Story Points - a thread.
SP's can work. But only if you don't try to explain them, define them or formalize them. If you do you'll instantly cause them to collapse into a costly mistake with an inbound paradoxical dimension that will suck all joy out of your existence.
SP's can work. But only if you don't try to explain them, define them or formalize them. If you do you'll instantly cause them to collapse into a costly mistake with an inbound paradoxical dimension that will suck all joy out of your existence.
SP's as envisioned is a cheap and simple way to do estimation and gauge delivery capability over time. They can't deliver on that. They either won't work at all. Or will no longer be neither cheap nor simple.
So let enumerate what most folks tend to think when they hear story points:
1) They're numbers.
2) Showing Relative Effort
3) Higher Numbers Have Greater Uncertainty
If you believe this then according to the principle of least surprise you've invented a horrible abstraction.
1) They're numbers.
2) Showing Relative Effort
3) Higher Numbers Have Greater Uncertainty
If you believe this then according to the principle of least surprise you've invented a horrible abstraction.
SP's are supposedly numbers. But you can't do math (the way you think) on them!
In the small the problem might seem negligible but every single team I've met using SP's have had superstitious beliefs around story selection that could be traced back to misunderstanding that fact.
In the small the problem might seem negligible but every single team I've met using SP's have had superstitious beliefs around story selection that could be traced back to misunderstanding that fact.
Addition of distributions is significantly different from addition of numbers.
This is why although your have a numeric "velocity" (a sum of numbers) most will also argue that 100 one points stories are different from one 100 point story.
...
...
This is why although your have a numeric "velocity" (a sum of numbers) most will also argue that 100 one points stories are different from one 100 point story.
...
...
Imagine yourself placing either many small bets with high certainty or one huge bet with very low certainty. Those are economically *very* different.
Each story is a bet. The size signifies both the cost of the bet and the certainty for a payoff. Problematic.
Each story is a bet. The size signifies both the cost of the bet and the certainty for a payoff. Problematic.
So they're numbers that can't be added.
They're also numbers that really can't be relative.
Or if they are relative. Then they can't be distributions.
And if they are indeed truly relative they're easily converted into hours.
They're also numbers that really can't be relative.
Or if they are relative. Then they can't be distributions.
And if they are indeed truly relative they're easily converted into hours.
In practice most teams establish a heijunka like system where they have polices like "max one X point story each iteration" and then they pad around that big bet with some controlled distribution of smaller bets.
We're now in a distinctively non simple scenario.
We're now in a distinctively non simple scenario.
