Wednesday, November 9, 2016

Election Betting Recap

Election betting - especially a close election - is an interesting exercise. In this day and age, gamblers and forecasters lean on models and ratings for horse racing, sports betting and otherwise that are tested and retested empirically. An election is sometimes like that, but oftentimes not. Last night was clearly the latter.

Here are some thoughts on the wagering, and data.

In preparing for wagering, the traditional polls, plus exits are informative. In this case, the exits confirmed the polls, which is usually good news. Pre-betting, I had a Clinton plus 3, 308 electoral vote projection. That, at the time, looked sound.

There were two clues I uncovered in the exits though, which I filed away. Trump was doing much better (than comprehensive polls from Wapo-ABC and others which I dug up suggested) with college educated whites, and the working class voter numbers were also quite good for him. The union vote was coming out for him, which most did not have pegged. There's a chance the plus 3 projection was high.

That data, to me, said Iowa looked like a lock. Taking 2-5 odds there seemed wise. Longshot fliers in Wisconsin, Michigan and Minnesota with these numbers seemed unlikely, but value.

At 6PM only parts of Indiana and Kentucky close and these micro-votes (that the networks tell you not to look at) questioned my pre-results bias, further.

Viggo county in Indiana was solid Trump and it should not have been.

Elliott County (well known to my friends in Kentucky) was not only a solid Obama country the last election, it has been a solid blue county for 136 years. After two polls, he was up 69-29.

Trump, a half hour after these early states, was 8-1 to win the election, and in hindsight, there were signs he should be half that, if not less. I, at that point, mentioned to my betting partner that maybe the polls were wrong and he was going to have a good night.

Why election betting can continually give us edges -- and in my view it's one of the only betting mediums left that can allow monster edges -- is because things are confusing. Throwing a wrench into the plans was the early vote. This year's was a record, so when cities came in Hillary was doing very well with that data dump. Was this data a trend, or simply her edge on the early vote and Viggo and Elliot counties were the real story?

The last issue that was needed to be factored, was inner city turnout. If it was much lower, you could feel good being long Trump, if it was in 2012 numbers, you'd go the other way. We had very little of that data at around 7PM.

We all have biases in betting, and mine was to be cautious, because it was confusing. I also had an internal bias  -- I apologize to my Trump friends reading this; I was not a fan, so picturing him winning was difficult.

Later on, this bias slipped away. Counties in Virginia were coming more like 2012, not like the 2016 polls suggested. She looked like she'd win Virginia (and I took a couple of positions) but she was running very soft.

At that point, Trump was still a strong underdog to win, and he should not have been. There was the value, at over +300.

North Carolina represented a true swing market, and again, it was confusing. Around Raleigh she was overperforming him - widely and above President Obama in 2012 - but there's that early vote phenomenon again. Regardless, we took some positions at about 4-1 on Trump (mainly because of the Virginia data and trends), which was a good trade.

Finally in Florida -- which I lost severely in 2012 because of Broward County numbers that came in late, and well above what anyone projected they'd be --  we had our answer about the inner city vote. It was not there. I still think - just like 2012 - taking HRC positions near the end (when she was over 10-1) was smart. It was close, and there was a shot (just like the late numbers came in, in MI)

Election betting - despite the craziness of the race, which at times for me was impossible to follow -  once again proved a challenging mental exercise. And if we dug far enough, and stuck with convictions of what we were seeing, (not what others told us we should be seeing) there was some money to be made. 


No comments:

Most Trafficked, Last 12 Months

Similar

Carryovers Provide Big Reach and an Immediate Return

Sinking marketing money directly into the horseplayer by seeding pools is effective, in both theory and practice In Ontario and elsewher...