A big thanks to Mike, billr for his help on Renko charts, nakachalet, Devil Man, NJAMC, aslan for his work in NT7 chart types, Sam for keeping the freeloading code-jackers out of the elitecircle, and drmartell for the awesome market replay downloader script for NT7 (software here () - I downloaded a little over 4gb of mrp data yesterday, and plan to do allot more today. I'll compress it hard and post it on the board for others to use. This really is a great online community, and I'm grateful I found it last Wednesday. Learning new things quickly becomes more and more difficult as one grows their knowledge, but this place is definitely a huge library of raw and fresh reading material.
r3723 is a system that has been profitable every day since March 1st in live trading. Last night I ran the system (and 6 other tweaked versions) via MarketReplay for 78 days on the SI-March contract only, with profitable results for all 7 subversions. The best version brought in a net profit of $17,550 during this period. Combined with the live results during the past 2.5 weeks, this equates to a net profit of a little over $25k for trading one $1k SI contract since the first week of December.
The secret sauce is a filter that effectively distinguishes trend versus range periods on multiple chart types. Currently, I'm using BetterRenko buy/sell signals, however some quick work using time based charts shows promise if you come up with your own buy/sell signals. The filter is a moving average of the derivative of volume (the rate of change of volume, or the slope of volume). 3723 uses the Hull Moving Average on the VROC. VROC is a generic indicator that is the slope of volume with a smoothing constant. The smoothing constant isn't needed, but does make things look a bit prettier on the charts, and it makes easier to program in NT7. This filter can be thought of as a backwards derivative. The thinking is a moving average of the rate of change in volume will effectively outline which big volume movements are the big shares (MM's et. all) initiating trends.
Although the system is ready for live trading today, it still could be even better, so during the next day, I'll be focusing on trying a list of hypothesis' I made in the past 18 hours based off of my analysis of the 3-month live executions.
Note that backtesting any kind of Renko chart is impossible - billr made a great missive ( on why this is so. The only way to test Renko charts is in live trading. Market replay is the effective equivalent of live trading. I've also compared real live results with market replay results by hand, and found them to be absolutely identical.
→ Why you should post
Although I'm the sole quant behind this system (okay, quant is an inappropriate word for retail trading! :p), the system benefits from the BetterRenko chart type created by EliteCircle member aslan, and it benefits from my new knowledge after burning through hundreds of pages of EliteCircle pages the past four days.
All of us are wary and skeptical of posting the code we write, the systems we build, and the ideas we concieve of; worrying that the edge will be reduced due to more people using it, or because of some freeloader exporting it from our circle.
Before you decide to never post original algorithms, systems or fresh ideas, consider this: firstly, our own edges will not be rendered less effective because of the small group of elite circle members will implement or build on them. Hundreds of trillions of dollars cycle through the derivatives markets each year, meaning even if each of us rack up a couple million during this new quarter after piggybacking off of each other's ideas, it's still considered to be small potatoes by Mr. Market's standards. Edges are not reduced by people jacking code, they are reduced by changing markets, making a system's ability to draft behind the big boys in the market less effective.
So, there is no doubt that it will be hugely more effective for our small group called nexusfi.com …