Anatomy of a Market Beater: Part VI
In this series, I’ll be exploring whether there are any useful characteristics that separate stocks that beat a given benchmark index (“winners”) from those that lose (“losers”).
Suppose the benchmark index you are using is the NASDAQ.
Also, suppose that we are comparing stocks against the NASDAQ over a period of 365 days (the “test period”).
I ran 1,000 randomized trials on existing NYSE and NASDAQ stocks whereby for each trial: (1) a random date was chosen, (2) every existing stock was compared against the NASDAQ index over the period beginning with that random date and ending 365 days later. The stocks for each trial were grouped into either “winners” or “losers” depending on whether a stock’s overall growth over the 365-day period was greater (“winner”) or lesser (“loser”) than the NASDAQ.
The winners and losers were compared by days a stock would be beating or losing against the NASDAQ, the amount of that gain or loss (margin), and the absolute gain or loss that stock was experiencing.
Here’s what I found.
VI. Progressive Marginal Gains of Winners versus Losers
If we map out the progression of median marginal gains of the winner stocks versus the loser stocks at each point in time along the investment period from day 1 to day 365, we would get something that looks like this:
No surprise here as stocks that beat the market overall tend to do progressively better than those that don’t day by day.
Remember, this is an aggregation of 1,000 trials, so each line above is the median of 1,000 trials of median marginal gains for each group of winners and losers. In other words, on each day, in each trial, each winner’s marginal gain is calculated. The median marginal gain out of all the winners of that day of that trial is calculated. Since there are 1,000 trials, there are 1,000 samples of that median marginal gain. So the lines above represent the median, median marginal gain of a winner or loser stock respectively.
Because Figure 1 above is an aggregate of 1,000 trials we’d probably want to know how variable the figures of that aggregate are. In essence, how reliable is Figure 1?
Let’s first look at maximum and minimum figures:
Here is the same graph as Figure 1 except this time the winners’ cumulative marginal gains (blue line) are bounded below by their minimum cumulative marginal gains (green line). The losers’ cumulative marginal gains (red line) are bounded above by their maximum cumulative marginal gains (yellow line). Reminder, my analysis is an aggregation of 1,000 trials, so for example, “minimum cumulative marginal gains” means the sample out of 1,000 that had the lowest median marginal gain of a winner group.
How reliable or stable is Figure 1?
The blue and red lines represent the spread of marginal gains within the winners and losers groups respectively. (Note because this is an aggregation of 1,000 trials, “spread” really means the median spread over those 1,000 spread samples). Losers tend to have a lower spread in marginal gain figures amongst their ranks while winners tend to have higher variability. As the days progress, that variability increases for both groups but at a higher rate for winners than losers.
The green and yellow lines represent the spread of median marginal gains of the winners and losers respectively over 1,000 trials. In short, they represent the spread of the lines in Figure 1 above.
What this says is (1) that identifying whether a stock is a winner or loser based on its marginal gain on a given day is not very wise because their spreads are so high (see Figure 3, blue and red lines) — I mean look at the range of possible marginal gains any given winner or loser stock might experience:
Winners and losers can experience nearly the same types of marginal gains as the other save for the last day, as that is the day that defines winner from loser: if you have a higher cumulative gain than the NASDAQ by period’s end, you are a winner, if not, you’re a loser.
…and (2) that identifying whether a group of stocks are winners or losers based on their collective median marginal gain is not that bad of a gamble because the spread is rather narrow (see Figure 4, yellow and green lines) while their regions are divergent (see Figure 1).
Here’s Figure 1 again but this time each line is bounded by lines representing +- 1.5* median absolute deviation of their medians:
If a group of stocks you have is experiencing X median marginal cumulative gain over NASDAQ on day Y, and X exists outside of the winner region bounded in Figure 6, then you should consider the possibility that at least one stock in your group is likely not going to be a winner by year’s end.
The point at which these two regions diverge (where the orange line and green line stop overlapping) is day 105 where the median marginal gain for a loser plus 1.5 times its median absolute deviation (MAD) is -0.25% and -0.24% for a winner’s median marginal gain minus 1.5 times its MAD.
If you like my content please follow me on all my socials. Donations are accepted too! It’d really mean a lot. (See my Linktree below)
WARNING: By using or reading my content in any way, you acknowledge and agree to the following: I am not a registered investment adviser nor am I licensed or qualified to provide investment advice. My content is for informational purposes only and is not advice (financial, investment, or otherwise). You should never invest based solely on my content. You should assume my content is not trustworthy unless verified by your own independent research. None of my content is a guarantee that you will make a profit if you use my content. By reading or using my content in any way, you do so entirely at your own risk. I am not nor shall I be liable to you for any damages or losses you incur directly or indirectly as a result of or arising from the use of my content.
