Spreadsheet Modelling Craft at SUU - A Living Textbook
Thursday, May 26, 2011
Transpose
Under "Paste Special" is an option to change columns to rows and rows to columns when you copy and paste.
Risk Optimizer
This ships with @Risk.
It is a program to simultaneously do Monte Carlo and an optimization routine (like Solver).
It is a program to simultaneously do Monte Carlo and an optimization routine (like Solver).
Wednesday, May 25, 2011
In Regression, More RHS Variables Make Each Estimate Less Sharp
The technical term is multicollinearity.
It's a good thing to start a regression big, but you have to remember to prune it down.
If you don't, what you get is larger standard errors for everything which makes them all look insignificant.
It's a good thing to start a regression big, but you have to remember to prune it down.
If you don't, what you get is larger standard errors for everything which makes them all look insignificant.
Power Curves Are Common
These are related to Pareto distributions ...
Power curves describe a lot of real world phenomenon like the distribution of talent, and tipping points, and the effects of advertising.
N.B. Go read Malcolm Gladwell's The Tipping Point.
Power curves describe a lot of real world phenomenon like the distribution of talent, and tipping points, and the effects of advertising.
N.B. Go read Malcolm Gladwell's The Tipping Point.
Deterministic vs. Stochastic Models
Most Excel models are determinimstic: the inputs don't fluctuate, so the outputs don't either.
But in the real world they both do.
Old School: using scenario manager to vary inputs, or doing it by hand.
New School: use Monte Carlo add-ons like @Risk or Crystal Ball to automate and keep track of simulations that generate probability distributions of outcomes.
But in the real world they both do.
Old School: using scenario manager to vary inputs, or doing it by hand.
New School: use Monte Carlo add-ons like @Risk or Crystal Ball to automate and keep track of simulations that generate probability distributions of outcomes.
Subscribe to:
Posts (Atom)