NBA Playoffs: Zig Zag Theory - SDQL Tutorial and Analysis

Saturday, April 18. Next weekend. We'd be watching the start of NBA Playoffs and winning plays if we weren't living in this messed-up, bizarro world. My heart goes out to all that are affected by this nasty virus and I hope all of you are staying safe. I am thankful for my health and safety; however, I am still saddened that we had to cancel a boys' trip planned with my best friends to visit the casino for the weekend to play cards, golf, and watch some NBA Playoff basketball. In a recent interview, Adam Silver has expressed that the NBA does not expect to make any decisions on potential returns until May 1, at the earliest. There is so much uncertainty, but I am still holding out hope that something can be worked out to finish this season.

This is the first NBA season where I have the SDQL knowledge and database in my arsenal, so I was eager to do some research on NBA Playoff trends and be ready regardless if it happens or not.

A general note: when sports finally return, I think we as handicappers will need to proceed with caution when analyzing these games. There have been so many external factors that have affected each one of these athletes, teams, fans and oddsmakers. Will games be played in empty arenas? Will players be as motivated? I think trends and situations will still be useful, but you may need to re-evaluate how much weight it carries in your analytical processes. For my post today, I will tackle is a common, popular NBA playoff betting strategy: the Zig-Zag Theory. Most seasoned NBA playoff bettors are very familiar with this strategy and the system has been around quite a while.

The basic premise is this: in a playoff series, bet on a team to cover the spread following a loss in the same series. The idea behind this is the losing team will be one game closer to elimination and will come out the following game more motivated and play with more intensity. Also, perhaps recency bias will force oddsmakers to favor the previous winner in their price as the unknowing public may want to ride the hot team.

I’ve blindly used this theory to both success and failure, but I wanted to use SDQL to see what variations of this theory has truly worked. I’ll extract data from 2011 for purposes of this analysis. This gives me a decent sample size while not going way back in time. This season was the first championship the Lebron-led Miami Heat won.

The big question: is the Zig Zag Theory profitable?

Let’s build a basic SDQL query to see:

SDQL Query: playoffs=1 and season>=2011 and p:L and p:round=round

Follow along here: https://killersports.com/nba/query?sdql=playoffs%3D1+and+season%3E%3D2011+and+p%3AL+and+p%3Around%3Dround+&submit=++S+D+Q+L+%21++

Tutorial on parameters:

• playoffs=1 : “1” indicates it is a playoff game. “0” indicates a regular season game.

• season>=2011: indicates we are pulling data for seasons 2011 and beyond.

• p:L: “p” looks up the previous game. “L” means a straight-up loss. Combine this together and you are looking for games where a team lost their previous game SU.

• p:round=round: again, “p” looks up previous game. This tells you that the round of the previous game is the same round of the current game you are querying. The intention is to bring in games from the same series. If you do not include this, the query will bring in Round 1/Game 1s of teams that have lost their last regular season game.

The results? The ATS Record is DEAD EVEN. Nothing special at all.

But, this query is rather broad. Let’s add in “series game” at the end of the query to pull in how losing teams bounce back in a particular game in a series. As you can see, this theory seems to work better earlier in the series as performance really drops off in Games 6 and 7. At this point in the series, teams already have their back against the ropes. The loss from the last game really gives their opposing team a huge momentum boost as they go to game 7 and try to finish the series.

Side bar: My beloved Rockets have been on both the favorable and unfavorable side of this Game 7 situation in the past several years. In 2014-2015, down 3-2, the Rockets made an amazing comeback in game 6 vs. the Clippers in (led by Josh Smith, Corey Brewer while James Harden on the bench) and the momentum carried over to an easy victory at home for Game 7. It was beautiful. Enjoy this glorious moment with me:

The unfavorable game came in the 2017-2018 WCF, where Rockets were up 3-2 against the Warriors 3-2 after Chris Paul led the Rockets a gutsy win in Game 5. He injured himself and had to sit out the remainder of the series The Rockets went on to lose Game 6 in Oakland and Warriors took it in 7 after the infamous 0-27 from 3PT fiasco, but let’s move on before I throw my laptop across the room. No highlights for this one and let's never speak of it again.

Back to our query - let's tweak it a little more and exclude any results from Games 6 and 7:

• series game<=5: pulls in all games played in Game 5 or earlier. The results are slightly better, but still not a worthwhile edge.

Maybe we should check to see if the team is a Favorite or Underdog coming off that loss. If they are a Favorite, chances are they “the better team” and the previous game was an upset and chances are the lesser team will not be able to pull an upset two games in a row.

• F: add in the “F” to indicate that is coming off the loss is the Favorite in the game. 58.3% ATS! Now we're talking!

Let's add one final parameter that looks at how a certain seed performs in this scenario. A quick glance at the query told me that the higher seeded teams (#1 and #2) perform better in this situation.

• seed<=2: this will pull in all teams that are a 1 or 2 seed.

Bingo! I like these chances a lot better. These results do not surprise me as there is not much parity in this league and the top teams are usually head and shoulders above the rest and tend to bounce back strong from a loss.

The point is, you can keep slicing and dicing this data to find how teams perform in certain situations. The possibilities are endless.

Finally, let's explore a scenario that is not so successful:

playoffs = 1 and season >= 2011 and p:AL and p:round = round and HD

• p:AL - this means that the team lost their last game on the Road

• HD - 'H' is a shortcut for Home and 'D' is a shortcut for Underdog

Following the system in this scenario will empty your wallets. This has gone 4-16 ATS since 2016 - yuck! Perhaps the public perception is that a team who has lost their last game on the road but will be in a bounce back spot and have motivation to play harder in front of the home crowd. The SDQL results show that oddsmakers adjust to this and shade the lines too much in favor of the Home team and they typically fall short against the spread.

Conclusion: does the Zig Zag Theory work or is a myth? It depends on the situation. In our 2 test cases today, it can be a very successful strategy as a top seed (1 or 2) whom is a Favorite coming off a loss early in the series. However, Home Dogs coming off an Away Loss usually fall short.

Hopefully this gives you some insight that blindly following a pick because someone claims the Zig Zag Theory is not always to best idea. Use the SDQL as a resource to help identify which situations are more profitable. Visit https://killersports.com and start practicing today.

What NBA Playoffs topic should I tackle next? I am open to suggestions. I am leaning towards searching for a Totals system.

Follow me at @awang_htx on Twitter.