Market Watch: US Presidential Election 2020
Everyone is sick of reading and thinking about this but I'm just dropping a brain dump here as a reference for future. In particular as a warning not to get involved next time.
NB: I dont think I am qualified to speak about these things, I am not any smarter than any of the other participants. These are simply my notes as I attempt to grow my own understanding.
The market usually comes to a fair price by way of market forces - diverse and independant opinions aggregated together in the exchange. One thought is that on politics that the market can be lead by an often biased media.
There was a growing opinion that this was a once in a lifetime betting opportunity similar to the McGregor/Mayweather event, the market was wrong/biased, the smart money had run out. The matched amount kept growing and growing presumably people with more money than sense having a flutter. I thought that I could beat this sort of person.
£500,000,000 was traded on betfair before election day.
Where has this money come from?
Matthew Trenhaile expanded on my understanding on the market participants but didn't answer the question that people wanted to know - was the market efficient at coming to a true price? Instead he opted to answer the question that nobody asked - he told us that the market was 100% efficient at being what it is (did this even need to be answered?). His point was its job is not to make accurate predictions but to maximise the profit for the market participants.Pro Gamblers
When you think of the exchange you often think of pros. They cannot get on with soft books so the exchange is definitely a place that they bet.
It is possible that some pro gamblers/modellers/syndicates had taken a position. Could they have been significant players? Could it be that they had run out of bank to balance out the small recreational bettors? Some models had Biden as low as about 1.1 while trading at 1.5. A big pro with a big bank (say £1M) with the opinion that Biden's true price is 1.1 would want to stake £700K+ at 1.5 based on full Kelly staking. Say they prefer 1/4 Kelly then its closer to £200K. The issue being that Politics markets like this only come along once every 4 years. Is that often enough to sustain a pro politics punting career? If so then for how many. If there are 100 at the mentioned stakes then they could account for £20M, if thousands then £200M+. Matthew mentioned different modelling approaches including tracking FX market feeds, Crypto currencies, Regional Banking ETFs indicate strong Democratic performance, Google search key words, Twitter trends, Polls, 538 model. So it is conceivable that these are diverse enough opinions to move the market to something close to true. But it seems unlikely to be 1000's of them and seems unlikely to be at such a big discrepancy of opinion with the market (so maybe they are not staking 20%+ of their banks).
Of course you also have pro gamblers without models but who understand market dynamics like chasing steam and finding arbitrage opportunities.
Bookies
Bookies will also hedge into the exchanges. They can lay the punter 2/1 on Trump and back it straight back at 3.1 on the exchange. It's guaranteed money. And the bookies did have big enough liabilities on this market. Although I suggest that they might let the risk run without balancing up - knowing that in the long term they'll win more that way even if they have to take the odd big hit.
Flutter mentioned liabilities of about 10M and GVC similar. This could not have been hugely significant in the over all 500M matched. It would have biased the market toward Trump though as you can see Ladbrokes and even a sharp bookie like Pinnacle were full up on Trump.
"Whether it’s that people don’t like to publicly support Trump or that the polls are more likely to be weighted towards Democrats, I don’t know. What I do know is that we’re taking huge amounts of money on Trump and in terms of the bets, we’re looking at the complete opposite of the market with around {70%} on Trump."
- Marco Blume (Head Trader Pinnacle Sports).
Recreational Bettors
These players will be insignificant in the exchange. They will prefer to bet at the bookies which offer a more attractive product for them. So, expect it to be significantly less than the volume discussed above. But there will be some for sure (like me on this occasion) who will have stuck some mug money into the exchange.
VIP Bettors
These are the same as rec bettor but with more money. The independently wealthy sort with a lot of cash. Usually bookies will love to accomodate them but if for some reason they want to bet bigger than the bookie wants to allow or the customer wants a bigger price than the bookie is offering then their money can wind up in the exchange. There were reports of some big bets in this market - $5M on Biden and £1M on Trump. It is impossible to say if these were sharps or mugs.
Market Makers
Market Makers will be incentivised by betfair to create prices on markets with little liquidity. Of course they will be trying to win money for themselves at the same time. It would seem less likely that these were a large participant in this market. There was little need to create liquidity with it being the most traded market of all time.
Traders
Most traders possibly work in trading offices with the bookies so this money is already discussed. You do have some independent traders (thinking of people like Caan Berry) that can read trends in the market. I suppose there are two types, a long term trader (swing trading) who can quickly respond to some market related news or trends, buy at the right price and hold on until he can trade out again for profit (or cut his loss if it goes against him). The other type of trader just wants to be at the top of each queue getting matched on the back and lay in equal amounts - scalping.
It is unclear how profitable any of these traders could have been in this market and there is high failure rate for exchange traders in general (which I suppose is true of pro gambling as well). There will surely have been some speculative trades executed at significant points in the campaigns. How big I am not sure.
Some traders might have access to scarce or inside information that they can trade on. For example internal campaign reports, strategies or early count tallies. All this sort of information should end up in the market even if it isn't accounted for by the models/polls.
Bots
Bots are all of the above really. You can have arbing bots that scrape the web and find arbs, betting both sides for gauranteed money. You can have trading bots that queue climb so that they can be on the top trading/scalping either side. These “queue climbing bots” get specific mention in Matthew Trenhaile's version of events. So, I suspect that these automatic bots are significant players on the exchange.
Politics
These campaigns cost Billions for the candidates. It is clear that there is ALOT at stake. There is much to be gained too for other countries/organisations. The betting markets are getting more press and considered almost as informative as the polls. If an organisation wanted to create a narrative that the man they wanted was a strong candidate then they could move the price to show that with enough money. If this is happening it is hard to detect as I expect it would be done as quietly as possible. I dont even want to guess how much money this accounted for the total traded. Its bordering on unverified conspiracy theory. I am not sure on the part these nefarious forces play.
I did make a short attempt to profile the types of stakes coming in on each candidate. I polled the "matched amount" for both candidates and recorded the increments. I will not read too much into this as its unclear to me how betfair update this display. Do they aggregate some channels/agents together before updating? It feels more or less real time though so I will go with it for a bit. I watched 12M in total get matched. Most of the increments that I observed were small (<€1000) bets spread evenly on either side. However by volume these small bets (recreational? oportunistic?) paled into total insignificance with some lumpy money coming in also on both sides.
It is one of the advantages that Matthew Trenhaile describes for being top of the queue. The bot that gets to the top of the queue receives information on bet amounts, frequencies etc that can reveal alot about where in the market you want to be. I thought it would be interesting to try and quantify this in my own way. The results are above.
I think with the frequency and size of the bets coming in that you would have to respect these opinions. It was this money I conclude that formed the market. This money seems very confident (its big) and smart (you dont get to have big money by making dumb bets). In hindsight I would find it very hard to disagree with this money. And indeed the markets were quite accurate. The race did feel a lot closer than the polls and models predicted when Trump won Florida. Even the less liquid state betting markets came to an Expected Value for Electoral College votes of 308 which is close to the actual 306 that Biden should end with (pending recount/legal action).
Conclusion
I think most of the market participants maximise their profit when the price is more or less right.
If the price is wrong the pro gambler will back it till its right (so long as their bank and staking plan allow them).
If the bot is scalping both sides of the queue it wants to get as much on either side, if the price is wrong then it is more prone to drift or shorten. If the price drifts or shortens the bot will only get one side - the wrong side.
The Traders will quickly react to significant news in a measured way, if they overreact or react late then they are on the road to failure. Survival of the fittest will mean that the savy successful Traders will dominate.
The recreational money and bookies hedging will be biased as we see towards Trump, but this is insignificant in volume.
So for the most part the exchange should somehow reflect reality (so long as the nefarious political forces don't have a big say).
Comments
Post a Comment