Quick And Random Stat Update
So here is a random stat that I came upon today.
*** update added VIX sell signal stats way below***
So the first quant math is how many times has the market not made a new low? Or how many times has the market closed higher than the previous day’s low consecutively. So how many times has the market not closed below the previous day’s low 19 times?
The stats do not have much of an edge (bullish or bearish)- I just though it was interesting that we got to 19 times. That being said, one way to interpret this data is that the market (since 1990) has only gotten to 22 consecutive non-lower closes 3 times. Or in English we are dealing with very very low odds of not having a lower close.
I will try to do stuff like this periodically as it literally takes me 8 seconds to run these stats. So for clartiy here are some explanations of the stats.
First off, all of the math excluding the “Ratio” presumes a long trade.
% positive= number of occurrences that the position made money on the long side out of the total- using the specified criteria.
—– Careful using this stat as if you win 7 out of 10- things look good, but does not mean a that it is a good idea. That is the reason most quants look a at skews and sharpe ratios. I prefer average max gain to average max loss. In any case “expected value” is much better than % positive. Many get caught in that scam/trap.
Mean= Average return for going long for the specified criteria. Mean simply means “expected return.” If it is negative- that is good for someone looking to go short.
Average Max Gain= The average of the max gain for a long trade for all occurrences that meet the specified criteria in the past- many would call this expected “run up.”
Average Max Loss= The average of the max loss fora long trade for all occurrences that meet the specified criteria- many would call this expected “run down.”
Now my favorite is the Avg. Max Gain/ Max Loss ratio. It simply means is my typical expected max run up better than run down. My calculator calculates the higher of the two and gives a risk to reward ratio. So here the max loss is greater than the gain- this spits out the risk to reward ratio of going short. Here it is about 2:1 over the next week, yet it is +2:1 going long over the next 2 week period. I presume that means a dip then a bounce is expected…. Again, there is not much of an edge here – well at least using this variable on a stand alone basis.
Skew simply means is there a tendency down or up- a negative number means downside skewness. As a trader I love to see a negative mean (negative expected value) and a negative skew. That implies the market has huge downside tails (great for lottery OTM nearby puts). Generally a -1 or +1 is significant- anything between -1 and 1 is “normal.”
Kurt means kurtosis or fat tailed. Kurt really means that the market has lots of little moves and infrequent, but huge large moves (up or down). Kurtosis only matters if it is greater than 2. If it is near 4 then it is “fat tailed.”
I hope this helps clear things up and if you have questions I will try to resolve them tonight or tomorrow.
*** ok here is another stat for you all. Here is the VIX sell signal with some extra sauce***
Ok notice that this is 31% positive in 1 week with a 12:1 Short sell ratio. Or average short selling max gain is 2.6% with a whopping 0.2% margin call. So I guess you know my position. So do not go bashing the VIX sell signal. I admit I have a small position- though I am long FEB VIX futures and short APR VIX futures (aka a bull spread or short VXV/VIX ratio ish) as well.
Here is the raw data for further inspection. 1-2 weeks is impressively good odds of more downside than upside in OHLC (open high low close) drawdown.