Latest News

## An empirical look at blowout risk in the NBA

• GPPs are often won and lost based on which games stay close, which games become blowouts, and which games go to overtime. Vegas Lines help DFS players predict which outcomes games are likely to see.

This article will use Vegas data from 2008-2009 to the present to gain a deeper understanding of how Vegas spreads can help us predict game outcomes. We’ll be looking for answers to the following questions:

• Quantitatively, how much does the spread affect the chances of a blowout?
• Can we use spreads to help predict overtimes and to what degree?
• To what degree does an underdog being at home mitigate blowout risk?

## Point Differential vs. Vegas Spread

Here are the historical score differentials based on spreads at the end of regulation:

### Some notes and general takeaways

• There is overlap in the ranges. For example, the 0-2 group and the 2-4 group both include games with a spread of 2. This is so this article can both look at as many ranges as possible while keeping the sample sizes large enough to be significant. 12 point spreads are not included in the >12 range
• The point differentials are not assuming the favorite is up. For example, an 8 to 10-point spread resulted in a 1 to 4-point game 15.6% of the time. This does not mean the favorite wins by this margin 15.6% of the time. The 15.6% includes times where the underdog wins by 1 to 4 points.
• NBA games with spreads between 6 and 8 behave very much like the average of all games. We see the 6-8 row very nearly matches the all games column. This gives us a benchmark when thinking of spreads. We should consider a spread around 6-8 an average spread, above 8 a high spread, and below 6 a low spread.
• The odds of OT are very much dependent on spread, but perhaps less so than one would think. The trend of lower spreads leading to more overtimes is very evident. About 1 in 12 games with a spread of 2 and lower will result in an overtime. 1 in 25 games with a spread between 10 and 12 going to overtime is not insignificant, however.
• NBA games are very hard to predict. Spreads very much do not guarantee close games or blowouts. 27% of games with a spread within 2 finish with a team winning by over 12. 5% of games with a spread greater than 12 finish within 8.
• Any Game can get out of hand. In every range, more games turn into 21+ point blowouts more often than they go to overtime.

## 3rd Quarter Point Differential vs. Vegas Spread

The 3rd quarter point differential can provide greater insight into the nature of blowouts than the regulation differential. Intuitively, teams that have the game in hand can let off the gas and give back points in the 4th. Blowouts are often fine in cash games, so long as the game doesn’t hit blowout territory too early.

This is especially true for a very high usage player playing as the favorite, who often will be the reason a game turns into a blowout at all. Late blowouts can also be managed well if a player’s minutes are staggered so they play at the beginning of the 4th quarter with the second unit before the rest of the starters come back into the game.

The 3rd quarter spread vs. Vegas spread can give insight into how likely a game is to stay close- at least long enough for our cash game plays to get some 4th quarter run.

Here is the 3rd quarter point differential compared to the Vegas Spread:

### Some takeaways from this data

• NBA games are very hard to predict. We’ve already mentioned this above, but it can’t be overstated. Nearly 26.2% of games with spreads between 0-2 had potential blowouts brewing with point differentials of 13 or greater.
• Point spreads over 12 are extremely risky. One thing that stands out here is the comparatively massive 23.5% chance of a game with a 12 point or higher spread being out of hand. Only 3.5% of all games have spreads this large, so it makes sense that these games would represent the biggest mismatches in the NBA. The sample size is relatively small, but a differential of 21 points or more going into the 4th happens 32.1% of the time in games with spreads over 14.
• Games tend to have similar risk of blowouts until the spreads reach double digits. Note how flat percentages of a game being out of hand or close to being out of hand are until the 10-12 range. The intuition that 10 point and larger spreads are when blowouts get a little worrisome is not just a bias towards fearing double digits, but backed up by the data.

## Blowout risk: Home vs. Road

DFS players commonly believe home underdogs are in position to keep games closer than underdogs on the road. As the following chart shows, it is true home dogs get blown out less than road dogs.

This chart is a bit misleading in the sense that teams tend to be bigger favorites when they are at home compared to when they are on the road. Below are the charts for games with a home dog, games with a road dog, and difference between the 2 in each spread tier.

After adjusting for spread, we see home dogs get blown out slightly more than road dogs, contrary to DFS’s general wisdom. This is almost definitely noise, but it shows there is likely no significant increase or decrease in the likelihood of a team getting blown out based on their home or road status independent of spread.

## The Golden State Warriors

With their fast pace and consistently large spreads, the Golden State Warriors are the team with the most potential benefits if one can accurately predict their blowout risk. While there will unfortunately never be a single-team sample size large enough to be sure, the Warriors produce a chart so odd that it is very much worth sharing.

This is since the Steve Kerr era in Golden State began. Note the hilarious amount of 21+ blowouts in games with spreads less than 4. Though the sample size is too small to make any serious conclusions, there does seem to be indication Golden State may carry a higher risk of blowout independent of spread.

• When looking to see if the increased volume of 3-point shots over the years has increased the likelihood of blowouts independent of spread, no conclusion was reached. If there is an increase in blowout risk it is very small, if not entirely negligible.
• When looking to see if pace increases or decreases the likelihood of a blowout, we discovered a somewhat significant correlation between high paced teams and blowouts. The sample size was not large enough to make any claims with a high degree of certainty, but it could be something to keep in mind.
• Blowout risk is relatively similar for all spreads until it nears and eclipses 10. Blowout risk ramps up quickly as spreads get larger than 10.
• Point differentials in NBA games are very volatile regardless of spread.
• The home-road status of the underdog does not significantly affect the likelihood of a blowout.