Pair Trading: Get Rich by Picking One Stock to Beat Another

3 min readNov 14, 2019

Because obviously Amazon > Best Buy

The funny thing about playing the stock market is that there are tons of strategies that seem like slam dunks, yet few (if any) that actually are.

Take for instance “pair trading.” All you have to do is find one stock that is likely to outperform another, buy the one you like, short the one you don’t, and you get the difference.

(To short a stock, you need a margin account and typically pay a high interest rate on the shares you borrow— but that stuff’s boring and I don’t really understand it so moving on.)

So our task is simply to find one stock that seems likely to beat another. Let’s see, what’s going on in the world today…

People aren’t using cameras

Apple’s the biggest smartphone company, so why not bet on Apple outperforming a traditional camera company. Let’s go with Nikon.

Well that worked! Apple +70%, Nikon -3%, that means we profit 73% in less than a year.

People are binge-watching Netflix

Another natural idea is to buy Netflix and short one of the big TV networks. According to Wikipedia the big 3 networks are NBC, CBS, and ABC, and of those CBS has the cleanest standalone stock (NBC is part of Comcast, ABC is part of Disney).

All right, we’re 2 for 2. Netflix +6%, CBS -18%, we made 24%.

People are flying more, driving less…taking trains even less?

Everyone loves Southwest so let’s use that for the airline. Pitting planes against trains seems like our best bet, so I Googled “biggest train company” and came up with Union Pacific. Here we go:

Southwest +23%, Union Pacific +30%, we lose 7%.

Conclusions

Railroads are making a comeback

Only explanation for Union Pacific outpacing Southwest.

2 trades worked, 1 was lucky

Apple beat Nikon, but we never could have predicted Apple growing 70% in a year. That level of growth was clearly not primarily due to people taking more pictures on their phones.

Hard to beat the market, especially after fees

Remember, the money tied up in a pair trade is money that could be sitting in an index fund, growing, say, 10% a year. If that’s our benchmark, we don’t just need one stock to beat the other, we need the difference to be considerable — even after paying margin interest.

So far in 2019, the S&P has gained 25%. That means only 1 of our 3 pair trades actually beat the market.

Slam dunks are already priced in

If you’re betting on an obvious winner and against an obvious loser, the obviousness of it is already going to be reflected in the share prices. Sad!

R code

# Install and load 'stocks' package
devtools::install_github("vandomed/stocks")
library("stocks")# Pair trade #1
plot_growth(tickers = c("AAPL", "NINOY"), from = "2019-01-01", 
            title = "Apple vs. Nikon")# Pair trade #2
plot_growth(tickers = c("NFLX", "CBS"), from = "2019-01-01", 
            title = "Netflix vs. CBS")# Pair trade #3
plot_growth(tickers = c("LUV", "UNP"), from = "2019-01-01", 
            title = "Southwest vs. Union Pacific")# Percent growth for each fund and the S&P
calc_metrics(
  tickers = c("AAPL", "NINOY", "NFLX", "CBS", "LUV", "UNP", "SPY"), 
  from = "2019-01-01"
)