Last week, I shared eleven questions on my mind about the MLB season as it approaches. Here they are again:
- Will any more players opt out before the season starts?
- What happens if a player wants to opt out in the middle of a season?
- What happens if a team has too significant a coronavirus outbreak to field a full roster? Will they forfeit?
- What has to happen for the MLB to pull the plug? Is it a nationwide death count? A particular level of impact within the MLB? One specific tragedy within the MLB? Action by the NBA? Action by state and local governments?
- How likely is it that the MLB finishes the season?
- Will players have to sit out games due to testing delays, or will the MLB figure this piece out that’s forced players to miss workouts? If the MLB doesn’t figure it out, who makes the decision on who plays if a test result hasn’t been relayed back to the team?
- What will the impact of low attendance be on the effect of home-field advantage?
- What will the impact of lessened travel be on rest and home-field advantage?
- What will the impact of the lower number of total games be on rest?
- How aggressively will managers manage out of the gate?
- How clearly will we be able to hear the verbal pieces of the game on TV and radio?
I wrote at the time that the first five of these are logistical curiosities, that the last one is an entertainment curiosity, and that the sixth through the tenth should impact projections and odds, most notably in the futures market where guesses need to be made a couple months out rather than taken day by day. In truth, though, they all impact futures bettors. Well, except for the eleventh. That’s still just an entertainment curiosity. The hope is we can hear players cuss out Joe West (in a socially distant manner, of course), but whether we can or can’t shouldn’t impact any bets.
Question 1: More Opt-Outs Before Games?
The first question of the eleven is being answered as we speak. The list of players who’ve opted out is somewhere around twelve, plus free agent Tyson Ross. It may grow this week, but that’d be surprising, with the first games in two and three days.
Question 2: More Opt-Outs After Games?
The second was answered in June, and I just hadn’t found the answer. Per Mike Axisa at CBS Sports on June 24th:
High-risk players may opt out of the season at any point—if they begin the season but later change their mind because they feel unsafe, they can do that and opt out—and still receive full salary and service time. Players who opt out because they live with a high-risk person are not entitled to full salary and service time, though the Operations Manual allows teams to be accommodating. Essentially, the team could still pay a player who decides to stay home because he has a high-risk family member.
This is significant, because it opens the door for players to give up on the season early, making a fast start potentially even more valuable. There are other implications, of course, but at a glance, this seems to be the most significant, and something models might not capture.
Questions 3 and 4: Could a Team Be Eliminated but the Season Go On?
It’s doubtful that there’s a definite answer to the third or fourth question right now. These are hypotheticals. The simplest bet seems to be that if a team has a large coronavirus outbreak and can’t put players on the field, the MLB as a whole will shut down. We don’t know, though, and this is a reminder that with everything this season, a higher degree of randomness than even that which the smaller sample size creates should be expected. For futures bettors, this should be part of the calculation, or at least the consideration. Personally, I’m including it in futures considerations in the following way:
The question at hand is what probability to assign each team of being straight-up eliminated from contention due to a coronavirus outbreak. Formally, this could happen by not fielding a large enough team, but the more likely scenario is that one team just becomes much worse by having a number of their best players test positive at once, while the MLB as a whole keeps enough of a lid on infections that the owners and players both choose not to cancel the entire season. This is a low probability because it requires threading what I’d imagine is a tight needle between one team having it bad and the whole league having it bad. It’s difficult to estimate. So many assumptions are required that finding the correct magnitude seems impossible.
So, let’s try this. If all teams have the same probability of suffering an outbreak, and one team does come to suffer an outbreak, what’s the probability no other teams do? In this calculation, we’re assuming that if two teams suffer an outbreak, the season’s over, which isn’t necessarily true. Having made that assumption, though, we get numbers like the following: If there’s a 1% probability of any given team suffering an outbreak, and one team does suffer an outbreak, there’s then a 75% probability none of the other 29 teams suffer one (99% to the 29th power is roughly 75%). If the outbreak probability for individual teams is 3%, the probability of one team, having broken out, being the only one is 41%. If the outbreak probability is 5%, and a team breaks out, the probability of being the only one is 23%.
So far in the MLS Tournament, two teams out of 26 have dropped out due to virus outbreaks. The setup is different in the MLS, of course, and we’ve now gone a week and a half since a drop out after seeing two in the span of a few days, and more teams will have outbreaks, but if we assume that two out of 26 is the probability of an individual team suffering an outbreak, we get a probability of roughly 8% percent that a team suffers an outbreak and 10% that, given its suffered an outbreak, it’s the only one to have suffered one substantial enough to require the cessation of play. Combined, we have a 0.8% probability for each individual team that its season is ended but the season at large continues, meaning all playoff probabilities, for futures betting purposes, should probably be multiplied by something like .992 before calculating expected payouts (an addendum here: other teams would benefit from one dropping out, but the benefit would be spread between four to 29 teams, making it a fairly insubstantial impact on the playoff probability for these purposes).
Question 5: Is the Season Going to Finish?
The prospect of the season finishing is relevant. My assumption is that sportsbooks will score futures as a push if they don’t come to completion, as they did with college basketball in my experience, but I’d recommend bettors do some work to see if they can get a guarantee on that, or at least accept the fact that they might get hosed. If a bettor’s worried about getting hosed, they’ll need to account for that in their futures bankroll, and even in the likely scenario in which DNF bets are pushed, they’ll need to account for the risk of investing money with no payout when they could have been investing elsewhere.
I could calculate this out, but instead, I’d just reiterate that anyone placing futures bets should seek confirmation they’ll be refunded if the season doesn’t finish, and make their futures bankroll small enough that they’re ok with it not earning any money in the event everything’s canceled (making one’s bankroll small enough that one is ok with losing it is always good advice with gambling).
Question 6: Will the MLB’s Testing Work Quickly Enough?
This pertains more to individual games than futures since the risk is probably equal across teams, or at least across teams in the same geographic area. The advice here is to wait and see. As recently as last week, the Cubs had more players sit out a practice because no results were available. It’s possible there have been others that I’ve missed. Keep an eye on this if you’re betting on individual games, and place your bets close to game-time if you can swing it and you think things are risky. Also, remember teams are affected equally by this with regard to individual games, too, as far as we know right now. So, maybe a little mean-reversion should be expected, but let’s wait until this issue actually affects a game before we worry too much about it.
Questions 7 through 9: Home-Field Advantage
I did a little work on this earlier today. Last year, the home team won 52.9% of regular season games. In 2019, it was 52.8%. The five-year average is 53.4%. The individual game probabilities on FanGraphs, judging by six games listed right now in which the teams and starting pitchers are the same but the location is different, assign home-field advantage a weight such that home teams would be expected to win 54.0% of regular season games, all else equal.
To really assess home field advantage, we have to question whether it’s a linear impact or an impact of magnitude—in other words, if a team would be a 60% favorite on a neutral field, should we add 3.5% to their odds or multiply their probability by 1.06 (or similar numbers) if they’re playing at home under normal conditions? I don’t know the answer to this. I’d guess it falls more on the multiplication/magnitude side of things. But because FanGraphs does it linearly, and I use FanGraphs odds as my first resource in examining individual-game bets, I’ll do it linearly as well when I adjust their adjustment.
What, though, should the reduction be to the linear number to account for how different home field advantage is this year? FanGraphs doesn’t seem to be making one, though this is something to monitor. FiveThirtyEight is reducing the advantage by 60% in European soccer, which is an imperfect comparison because road trips are shorter, travel time is shorter, and the impact of the crowd upon the game isn’t identical. With this, as with the will-a-team-drop-out adjustment, it’s better to err on the side of caution for both teams, but generally, it seems fair at this point, before we know more, to say that home field advantage should only be worth a roughly 1.5-point linear boost to win probability for the home team and a 1.5-point subtraction from win probability for the visitors. Personally, when looking at FanGraphs, I’ll be applying this by subtracting 2.5 points from the home team’s win probability and adding those points to that of the visitors. I’ll also be checking every now and then to see if they’ve changed their calculation.
Question 10: How Managers Manage
The obvious answer here is that they’ll be more aggressive right away, but that we don’t know what the impact of that aggression may be. We really don’t know. I have some anecdotal evidence from the last two months of last year that implies that better teams outperform the FanGraphs model down the stretch, but it’s only anecdotal evidence. Perhaps the best way to adjust for this is to consider which team has the better-rested, stronger bullpen in each game, and assume that team will get a slight bump thanks to their manager’s ability to pull the bigger lever than his opponent. How large a lever is this? Probably not more than a few percentage points, relative to the average regular season game upon which I’d assume FanGraphs’ model is based, but also not zero.
***
These aren’t answers to the questions. They’re estimations, hedges, and assumptions. They’re applications of the questions. We’ll get some of the answers soon. We might never get others (hopefully, we never get the worst-case answers). Whatever the case, the baseball season is starting. Whether it finishes is anybody’s guess.
