The Currs column shows the signal for 10/20 for each strategy. That was surprisingly tricky to get because the accounting part of the algorithm has no interest in today until the close. The win/loss numbers are pretty consistent between 1x and 3x equities. Zomb1 goes long when at least 12 of the other strategies are… Continue reading 10/20/20 Current Signals
Category: Methodology
10/19/20 Win Percent
The table shows log returns for the past year for the 3x Significant 6 as usual with the probability of having a positive day with the various strategies. For example, SPXL for Buy and Hold had a gain on 59% (590 out of 1000) of the days. TQQQ at 64% (641) is pretty amazing. Long… Continue reading 10/19/20 Win Percent
10/12/20 – Zombies, etc.
I was trying to write on article about Zombies for SA this weekend. Zombi is an acronym for Zero Optimized Median Based Inferentiation. Consenso 1 and 2 are now zcon1 and zcon2. zomb1 is a pure zombi. Zombies reanimate the corpse of previously calculated strategies. It seemed that the difference in performance between 3x and… Continue reading 10/12/20 – Zombies, etc.
10/8/20 – New Methodology
The trend following strategy is em3. That 5451 number for SOXL is difficult to grasp, especially when you consider how annoying the return stream is. Seems to be right though. The consenso strategies con1 and con2 derive the signal by combing the signals of all the other strategies and then following the resulting signal. I… Continue reading 10/8/20 – New Methodology
New Concepts in Garbage Cans
This is a new type of study. It shows 26 strategies and Buy and Hold sorted by Total. The natural log numbers can be transformed into current value of each $ invested. The single garbage can structure described previously can't handle this, so it now looks like this: 1 in a strategy column means long.… Continue reading New Concepts in Garbage Cans
9/21/20 – x00
There is a clear move down out of the 8 to 9 day congestion. Avoiding being long over the weekend wasn't real difficult despite the x00 buy signal. The decline has been relatively orderly if nothing else. These are the numbers that Specter/FF5 looks at over the last 8 years. Actually just E and M,… Continue reading 9/21/20 – x00
Moving Average Transformation Calculations
I decided to change the moving average calculation to use natural logs instead of the formula: (Weighted Price - Moving Average) / Weighted Price * 100. The returns are identical but the numbers are easier to understand with natural logs. Once the given moving average is determined, the new formula is: WorksheetFunction.Ln(Arg1:=(WPrice) / Mx(xw) *… Continue reading Moving Average Transformation Calculations
3x Significant 6 Bulls From 2018
This is something you won't see anywhere else. I've been looking at multiple strategies and thought this was a bug when I noticed ACC Total, the natural log buy and hold return, was almost exactly zero for this period. If an investor had invested equal amounts in these guys at the end of 2017, he… Continue reading 3x Significant 6 Bulls From 2018
9/9/20 Consenso x00 and Next Day
The four possibilities for x00 resolution are shown in the top table. xC0 does not occur in nature, the same as the other side, x3F. The only possibility for a long signal tomorrow is x00. x30 is the big disaster if trend following is not on. The table at the bottom says it all regarding… Continue reading 9/9/20 Consenso x00 and Next Day
9/8/20 Probable x00
x30 is the worst signal return, x00 is generally in the running for best. This is the output of the new signal processing engine. There are 4 x 6 columns = 24. xDS is the Consenso signal to make 25 for xFF5. Once you get here, it is possible to build the Single Garbage Can.… Continue reading 9/8/20 Probable x00
