If I go to the Burger King (for some reason it sounds better that way, by including the ‘the’), I kind of know what I’ll get. The burger, will likely be a Whopper. Whether I’m in London or New York, the whole Burger King experience is likely to be very similar. The burger should taste quite similar. The restaurant will be adorned with the Burger King logo etc. I basically know what to expect when I go to a Burger King. To use market lingo, my expectations are pretty much priced in. It’s unlikely the true experience will deviate much from expectations. By contrast, if I go to an untested burger joint, my expectations are going to be much more fuzzy. I have information about other burger joints in a similar price range, which I’ve been to. I might read reviews about the burger joint, to assess what others thought of it. I might use these to price in what I expect. Each additional piece of information help me price things that bit better. Will it be a good burger or not? However, it’s difficult to price my expectations properly, because, I’ve never actually been there before.
This is somewhat of a trivial example, it’s unlikely anybody (even me), would try to quantitatively assess their expectations of what a new burger joint could be like. Obviously, within markets, traders are continually trying to price in what will happen. When there’s a known event, we can attempt to price the outcome. One common case, can be trying to price in the outcome of a central bank monetary policy meeting. Will the Fed hike or not? The first point to make is that, it is difficult to make this prediction (ok, obvious point!). However, in a sense, we do not need to predict the outcome perfectly. Instead, we need to predict it better than what the market is pricing when we enter that trade.
If it is a central bank, which fewer market participants follow, we might conjecture that information flows might make it easier to find mispricing (against our own views). It might be more challenging too to ascertain precisely what the market is pricing for these too. Even for central banks like the Fed, which everyone follows, if we are trying to forecast monetary policy a long way out, what the market is pricing might be somewhat more interesting to speculate upon. Furthermore, in practice, we don’t necessarily need to predict the event itself, more the shift in market expectations after the time we enter that that trade. We might well choose to avoid taking the actual event risk in any case, and exit the trade before the event happens, profiting (or not profiting) from market changes in pricing. I recently read a note by Cameron Crise from Bloomberg News, elaborating on this point in a bit more detail. We could come up with other examples, such as when trading long dated vol. It is very difficult to estimate how volatility will play out over the very long term, so again, it’s not about getting it perfectly right.
None of this stuff is particularly easy, but I would argue that trying to predict every event you trade is never going to happen. Instead, trying to get results which are “good enough” is sufficient. You’re not going to make more money if your prediction of payrolls is 1k better than the next person, if you’re very far away from market expectations! That is simply the wrong thing to optimise for, if you’re trying to make money from your trading (as opposed to trying to win the competition for best forecaster!!).