## Wednesday, September 10, 2014

### Probability Zombie Hunting: Mark Gilbert Edition

There are many ways to prognosticate what will the markets do

Getting the maths wrong is one of them. A bit lacking in intellectual support, but in any case as good as any other ways if you believe you cannot forecast the markets in the first place!

Here is an example. BTW, I have already written more on the underlying issue here

Unfortunately, Scotland has no bonds (or currency) of its own by which we might more clearly gauge the attitude of investors to the outlook for the country itself. Perhaps, then, the best guide is to be found among the bookmakers, still unanimous in backing the view that Scotland will reject autonomy, despite the recent YouGov poll that put the Yes campaign ahead for the first time. Betting five pounds on Scotland saying "no" will win you just two pounds plus your stake returned; a "yes" vote offers a potential profit of 9.38 pounds.
So on the basis that the house never loses, you'd back Scotland to stay in the union -- and you might also be tempted to buy pounds.

Well. Yes, the house never loses. That means they win if you get it wrong. Or if you get it right. And that is precisely why it gives no indication what so ever which outcome you should bet on! So you CANNOT ever bet on the "basis of the house never loses"!

Just to refresh your maths in case you need it (the author does, definitely!), The above odds means a (5/7) 71% probability of "no" and (5/14.38) 35% probability of "yes". The house always wins because the probability adds up to 106%, instead of fair 100%. So the house have a 6% advantage

All he probably meant that stick with "no", as "no" is priced in with a higher probability. In fact, that is also quite a useless investment thesis. It may have a higher probability, but it is compensated by a lower pay off as well. So your expected win (probability times the winning amount) should be same if the odds are fair (i.e. they add up to 100%). And since they are not in this case, actually the expected win is higher for betting on "yes"

And that is exactly the opposite of what the Mr Gilbert suggests!