A Casino player That Chipped the Horse-Racing Rule
The probability of the outcome is exactly the same every time you flip it, so the probability of being right is exactly the same. The reason it looks different is that there are many flips. At some point, you have to do it at least 50 times.
What causes the “experience” of flipping the coin to differ from flipping the same coin at some point in the future?
This same principle applies to a big jump on the J curve, or in the case of updating a financial model, running a new and different simulation to estimate the price. All things being equal, the exact details of what we are doing can’t change.
A big jump in something can mean the details of what we are doing are different. For example, the pricing forecast could be a different time horizon or the forecast could change. In this case, if we change a few numbers, the updated forecast will look different than the original forecast.
A big jump in anything can also mean the details are different. If the forecast is wrong, it means we are using a different outcome. For example, using the same data but tweaking one variable and using it differently. This can happen at a very granular level. A few changes to the input values can lead to a huge difference.
At any rate, when the details are different, the J curve will look different.
What are the risks of making a big jump?
A big jump is certainly risky if it causes the estimates to change. While modeling is an art and not a science, it’s still something that’s good to know how to do. We want to be comfortable with the results, which are essentially a snapshot of what we think will happen.
The flip side is, a big jump can be painful, as the amount of uncertainty grows with the size of the jump. So it might be the case that we aren’t that comfortable with the forecast and don’t want to go through the trouble to start with. In that case, it might be best to start with something smaller.
Lastly, if we have a lot of information available, we are able to tweak things and obtain a better forecast, or provide better confidence in the forecast. In this case, maybe we shouldn’t go to the big jump. There is really no downside to this, it’s only to our comfort and confidence in the forecast, and the uncertainty in the outcome.
I just want to quickly check some of the risks of a big jump.
1) The biggest risk is that the result doesn’t match your expectations. In this case, we are assuming that we have data that we can trust, but if it is wrong, it can be painful. While this can happen for a number of reasons (both internal and external to the model), it’s a tough one to be ready for.
2) The second biggest risk is that it causes a panic. There can be some strange reactions to big jumps in a situation like this. A lot of things are jumping around at the same time and it can be very unsettling. When making a big jump, it’s important to take the emotion out of the calculation. It’s much easier to evaluate your estimate when it’s measured against the expected value.
3) Finally, it can cause a mental bias in the population.