Detecting Credit Card Fraud: A Machine Learning Approach
How to make the decisions that will define you
1. How to Make the Decisions That Will Define You
June 04, 2013 AHaller Commentary
Pretend you need to make a choice between doing this or doing that. Both options have their
merits. Both have upsides and downsides. Absent a clear winner (which is often the case) what
should you do?
Always go with the hard choice. Here’s why:
Effort creates its own reward.
Hard work – especially incredibly hard work – rarely pays off in the short term. But without
incredible effort there is almost never an incredible payoff.
The hard choice usually requires the most effort and the greatest personal investment on your
part: And when you put in the time, you learn more, grow more, and achieve more.
Even if you don't hit the target you aimed for you will have hit a lot of other targets along the
way... possibly some you didn't even know existed.
Always choose to work harder. It always pays off.
Hard choices build outstanding reputations.
Staying late to complete a project, making a tough call to a customer, tackling an employee issue
head-on, biting the bullet and taking responsibility when you make a mistake... you don’t have to
do any of them. In a crisis there are always easier options.
But there is usually only one right option -- even if it's the least attractive option.
We all admire people who sacrifice, who compromise, who stand tall in the face of adversity –
so do the right thing, even if the right thing is the hardest thing, and in time you may become
someone other people admire.
Luck is occasional, but intent lasts forever.
Deciding to take the easy way out usually means you hope luck will play a part.
"We'll go ahead and ship this... if we're lucky the customer will never notice the problem."
(Almost everyone who has worked in software or manufacturing has decided to let a quality
problem go so they can meet a ship/release date and hopefully avoid the cost of rework.
Sometimes you get lucky… and sometimes you don't.)
2. While it's painful to make the, "I'm sorry, but we're going to be a day late but we found a quality
problem we need to take care of,” call, it's a lot worse to answer the, "How could you ship us this
garbage?" call.
The angel lies in the details.
Shortcuts, high level decisions, quick fixes... sometimes they work out, but they also mean you
lose the chance to spot other problems, identify other solutions, or find different ways to
improve.
"Quick and easy" creates an illusion of success. Effort and application – and a willingness to do
what others are not willing to do – builds the foundation for lasting success.
The hard choice is always binary.
It's easy to convince yourself that a black-and-white situation is actually gray. Usually it's not:
Needing to fire the employee who doesn’t fit; needing to bypass a senior employee for
promotion for a person less tenured but more deserving; needing to call investors to let them
know results are falling short of forecast… you can talk yourself into thinking there are reasons
not to make the hard decision, but in the end you're just rationalizing.
Usually there are a host of wrong answers, and one right answer.
Think about a tough decision you face. You can probably list a number of easy answers – and
one very difficult choice. Bite the bullet and pick the hard choice.
Maybe you won't now… but later, you will be glad you did.
Randomness is uncertainty
Success and failure arise neither from great skill or great incompetence. Past performance is
only an indicator not a 100% guarantor of predication of future performance. Babe Ruth struck
out as much as he hit home runs.
Uncertainty arouses innate emotions of apprehension and fear.
Many great books suffer multiple rejections. Successful people persevere + external factor
beyond one’s ability converge to = success
More complexity more chance for misinterpretation. Bias cannot be avoided. Mind tends to
recall the vivid, the lively, and recalls less the mundane.
Isomorphism – one problem is another in disguise.
Monty Hall
3. Case 1
Door 1 Right
Door 2 Wrong
Door Three Wrong
Right 1 out of 3 times
Causation blamed on superstition or fate. No pattern recognition.
Pascal’s wager
God : True of false – if false eternal outcome 0 if true everywthing
People wrongly assume conclusions on too small a sample size
“Gamblers illusion” Expected value foundation of gaming theroy.
One school of uncertainty: Objective and logical – errors will exist not caused by bias but
randomness, Based on reliable measurements
One school of uncertainty – Emotional our wants exceed our needs. A hierachy of values. How
much risk are you willing to take?
Utility theory. Pain vs pleasure. Value depends on utility.
Insurance a negative sum game. It pays to have in case of a catastrophe. Gambling more
entertainment than risk.
Insurance requires law of large numbers and independence.
Reducing uncertainty is costly.
How do we make decisions under uncertainty? Much of the information we have is incorrect or
incomplete. We never are certain? Certitude is elusive,
Maximin versus minimax.
Game theory – 2 or more people try to max their utility at the same time each aware of what the
other will do. Uncertainty lies on the intention of others
Risk management – optimize the outcome positive – deciding has some control and minimize
the negative outcome – little control