“When I was back there in seminary school / There was a person there / Who put forth the proposition . . .”
— Jim Morrison, The Doors, “Soft Parade”

The body of this piece was originally going to start with the phrase “When I was back in grad school,” which made me think of the beginning to the “Soft Parade,” because when I was back in middle school I had a slight Doors obsession, which is totally normal for an 11-year-old ambivert growing up in the 1990s in a small Texas town just outside of Fort Worth. Totally normal.

Anyway, when I was back in grad school pretending to work on a Ph.D. in English I had a couple of hobbies that occupied my attention:

  1. Fantasy sports
  2. The stock market

This piece is about applying one indicator from the stock market — the moving average — to daily fantasy sports.

What Is a Moving Average?

In statistics, a moving average (MA) is a series of averages based on a set number of successive units of time. For instance, if I took the separate noontime temperatures in Cedar Rapids, Iowa, for the last 50 days and averaged them all together, the resulting number would be the 50-day MA for noontime temperatures in Cedar Rapids. With each advancing day, the oldest day in the data set would be removed, the most recent day would be added, and the 50-day MA for noontime temperatures would be updated.

Eventually (if we wanted) we could plot the 50-day MA and the actual noontime temperatures on the same graph. When the weather would warm, we’d expect the temperature to be above the 50-day MA. When the temperature would cross below the 50-day MA, we might expect a ‘downturn’ in the weather with cooler weather prevailing.

A few thoughts:

  1. What I just described is a smidgen (very?) unscientific.
  2. With a few tweaks, a big database of information, and the right people analyzing the data, MAs of various lengths when used combinatorially can be useful for the identification of trends. For instance, MAs are what the archmaesters use to determine that winter has finally come at the end of the sixth season of Game of Thrones. That’s extratextual, but it’s my suspicion. The chain-wearing maesters in an HBO fantasy drama series seem like the type of people to care about facts and data.
  3. A lot of people in the stock market — technical traders who often don’t care much about company fundamentals or particular catalysts — tend to use MAs as indicators about when to buy into or sell out of positions.

I can see you’re enjoying this conversation, so let’s keep it going.

Moving Averages in the Stock Market

Since I’m an up-to-date person who uses words like “hip” and “gnarly,” here’s a 2005 price chart (including the 50-day MA) for the Nasdaq 100 Trust — an exchange-traded fund (ETF) that basically tracks the 100 largest non-financial stocks traded on the Nasdaq Stock Market. Sometimes called “cubes,” the ticker symbol “QQQQ” has since been shortened to “QQQ,” and the ETF is now known as the PowerShares QQQ Trust.

Anyway, here’s its 4/29/05 price chart and 50-day MA:

If I were an amateur technical trader — and I’m not — I’d draw your attention to the points in the chart where the price (in red) moves above or below the 50-day MA (in blue). In late May 2004 QQQQ breaks above the MA before dropping below it in early July. After breaking down, QQQQ continues to drop by double-digit percentage points before rebounding and crossing above the MA in early September. From there QQQQ moves from about $34 to $40 — with all of that movement coming above the MA — and then QQQQ eventually drops again, crossing below the MA in January 2005 and (for the most part) staying below the MA as the price dropped from around $39 to $35.

You can see the vague outlines of a trend here: When the price crosses below the MA, it tends to drop further. When the price crosses above the MA, it tends to rise further. There’s not an exact science to this — but you can clearly see that, although there would be inefficiencies in a trading system based solely around this MA, such a trading system could have limited an investor from the downswings in QQQQ in 2004-05 and still enabled the investor to capture most of the benefit of the upswings.

If someone had bought QQQQ in late April 2004 and sold it in late April 2005, the person would’ve bought around $35 and sold around $35 for no profit. If, though, the person had throughout the year regularly bought QQQQ shortly after it crossed above the MA and sold shortly after it crossed below the MA, then the investor would’ve booked about $5 of profit from trading into and out of the stock strategically.

That basic concept — trading in and out of a position based on the MAs (and other technical indicators) — is what a lot of stock traders attempt to do with the market in general.

A Quick Aside That Might not Be Quick

Not all MAs are created equal. They all function differently, and some are likely more predictive than others. They can differ based on duration and also composition.

For instance, the 50-day MA is more of a short-term tool, while the 200-day MA is a long-term indicator. On top of that, there are MAs for basically anything that can be quantified over a period of time — and these MAs can be calculated in various ways, with either all units of time weighted equally (simple MA) — which is what we’ve been dealing with so far — or the more current time units weighted more heavily (exponential MA). I suppose there could also be an EMA that overweights the older time units, but that type of MA wouldn’t make a lot of sense in most contexts.

The point is that not all MAs are the same. Each of them has potential uses and drawbacks and should be employed accordingly. In our Player Models we often break data down into de facto MAs based on the last year and last month of fantasy production.

Right now it’s the morning of Thursday, 3/23/17. This piece is coming out on Friday, but I’m getting a head start on it because I plan to kick off my epic weekend with an all-day Friday pre-party of relatively uninterrupted reading. So I’m looking right now at our Phan Model for the five-game NBA main slate on Thursday night. If I scroll all the way to the right, I can see historical production data for the last fantasy month and year:

People might not look at these numbers and think, “Hey, cool, those are a bunch of short- and long-term SMAs,” but that’s what this data represents.

Big picture: Almost any data point that updates on a trailing basis can be contextualized as a moving average, and sometimes it’s useful to think of these data points not as individual specks on a graph but as lines against which current fantasy production can be measured.

And here’s an aside within my aside: In investing, DFS, deciding what to eat, whatever, no one should use just one indicator. Someone who trades the market based solely on one MA is someone who probably won’t be trading stocks for long.

A Quick Anecdote That Might not Be Quick

When I was back in grad school and dabbling in the stock market, I used to read articles and watch videos on a number of so-called advisory sites to get ideas.

Some of these sites were so spammy and dispensed such sh*tty information that I actually considered tracking the stocks they touted, waiting for them to shoot up, and then shorting them as they inevitably dropped. I didn’t do it because I’m 1) a wimp and 2) fundamentally opposed to the mechanics of shorting, since the practice offers theoretically unlimited downside and capped upside — but, still, I’m somewhat convinced that private investors more sophisticated and dedicated to monitoring their protfolios than I am could make this strategy work.

Anyway, you didn’t come here to read about how I decided to bypass millions of dollars (probably).

So let’s get to another anecdote you don’t care about: I still get a couple of free financial newsletters, podcasts, etc., emailed to me each week, primarily because I’m too lazy to unsubscribe from all of them. My wife — the one who thinks my British accent sucks — was out of town this last weekend, so on Saturday morning I saw one of these emails in my inbox and thought, “Watching that videocast is probably more productive than consuming live performances of ’90s bands on YouTube,” so I watched it. Also, “live performances of ’90s bands on YouTube” might be euphemistic.

In the videocast, a guy named R.C. Peck (who seems less shystery than most independent advisers) discussed a section of Unshakeable, the new financial book by motivational speaker and writer Tony Robbins. A word of caution: I haven’t read the book. I’m relying on Peck’s characterization and report of the book.

In Unshakeable, Robbins (reportedly) presents data indicating that over a long period of time a huge percentage of the market’s total upward movement is the collective result of the gains made in the market’s 10 best days over that time. According to Peck, Robbins uses this information to argue that people should use a buy-and-hold approach to the market. Otherwise, they are at risk of missing the market’s most substantial gains.

Of course, what Robbins doesn’t note — and what Peck does — is that if over that same period of time an investor missed the 10 worst days of the market (and even the 10 best and 10 worst days) — the investor would do significantly better than if (s)he had used a simple buy-and-hold approach.

[Random DFS Application: Don’t have a buy-and-hold approach when it comes to investing in DFS assets. Don’t roster a guy simply because “he’s a stud.” Be judicious by investing in players when they have significant indicators in their favor.]

And then Peck presented a price chart of the S&P 500 since 1960 with a 200-day EMA and pointed out that 83 of the market’s 100 worst days in that time occurred when the price was below the EMA.

In summary, MAs matter, even if they’re relatively straightforward.

Some DFS Thoughts

I’m not 100 percent sure that the idea of MAs is applicable to DFS, but I’m fairly confident it is, especially for positive expected value FantasyLabs subscribers who research with our Trends tool.

For instance, here’s what FanDuel pitchers have done when they have a ’12-month SMA’ of 4.0-4.9 strikeouts per nine innings:

And here’s how a subgroup of those same pitchers has done when ‘above the MA’ with a K Prediction of 5.0-5.9 K/9:

The cohort is small, but I picked these ranges out of thin air for the purpose of making a point. The Plus/Minus and Consistency Rating are much higher than those of the baseline group, and the ownership is almost nonexistent. (By the way, developments in tournament ownership across various stakes can be monitored via our DFS Ownership Dashboard.)

It took me about two minutes to create these trends. I could probably do more MA research with the Labs Tools across various sports and find similar results.

I don’t want to suggest that DFS is as simple as staying away from players below their short- or long-term averages and rostering players above those averages — but there probably aren’t many cash game heuristics better and more intuitive than that.

Again, I’m not telling you that you should employ this strategy. I’m telling you to be your own DFS general, research the idea, and determine if it can work for you.

The Labyrinthian: 2017.29, 124

This is the 124th installment of The Labyrinthian, a series dedicated to exploring random fields of knowledge in order to give you unordinary theoretical, philosophical, strategic, and/or often rambling guidance on daily fantasy sports. Consult the introductory piece to the series for further explanation. Previous installments of The Labyrinthian can be accessed via my author page.