Saturday, 29 February 2020

Signs of Cheating

A Lucky A Day asked about the impact of the cheating Houston Astros on the results for hot favourites in baseball I looked at in my last post.
For those who may not know, the Astros were caught stealing signs, something which in itself has always been an accepted, almost admired, part of the game, but the use of technology to steal them has always been considered cheating, and the use of technology in real-time was especially egregious.

Sign stealing in baseball has a long history and goes back to 1876:
when the Hartford Dark Blues hid a person in a shack to tip off their hitters when the pitcher would throw a curveball.
The team, supported incidentally by Mark Twain, folded after the 1877 season, after they had moved to Brooklyn.

The Boston Red Sox are also under investigation for cheating during the 2018 season, and the New York Yankees are also reported as being suspected of cheating in recent seasons.

The Astros cheating has been proven to have occurred during the 2017 season, which ended with the Astros winning the World Series, and into the 2018 season. No evidence was found of cheating in the 2019 season, although it has been suggested that it began as early as the 2016 season. While the impact of Home advantage has declined in recent years, this is one example where playing at Home was definitely an advantage.

As Winston pointed out, the cheating at Home probably had something of a carryover effect on Road games as well.
Possibly, but I think from a betting perspective, previous results would already be factored into the price, albeit with too much weight on recent results.

For non-baseball readers, the catcher signals (signs) to the pitcher what type of pitch to throw. By ‘stealing’ the sign, the information can be conveyed to the batter so that he knows what pitch to expect. One extreme example of this advantage is from Game 5 of the 2017 World Series when the Los Angeles Dodgers pitcher was Clayton Kershaw, well known to readers of this blog.

In that game, Kershaw threw 51 curveballs and sliders, and not one pitch resulted in a swing and miss. With the benefit of hindsight, we now know why.

Anyway, how did the Astros fare in Home (regular season) games during the period in question?

From 2014 to 2016, they were 1.5 or shorter on 15 occasions, winning 12 for an ROI of 15.8%.

In 2017, not surprisingly they were hot favourites on 20 occasions, winning 17 with an ROI of 20.3%.

In 2018 they were big favourites 55 times, although the 35 wins resulted in a negative ROI of 10.5%.

The 51 games in 2019 resulted in 41 wins and a 9.3% ROI.

Overall the difference over six seasons with the Astros excluded is small:
For Home games only, the numbers are:
Green Sea asked me what I meant by -200 or shorter. Poor wording yet again on my part perhaps, but by ‘shorter’ I mean any price that gives a lower return to a bettor, i.e. the implied probability of a win is greater. 

-200 is 1.5 in decimal odds, so anything shorter would include everything in the 1.01 to 1.5 range, although not many baseball teams start at 1.01 even if they are cheating!

Although nothing has been proven regarding cheating in the 2019 season, there are a couple of interesting observations. One is that in the World Series, the Washington Nationals won all four games in Houston after being warned by players from other teams about what was going on. As a result of these warnings, the Nationals “developed a complex system of mixing signs to thwart any attempts to steal signs by the Astros.”

Also notable was that:
Before the 2019 World Series, Joe Torre and other MLB officials had an unusual pre-series meeting with the GMs and managers of both teams to warn them against using cameras on the dugouts or catchers or using electronics in the dugout.
One might have expected the Astros to be stripped of their, already tainted, title, but MLB seem anxious to move on. The Dodgers can certainly feel they were cheated of a World Series, but then other AL teams were also cheated of the chance to play in the World Series

I did actually mention the Houston Astros a few times last season when their price was at, or close to, a record for the shortest ever in database history, but apparently they weren’t cheating.
Coincidentally, given that we are discussing favourites, the Houston Astros today became the shortest ever (since 2004 anyway) priced road team at around -430 but they lost 7-8. I saw that the price touched -500 at some places, but in a game like baseball, you really are playing with fire at those odds.
Hmmm. Someone certainly had confidence in them although as I mentioned, when you’re playing odds this short you are really playing with fire.

The Astros comprise five of the shortest six prices in database history, but backing into these lines isn't the path to riches:
Looking back, these 2019 Astros prices really were unusually short, but perhaps not surprising if cheating was going on last season. 

Insiders would have known that there was an edge on the Astros at home and the market will adjust to the money coming in, an illustration of why sportsbooks are taking a big risk if they use their models to take a significant position on a game.

Here are the results by season (including post-season) for backing the Astros both home and away:

No comments: