but the CAR did not increase that much. The max drawdown increased dramatically. Definitely in the realm of very few people could handle. The year 2008 is still positive without the MA100 which is interesting. I expected […]
[…]which started on October 2023 has been strong. My trading buddy and I started wondering how this compares to the past. Is this a “normal” strong rally or an “abnormally” strong one? Determining this is always tough because it depends on the indicators you use. Because of that, I tried lots of them. This will be a post short on words but with lots of tables. Where are we now? Data is from 1/1/1980 to 12/19/2023. The close of 12/19/2023 was the 36th day of the rally. These are the stats based on these last 36 days. 18 days with […]
[…]4, 6, 8) exit after this many months if still in position (NA, 50) stiffness exit. For NA waiting for one of the stops, loss, target or time to hit. Comparing to the CAR of 10.85, only these 4 runs did better and not by a lot. What is good to see is the low DD. Not shown but in the spreadsheet, is using a Stiffness value greater than 95 improves the CAR a bit more to over 12. Spreadsheet & Code File the form below to get the spreadsheet with lots of additional information. See the results of […]
[…]Heikin-Ashi Candlesticks use the open-close data from the prior period and the open-high-low-close data from the current period to create a combo candlestick. The resulting candlestick filters out some noise in an effort to better capture the trend. … Heikin-Ashi Candlesticks are not used like normal candlesticks. … Instead, these candlesticks can be used to identify trending periods, potential reversal points and classic technical analysis patterns. Testing Details For these tests, the focus is on ETFs, for a change, with the testing dates from 1/1/2006 to 6/30/2014. Daily Timeframe Searching around, one common way to use Heikin-Ashi is by pairing […]
[…]part of the first class with no credit card needed. Two hours of professional instruction plus bonuses for $175. If you like the first part, email me and I will send you a PayPal invoice. This course assumes you have a good working knowledge of AFL. If you don’t yet have those skills, please consider taking the AmiBroker and Backtesting 101 course first. Thank you, Cesar […]
[…]I need to know that a strategy is not sensitive to changes in the parameters and did fine in out-of-sample testing. It would be hard for me to trade this strategy even with really good results. Now, there is nothing wrong with camp 2. The numbers would have to be really compelling for me to take this strategy seriously. Now does this mean, I would not trade a black box strategy? It depends on how opaque it is. Strategies that use machine learning can be very opaque on understanding how parameters were selected and are usually parameter sensitive. This I […]
[…]A CAR of 28% compared to -8.47% with using a $.01/share. A huge difference. Your assumption on what to use for commissions would change whether you would consider trading this strategy or not. When your edge is only .35% for an avg. % profit/loss, what you assume for the commission will have a large impact on the results. This definitely surprised me on how much the numbers changed. Original Strategy How much does the original strategy change with different commission? Here are the results of the original strategy with the various commissions amounts. I would say most of these results […]
[…]an optimization, we can now identify the “best” variation for each market regime during our in-sample test period. This is an exercise best accomplished with some help from Excel. We’ll then plug the adaptive parameters into our AFL without breaking the static (non-adaptive) functionality. Video 3 In our final session, we’ll see whether our adaptive strategy outperformed the static version. The most apples-to-apples comparison would be between the Adaptive OOS test and a Static OOS test using the best parameters from the IS optimization. Of course, we’ll also want to see how our Adaptive version measured up against the full […]
