Love reading anything by @mjmauboussin, but his latest white-paper might've been my favorite.
"BIN There, Done That"
Here's the link: morganstanley.com/im/publication…
I read it, but you should too!
Here's a thread breaking down my favorite thoughts and graphs ...
"BIN There, Done That"
Here's the link: morganstanley.com/im/publication…
I read it, but you should too!
Here's a thread breaking down my favorite thoughts and graphs ...
1) Defining the BIN Model
- Bias: inappropriate application of a heuristic to come to a decision
- Information: measures the subset of signals that forecasters use relative to full information.
- Noise: chance variability of judgments.
Let's break each down further ...
- Bias: inappropriate application of a heuristic to come to a decision
- Information: measures the subset of signals that forecasters use relative to full information.
- Noise: chance variability of judgments.
Let's break each down further ...
2) Understanding Bias
Mauboussin identified two major biases in investing:
- Overconfidence: overestimation, overplacement, and overprecision.
- Confirmation: tendency to dismiss or discount information that is inconsistent with your prior view
Mauboussin identified two major biases in investing:
- Overconfidence: overestimation, overplacement, and overprecision.
- Confirmation: tendency to dismiss or discount information that is inconsistent with your prior view
3) Understanding Information
Mauboussin identified a few factors for divergences in information between varying forecasters:
- Discovering new information
- Speed & accuracy of updating views (i.e., Bayesian)
- Weighting of information
- Being better with signal complexity
Mauboussin identified a few factors for divergences in information between varying forecasters:
- Discovering new information
- Speed & accuracy of updating views (i.e., Bayesian)
- Weighting of information
- Being better with signal complexity
4) Understanding Noise
Key features of noise not shared by the other two (B or I):
- Noise isn't systemic
- You can calculate it without knowing the right answer
Key features of noise not shared by the other two (B or I):
- Noise isn't systemic
- You can calculate it without knowing the right answer
5) Something that blew my mind:
- A reduction in the amount of noise in a forecast improved the "superforecaster" edge by 50%.
- Compare that with Bias Reduction (+25%) and Better Information (+25%)
- A reduction in the amount of noise in a forecast improved the "superforecaster" edge by 50%.
- Compare that with Bias Reduction (+25%) and Better Information (+25%)
6) Noise Traders & Noise Allocators
- Noise Traders: Make random trades in financial markets and trade on noise as if it were information
- Noise Allocators: make random allocations to asset managers
So, more noise = less returns.
How do we fix that? Here's how:
- Noise Traders: Make random trades in financial markets and trade on noise as if it were information
- Noise Allocators: make random allocations to asset managers
So, more noise = less returns.
How do we fix that? Here's how:
7) How To Reduce Noise:
- Invest on signals
- Understand the investor on the other side of the trade
- Invest on signals
- Understand the investor on the other side of the trade
8) How To Reduce Bias
A few methods:
- Include base rates in forecast
- Create mechanisms to open mind to alt. scenarios
- Create process to provide accurate and timely feedback
A few methods:
- Include base rates in forecast
- Create mechanisms to open mind to alt. scenarios
- Create process to provide accurate and timely feedback
9) The Secret To Forecasting:
Mauboussin highlighted the single most important factor of success for these superforecasters:
- Updating beliefs based on new information.
Mauboussin highlighted the single most important factor of success for these superforecasters:
- Updating beliefs based on new information.
