Yesterday's post led to an interesting Twitter thread last evening, which included a suggestion to reanalyse the data to determine whether price movements in the Pinnacle head-to-head market might have predictive value in other markets for the same game, specifically in the line market.
So, after classifying games in exactly the same way as I did for the previous post, I calculated the returns that would have accrued to someone wagering in the line market on the Favourite or on the Underdog in each of the different game types. The results are in the table below.
What we find is that, as we found in the previous post, wagering on Underdogs whose head-to-head price shortened between the opening and closing of the market is a profitable proposition. Here, in the line market, we find that it provides a 1.3% ROI, which is slightly less than one-third of what the same strategy provides in the head-to-head market.
In the line market however, there's also a positive ROI to be had by wagering on Favourites whose head-to-head price has shortened. That strategy provided a 3.1% ROI over the period analysed. This contrasts with the small loss that such a strategy was found to incur in the head-to-head market.
If anything then, it appears to make even more sense to follow the money in the line market and wager on whichever team has seen its head-to-head price shorten (provided, of course, that this description doesn't apply to both teams in the same contest). For the Pinnacle data that would result in collects about 53.4% of the time and an overall ROI of about 2.3%.
I'd make a similar caveat here about the generalisability of this result into the future that I did in the previous post: it's based on a smallish sample and there's no replication/holdout. One aspect of today's result that is more appealing than yesterday's, however, is that the returns in the line market are, by their nature, less variable than the returns from wagering on underdogs in the head-to-head market. Still, caution is highly recommended.