-Matt Cox

Thanks to all who took a swan dive into our home court value analysis refresh last week. If you stumbled upon this article without consuming those pre-read materials, revert back to part 1 here.

Double thanks to Jim Root, my fellow Weave cohort, and Chris Molicki, co-host of the Sports Brunch podcast, for raising their hands and pointing out two critical holes in that part 1 rendition, which left the following questions unanswered:

1. How did the difference between actual home court value and oddsmakers’ estimated home court value translate to against-the-spread (ATS) results?

2. How should the actual home court value and oddsmakers’ estimated home court value calculations be used in practice for betting / handicapping during the season?

Overlaying Actual ATS Performance

A key takeaway from part 1 is that home court value, on average, varies greatly by conference. This is not earth-shattering news by any means, but the relative variance across conferences is insightful – and can be weaponized in your own personal betting and handicapping.

The chart below juxtaposes 2019-20 ATS records by conference with our previous calculations of actual and oddsmakers’ home court values by conference (already shown in part 1). The data is sorted by the shaded left-most column, ‘2019-20’ Actual HCA’, which is defined by the average scoring margin differential in conference head-to-head matchups.

To contextualize what we’re looking at, let’s highlight the Big-12, an appropriate example due its true round robin scheduling format in which all ten teams play each other twice (the Big East is similar in this regard). The way to interpret the 5.89 actual home court advantage figure is that across all intra-conference, head-to-head matchups, the home team won by an average of 5.89 points a game.

Now, compare that to the oddsmakers’ estimated home court advantage, 3.62. This is the average home court value derived from the oddsmakers’ final spreads, which is calculated as the average difference between all intra-conference head-to-head matchups (divided by 2). For example, if Kansas was -9 at home against West Virginia and -1 on the road against West Virginia, the implied oddsmakers’ home court value is 4 points (9 minus 1 divided by 2).

The positive variance between the Actual HCA (5.89) and Oddsmakers’ HCA (3.62) implies that the bookmakers, on average, under-estimated the value of home court advantage in Big-12 conference games by roughly two points (5.89 - 3.62 = 2.27).

Obligatory Reader Beware: There is A TON of noise littered throughout the underlying data, as detailed in part 1, so please, please, please understand the pitfalls associated with this quick-and-dirty approach (there are many).

If you scan down the list, you’ll notice the actual 2019-20 HCA figures are larger than the oddsmakers 2019-20 HCA for 16 of the 32 conferences (denoted by the horizontal line below the NEC and above the A10). Unsurprisingly, home teams covered more frequently than road teams in 12 of these 16 conferences, with the exceptions being the Summit, Mountain West, Big East and MAC. What’s more insightful is that of those 12 conferences, eight had home ATS records greater than 53% (refer to the ‘Win% column’), which is the win rate threshold needed to cover the juice. In these eight leagues, blindly betting the home team would’ve been a profitable strategy and generated the corresponding returns depicted in the far-right column (‘ROI’).

The bottom half of the chart tells the inverse of the same theme as above. In 13 of the 16 conferences with actual HCA lower than oddsmakers’ HCA for 2019-20, home teams covered less than 50% of the time (again, to be expected). In all 13 of these conferences, auto-betting the road team would’ve been a profitable strategy – though, only the SoCon (SC) and Atlantic Sun (ASun) would’ve generated significant returns by doing so.

Before getting into some interesting observations, the key data point omitted from this chart is ‘average cover margin’, the average value by which home teams covered the spread in each conference. Last year, in situations where a home team obliterated the road team – for example, when South Dakota beat Denver 72-45 on February 20th – the lopsided outcome would’ve had a substantial impact on the actual home court advantage calculation, but only result in one win from an ATS record perspective. Thus, instances like these explain how the Summit was only 32-39 ATS at home, despite showing a one-point actual home court value edge over the oddsmakers’ implied home court value (4.38 – 3.38 = 1). The home teams who did cover tended to cover by a large margin, while road teams tended to cover by smaller margins. Conversely, the A-10 showed a slight negative actual home court value to oddsmakers’ home court value difference (2.96 – 3.39), but the home teams still went 65-58 ATS – thus, auto betting road teams was not a profitable strategy (-13.5 units).

A few additional observations and takeaways from the chart above:

• We’ve always perceived the Pac-12 to boast one of the stronger home court advantages in the country, but credit to our former colleague Max Meyer, a long time Pac-12 patron, for beating this into our brains. Home teams covered at a blistering 59.4% rate in conference play, second only to the Missouri Valley Conference at 60.5%. Oddsmakers historically appraise Pac-12 home court advantage value in the 3.75 to 4-point range, but that figure dropped to 3.52 last season, presenting a juicy opportunity for bettors to exploit an ill-timed adjustment. Per the 2018-19 actual HCA column, the Pac-12 and Missouri Valley were two of only five conferences with an actual home court value above 5 points. The other three, Big-12, SEC and Big Ten, didn’t produce the same home team betting returns, as oddsmakers, on average, baked in a larger home court value – notice the slightly higher figures in the 2019-20 Oddsmakers’ HCA column for the Big-12 (3.62), SEC (3.63) and Big Ten (3.78), compared to the Pac-12 (3.52) and MVC (3.29).

• If I were an oddsmaker, I may consider reexamining my home court value inputs for the Pac-12 and the Missouri Valley. However, I would not overreact to the MAAC, OVC and MEAC home team dominance seen last season, which seems like an aberration when compared to their multi-year historical home court values. Anecdotally, 2.5 to 3-points feels like a more appropriate range for average home court value in these leagues, which is more in line with the historical data.

• The A-10 is another example of a one season outlier where I’ll largely be ignoring last year’s small sample results. The A-10 not only spans a fairly large geographic footprint but features a bevy of passionate programs with hostile home court atmospheres. Remember, looking only at single-year figures for average home court value is risky. The small number size of intra-conference, head-to-head matchups are well below the sample size threshold for making valid statistical inferences. This is why the 9-year average figure of 3.86 is much more meaningful and a better long-term indicator of how strong home court advantage is in the A-10. Thus, even though the actual HCA fell to 2.96 last season, home teams still covered 52.8% of the time, which is why auto-betting road teams would’ve been a foolish endeavor (-13.5 units).

• As mentioned in part 1 last week, The Southland (Slnd) is a fascinating case study. We detailed how, despite being a low-major, it has consistently held one of the stronger home court advantages of any conference in America over the last decade – refer to the 3.83 under ‘9-yr Actual HCA Average’. Based on that historical trend, the oddsmakers inflated their average home court input to 3.16 last season, a slight uptick from 3.09 the year before.
  
  The historical data confirms the oddsmakers made the right directional adjustment last year in the Southland, but their timing was simply unlucky. Last year, home teams only won by an average of 2.50 points, 1.33 points below the historical 9-year average of 3.83. Thus, the culmination of home teams severely underperforming relative to historical standards, combined with an ill-timed, though justified, adjustment from the oddsmakers, resulted in home teams covering at such a low rate. Going forward, I would bet large sums of money the actual home court value converges toward the mean and we see that actual HCA figure tick back up toward the 3.83 historical average.

Putting Home Court Value Data to Work

Up until this point, the crux of this home court advantage series has been told from the bookmakers’ vantage point. So, how do we make sense of all this from the bettor side of the aisle? Can we turn this data into an information advantage?

Well, yes… but not in the way you’re probably thinking…

Last year, we observed only a few instances where a bettor could’ve exploited market inefficiencies at the conference level, either by betting all home teams (e.g, the Pac-12) or all road teams (e.g., the America East) to turn a profit. I cite the Pac-12 and America East (AE) as proper examples, as these are two cases where the historical data coming into last season suggested home court was overvalued (America East) or undervalued (Pac-12), based on a significant sample of historical game results. In both cases, those trends carried over into last year.

Other conference examples where an auto-bet or auto-fade strategy produced significant profits last season, such as the Ohio Valley (OVC) and the SoCon (SC), were more coincidental. There was no data-based evidence to suggest oddsmakers were consistently overpricing or underpricing the values of home court advantage in those conferences.

The broader point is this: there might be a handful of conference-level or team-level home court driven market inefficiencies, but they are few and far between. And the ones that do linger (e.g., the Southland) appear to be evaporating quickly.

Referring back to the chart above, if you subtract the 9-year oddsmaker HCA averages and 9-year actual HCA averages by conference, only five leagues have a difference greater than 1 point: the aforementioned Pac-12 and America East, along with the WCC, Big West and NEC.

Without diving into the significance of these variances, that leads me to believe the books have been pretty damn accurate in their appraisal of home court value. Because the oddsmakers are quick and precise in adapting to these historic trends, I’m not in the business of beating the books based on mispriced home court values. In fact, in my personal handicapping, I tend to gravitate towards using the oddsmakers’ historical HCA averages (based on head-to-head matchup spread differentials) for my own home court value inputs, rather than the actual historical HCA averages (based on head-to-head final score differentials). Worded differently, I aim to align my home court value inputs with what I think the oddsmakers are pricing in.

In my power rating handicapping approach (home team rating – away team rating + home court value = final spread), home court value is the third and final piece of the puzzle. By normalizing for potential differences in my own home court value estimations and the oddsmakers’ home court value estimations, I can zero in on pure team-based pricing inefficiencies – that is, discrepancies in how high / low I have two teams rated, compared to how high / low the oddsmakers have those same two teams rated. The way I see it, the secret sauce to my personal handicapping success is my knowledge and power rating precision for all 353 teams in college basketball. Obsessively tracking every team, their players, and the ebbs and flows of each team’s season is where I dedicated my time and effort – thus, this is where my edge lies.

Now, there are situational exceptions to this approach, ones where I might be inclined to use a more extreme home court value number based on specific situational nuances (I’ve discussed a few instances, both conferences and specific teams, where the books may be slightly behind the curve). But, for the most part, I prefer to lean on the historical oddsmakers’ HCA data as a guidepost, rather than the actual HCA data.

I know there’s a lot to unpack here and this article fails to tie up a lot of the loose ends likely rattling around in your brain right now. Just be aware that the number of unique home court situational advantages are endless. I would not advise blindly plugging in the conference-specific or team-specific historical oddsmakers’ HCA averages if you utilize a similar power rating handicapping approach. Be diligent in adjusting for proper situational circumstances, including travel, rest between games, experience, among a myriad of other factors. There’s a fair amount of debate as to which of these elements holds the most weight in determining the ultimate value of home court, but the final product lies in the eye of the handicapper – the data can only provide the guide rails.