27 Apr, 22 tweets, 7 min read

You have \$1,000
You leave it alone
The volatility of SPY over the period was 18%

What is the volatility of your portfolio?

Not a trick question. It's 18%

1/n

At the start, you have half your money in cash and half in SPY.

What is the volatility of your portfolio now?

It's 9%: half what it was before.

2/n
Now, let's say you could buy \$2000 of SPY in your \$1000 account (and don't pay anything to borrow)

What is the volatility of your portfolio now?

It's 36%: twice the figure when you were fully invested.

This is a useful result. You can prove it to yourself easily in Excel

3/n
If you half the size of your position you get half of the volatility contribution.

If you double the size of your position you get double the volatility contribution.

This is very useful when it comes to *sizing positions*

4/n
Asset volatility is quite easy to predict.

And here are some scatterplots to illlustrate.

I've plotted annualised volatility over 20 days against the vol over the previous 20 days.

(estimated from the standard deviation of returns)

5/n
Simply assuming volatility stays the same as your last estimate of it works pretty well as a forecast.

Just like the weather.

This, and the fact that volatility increases in linear proportion to size, suggests a simple approach to "targeting" a certain level of volatility.

6/n
If you want a given position to contribute 10% volatility to your portfolio.

You can:
- Observe the vol it contributed over the last 20 days (15% say)
- Scale its sizing by vol_estimated / vol_you_want:

So you'd scale the position up 15 / 10 = 1.5

7/n
Why would this be a useful thing?

Why would you target a certain level of volatility?

Imagine you have two assets:
- a volatile orange asset
- a less volatile yellow asset

8/n
If you hold these assets together with equal size.

The portfolio returns are going to look a bit like the black line here.

It will be dominated by the volatility of the orange asset.

9/n
Is this what you want?

Probably not, right?

You're allowing our portfolio to be dominated by the most volatile asset, simply because it happens to be the most volatile asset.

10/n
Unless you have a good reason to prefer one asset over the other, you'll want each stock to contribute about the same amount of volatility to your portfolio.

You want the movements in your portfolio to be equally dependent on both assets. Probably.

11/n
To do that, we'd buy more of the yellow one - intentionally making it more volatile in the context of our portfolio (than when we equal-weighted it)

And we'd buy less of the orange one - intentionally making it less volatile in the context of our portfolio.

12/n
If you scale yellow up and size orange down to targe equal vol... it would look something like this:

The portfolio (black line) is less volatile than the constituent stocks. This will always be the case as long as they don't wiggle in sync.

Diversification 101

13/n
One objection you may have to this example is:

"Why would I give them equal volatility weight? The yellow one is better"

Yeah, but only in the past... We have no idea what's going to happen next.

Predicting returns is super hard. At least in the future.

14/n
Now if you're convinced this is a good idea, you already know how to do the scaling, cos I told you earlier...

But let's go thru it cos repetition is good...

We'll assume:
- orange shows 30% vol over the charted period
- yellow 10% vol

15/n
Remember volatility scales in proportion to size?

Given a \$1k account...

At \$1k we realize 30% vol
At \$500 we realize 15% vol
At \$250 we realize 7.5% vol

16/n
So let's size "orange" to 7.5% vol contribution by buying \$250 of it in our \$1k account.

17/n
Now let's do yellow.

Given our \$1k account...

At \$1k we realize 10% vol
At \$750 we realize 7.5% vol

So we size "yellow" to a 7.5% vol contribution by buying \$750 of it in our \$1k account.

18/n
Now, assuming our volatility "predictions" were reasonable, we can now expect both assets to contribute about 7.5% volatility each to our portfolio.

And, to the extent one zigs whilst the other zags, we'd see portfolio vol to be less than the sum (<15%)

19/n
This is a really useful sizing technique. And it's useful to think in these terms.

Managing volatility can also increase your risk/adjusted returns. Because although volatility is linear in size, compounded returns are not.

A discussion for another time...

20/20
I took these examples from this simple retail-focused quant trading course I'm teaching here: robotwealth.com/trade-like-a-qβ¦
Tweet 7 is the wrong way around here. Thanks @cyberSM7. Should say...

You want to target 10% vol

You can:
- Observe the vol it contributed over the last 20 days (15% say)
- Scale its sizing by vol_you_want / vol_estimated:

So you'd scale the position up 10 / 15 = 2/3

β’ β’ β’

Missing some Tweet in this thread? You can try to force a refresh
γ

This Thread may be Removed Anytime!

Twitter may remove this content at anytime! Save it as PDF for later use!

# More from @therobotjames

28 Apr
- Buying from someone too cheap
- Selling to someone too expensive

At least on average.

To do this, you need to know who you are playing against.

π§΅on "edge", where to find it, and how you can compete π

1/n
If you are a market maker, it is relatively clear to understand who you are trading against.

If you're a positional trader, it is perhaps less clear.

On a trivial level, you're probably trading with a market maker.

2/n
But understand that "the market line" is set by the supply/demand pressures of other aggressive traders.
- End users (wealth mgmt, retail)
- Aggressive prop traders doing short term risky arbs

3/n
22 Apr
Tips for doing financial analysis with OHLC bar data.

Many of you doing quanty analysis with OHLC bar data.

Here's some boring but crucial stuff you need to understand if you're doing that. πππ

1/n
An OHLC bar represents a summary of trades that happened in a certain period.

Open -the price of the first trade in the period
High - the highest price traded in the period
Low - the lowest price traded in the period
Close - the price of the final trade in the period

2/n
For daily stock data, the Close price will be the price arrived at in the closing auction.

This is set by balancing the supply and demand of MOO (market on close) and LOO (limit on close) orders to maximize the amount of stock traded.

3/n
20 Apr
In Australia, if you're serious about getting the job done effectively and efficiently, you might say:

"I'm not here to f*** spiders"

Many traders act like they are, indeed, here to f*** spiders.

1/n
If you're making soup, you first need a good stock.

Stock isn't exciting. Everyone has stock.

Garnish is exciting, but you can't make soup from just garnish.

You need at least one reliable, stonkingly obvious way to get paid

2/n
Here's a non-soup analogy...

If you start a business venture, it's clear that you need an obvious, reliable way to make money.

You wouldn't just try to blag it.

"I am smart and hard-working" is not a business case.

You need a stonkingly obvious way to get paid.

3/n
19 Apr
In my 20 years of trading I have noticed this cycle play out again and again with traders that "make it":

1. Overconfidently reach for returns
2. Get humbled by the market
3. Simplify + concentrate on clear, high probability edges.

1/n
Nearly everyone starts with a lack of respect for how hard it is to consistently make money trading.

That leads them to pass over high-probability sources of returns in favor of more marginal ideas.

Or they overcomplicate the trading of a good edge.

2/n
Here's an example...

It is 2015. You look at a simple strategy.

You hold an equal dollar exposure to:
- Cap Weighted US Stocks (VTI)
- 20+ Yr US Treasury Bonds (TLT)
- Gold (GLD)

And rebalance each month.

(I've extended back a bit with mutual fund prices.)

3/n
14 Apr
A common mistake is to make implementation decisions or parameter choices based on "what improves the summary performance of a backtest".

Quant research is not "changing random stuff and picking the best performing backtest"

1/n
A backtest is a very complicated thing.

In the best case, the cause -> effect relationship between what you are changing and the performance of the backtest (say) is highly non-linear.

More likely there is no clear relationship.

2/n
Quant trading is not "changing stuff until you get a backtest you're happy with".

You need to split what you're doing into small component chunks and model those chunks as best you can.

3/n
11 Apr
It's easy to lose money trading if you:
1. Trade too much (paying fees + impact on each txn)
2. Size positions too big (high vol hurts compounding ability + gets u rekt)
3. Shorting positive drift/risk premia

It's hard to lose money consistently if you avoid these things.
However clueless you are, you get to trade at market prices.

Imagine we can know that an asset has a fair value of \$100.

You might think it's worth \$150.

But if it's quoted \$99 / \$101, you can buy now at \$101.

You were totally wrong but you still bought close to fair value.
The same mechanisms that make it hard to get an edge also make it hard for you to trade at really bad prices.

In a simple model, you might say that prices are set by:
- (risky) arbitrage and relative value in the short term
- pricing/valuation models in the long term