Markets. Putting my bet out there.

Oh what to do. I’ve been pretty long tech for a while. The last week has been brutal though with my portfolio dropping 10% from highs. I’ve only just put on my hedges today. Sold short 2000 shares of QQQ as well as a few OTM puts for Feb and Apr.

The hedges saved me a bit of money today, but still not enough to offset the loss of my overall long tech position.

I’m sure we’ve all been curious about what capitulated such a sharp decline so fast. I feel like pundit discussions aren’t very satisfying. Inflation? rising wages? Hardly seem fit to justify  such a steep decline.

Maybe that data starting it, but I feel like something else must have magnified it to the drop we’re seeing. I have no basis of saying this, but I imagine it’s a lack of humans. I feel like the only humans exiting will be the ones that decided to lock in some profit after the run, but even they would only get spooked after sharp spiking of a couple risk-parity algos rebalancing for hiring vol and maybe even some stop losses getting triggered. XIV probably acted as the catalyst for that risk parity rebalance.

If that’s the case, there should be a few more days of bot reduction ahead of us. I’ve heard that instead of using just a spot VIX to determine Vol, they have a look back window. Who knows what the various players are using. 1 month? 2 months? 3? 6?. Whatver the case, the last 5 days of spiked Vol still are a minority in that average. As the low vol days drop out of the window, that increasing average might continue to necessitate more RP algos further reduce positions. As much as it pains me to “trade” against such strong fundamentals, I’ve got to place more short QQQ positions. I’ve played with various possible VIX look-back windows and have decided the following:

Risk Parity Play Planned on Feb 9

Hopefully, as the VIX spike falls off the look-back window, this means algos will add to their positions. Assuming no further terrible news, VIX should slowly decay back  to lower levels. That being said, until the end of February, the falling off of the lower VIX days of January from the look-back Window means that the average VIX for the algo will still rise even if the day’s VIX is dropping. Not buying the dip till March starts.

That’s my call. put it out there. Let’s see how it goes.

Short till Feb 26

QQQ Calls for the Open of Feb 27, exit at close.

Start buying dip at March 1.

Let’s see how dumb I look.

Yearly Deaths Amended Table

Hi all,

I’ve taken some suggestions from the comments received on the last one and amended to the following table:

  1. I’ve removed the month detail so it’s less cluttered a wall of numbers and reduced things to their whole year totals / averages
  2. I’ve included a normalised value for age which is its ratio to the life expectancy in the USA for the year. I understand that the deaths are global, but this seemed a simpler proxy
  3. I’ve included a normalised value for the death count relative to the population of the world for that year.


2016 Deaths. How do they stack up?

I was curious to see whether 2016 was really as bad as it felt in terms of notable deaths and created a heat-map of notable births, deaths, and average ages of deaths as indicated on Wikipedia’s various year pages from 1900 to 2016.

BoarderFixed 1900-2016 Age - Deaths - Births.png
In looking at the first third, we see that average age of notable death is going up which seems in line with the increasing life expectancy over the years.
Notable deaths were pretty sparse in the early half of the 1900s, but one has to wonder if this just a function of how sparse information was for the period and how difficult notability would be to garner at that point in time.
While the latter half of the century had more deaths, by the early 2000s, things seem to decline.
With that decline in mind,  that’s probably why this year has seemed so terrible in recent memory
Looking only at 2000 – 2016, notably deaths seem to be drastically increasing.
Something that might stand out to you is 2011s drastically low average age of death. A close up of the 2000 – 2016 period further highlight the tragedy of September 2011.
On September 7, 2011 a plane carrying a Russian Hockey team crashed killing all players on board.
The final section, “Notable births”, seems to generally increase over time, possibly again due to the easy of gaining notability through quicker spread of information. Unsurprisingly, there are very few notable entries for those born after 2000 due to the fact that these individuals have yet to have live enough of a life to do anything notable.
Those born in the 90s are starting to turn up on the wikipedia year summaries, but there is a strangely quiet space in 1994.
Other interesting things to note are the general trend across the months for death and age.
The general idea that winter months tend to be harsh on the old seems to echo in the data with January and February showing higher deaths as well as higher average age of death for those months.
So how unusual was 2016?
I decided to restrict the look back to what I would consider people’s recent memory. The early half of the 20th century hardly seemed appropriate.
From a simple chart, 2016 doesn’t seem too far out there, particularly standing next to 1979.
It is still almost 2 standard deviations away from the mean for the period though.
When we further restrict the window to exclude the tumultuous 70s and 80s, it becomes even more unusual.
Finally, for the 2000s, it is incredibly unusual.

Count ’em Up: Criteria Fitting Guys

I combined income and height distributions to the marital status data from the last post to create a filter to estimate how many unmarried men exist in Hong Kong that satisfy age, income, and height criteria. Please don’t rip into me too hard, I don’t by any means assume this is an actual, exhaustive list of filters a woman would use for evaluating men. It’s just for fun. 

How to Use:

Due to how census data brackets ages and income, you have to select from the drop down menu the minimum and maximums for age and monthly income.

I pretended height was approximately normally distributed and just used some mean and standard deviation online – you can enter any number for this field.

Anyway, the bottom result is the joint probability of height and income filters applied to the subset of single men in the age range. So it’s:

(# of single men in age range) * (% of people in income range) * (% of men in height range)


Number of single males in HK that satisfy the specified age, income, and height criteria

Big caveat: there’s a lot of reason to believe that some of this data is not independent. Higher salaries will probably correlate with higher ages, even height is said to correlate with higher income. I’m not really going to delve into those depths. I treated them as independent. Again, it’s just for fun.

Also, I decide to throw in the figure for single women of an age less than the maximum male age stated (and above the age of 20) just for reference. Again, I know height, income, and the man being older aren’t set in stone rules for dating, but they are at least joked about as conditions. 

Count ’em Up: Guys vs Gals

There was a funny exchange that went on last week regarding an article written by Alan Lee on EJ Insight titled “Hey sisters, get real or die alone”. It drew a few responses due to its general sexist tone, most notably by an article by Laurel Chor at HK Coconuts. While the original article from EJ Insight seems have been taken down, the HK coconuts page preserved much of the hilarity, particularly with a few snapshots of FB comments in which the Alan tries to defend his original article along with himself.

Anyway, I’m not here to take a side on that little drama, I’m here to comment on some numbers brought up by Alan’s original article:

“the proportion of women to men in the city was at 1000:856 last year, compared with 1000:904 in 2013.”

I gave the 2015 census data a quick check and got the following numbers:

Male: 3.370,100

Female: 3,954,200

Or roughly, 1000:852 🇭🇰 women to men. Seems close enough.

The thing is though, that ratio is for all of Hong Kong’s entire population, at every age range. 

The first way this skews the data is that women naturally live longer than men, thus the population at the older brackets will be weighted heavily female even though I don’t think that this segment is particularly relevant for examining the dating and marriage market.

We should probably further narrow to exclude people under 20. While 18 probably should be added to the pool, the data is cut into buckets of 5 years; I figured I definitely wanted to exclude 15 year olds from the dating pool that goes all the way up to 39.

All in all, I confined the population to be examined to be between the ages of 20 and 39 (I don’t mean to offend anyone outside this range, it’s kind of a line drawn in the sand by the original article). 

When we confine to this population, we get the following split:

Male (aged 20 – 39): 903,000

Female (aged 20 – 39): 1,195,400

Ouch. That looks even worse. That’s a female to male ratio of 1000:756 

Here’s where it turns around though.

That’s just the raw population counts, if we look at the marriage numbers for that same range, we have:

Married Males (aged 20 – 39): 292,400

Married Females (aged 20 – 39): 543,400

A lot of females have been taken out of the population. This yields an implied single count of:

Single Males (aged 20 – 39): 611,100

Single Females (aged 20 – 39): 652,000

1,000:937 single females to single males. Not as bad right? That’s still 40,900 females hanging though. Will it get worse? Let’s try and guess…

New singles in the 20 – 39 range don’t often come out of the blue, a lot of them come from… people in the 10-19 age bracket. What does that look like? Well, it’s a lot more Male heavy. 

(This next part is a caution to guys in their late 20s and early 30s thinking they will wait 5 years and catch a young 20 something girl.)

The 10 – 19 age range is:

Male: 311,300

Female:  294,000

While not a terribly earth shattering shift on it’s own, sliding the population table 10 years (ignoring immigration) and taking into account very negligible mortality differences at this point, we get this ratio:

Year 2025: (20 – 39 year old) Female to Male Ratio – 1,000:940

Just as a reminder:

Year 2015: (20 – 39 year old) Female to Male Ratio – 1,000:756

2025 is looking way better for the females. What’s more tilting though is that if we keep the current percentage of singles for each gender-age category constant and reapply it to this new aged population in 2025 (remembering that large proportions of men stay single in this age range), we get the following ratio:

Year 2026 [20 – 39 years old, Unmarried] Male to Female Ratio 

1,000 Female : Male 1,160

Not to detract from that punch line, but if men really are the “fine wines” that remain eligible until the 40s. I really should be including the single men at the 40-49 bracket in this, further heightening competition amongst the men for limited women.


Despite all that though, can anyone really get complacent with those numbers? Men in this decade or women in the next? Our individual lives aren’t statistics. It’s not like dating is ranking and matching the entire population of singles until you run out of one gender. You don’t get a date as the shittiest of 600,000 just because the other gender has 600,001.

Hello World

I’m not sure with what to populate this blog. Last one I did was some quantitative analysis on random things. Maybe some travel stuff for this one? Maybe The renovation that’s currently taking up my time? Maybe all. Oh well, wait and see.

Thanks for stopping by.