Hacker News Reader

Comment by crististm | Hacker News Reader

where did you get those figures?

I’m gonna go out on a limb here and say: they’re guesses, feel free to post better ones. Especially if it changes the conclusion.

So you are suggesting that (from their behavior) people are oblivious to the low risk/high reward indicated by those figures?

Or maybe they don't like assuming risks based on other's bogus data when they already have their version of bogus data that is probably closer to reality?

I’m not really sure who “they” refers to (parent, GP, the company, the employee?), but let me take a step back here and illustrate my point more clearly. Maybe it will help. What I’m trying to say is: this isn’t a court of law, it’s a collaborative conversation. Someone posits a hypothesis based on some estimates. Others are invited to build on either of those: the estimates themselves , or the hypothesis built upon it.

This is a healthier way of looking at online discourse than constantly asking people to provide sources and citations. If they didn’t mention them, you can safely assume it’s a guess. It’s implicit. Don’t like the guess, great: help us improve. We’re all in this together. It’s not a battle of “who has the best opinion”.

Maybe that’s what you were trying to do, in which case I’m sorry for misunderstanding. I genuinely didn’t understand your comment, please forgive me :)

The figures discounts the downside too much for them to work.

I was on the receiving end of no-reply and didn't like it. But if I were to switch sides to a company, then after a couple of honest attempts at feedback I will most likely say fuck-off and send a stock letter instead.

I made them up.

Of course they're made up. That's why I asked.

If you make a conclusion on made up data you get bogus conclusion.

It may be a good conclusion, but not for real life where facts don't match made up data.

> If you make a conclusion on made up data you get bogus conclusion

No. You work with made up numbers to understand the problem. Then you can make conclusions even without knowing the precise numbers.

For example, my analysis doesn't change much wheter the rate of lawsuits is 0.1% or 0.01% or 0.001%. It would change if the rate of lawsuits is 1%.

But I am pretty sure that the rate of lawsuits after interview rejection is much less than 1%. So I can make a conclusion without knowing precise numbers.

Calculations based on estimates come up all the time, and they are very valuable. They make it clear what assumptions your decisions are based on.

What's the alternative? You have to make a decision. If you don't want to use estimates, what are you going to base your decision on? Whatever feels right?

What I'm saying is that even with the lawsuit rate that low, there is no real incentive for the company (really the people sending the emails) to behave otherwise than they already do. Actually their benefit is that low, that even a slight error of that 0.1% guess would make the whole do-good business a really bad proposition.

Your (guess) data is probably right but it discounts too much the downsides. When you present your hypothesis to them (e.g. me) they will tell you (rightly so) to try it yourself first.

The incentive is “be a good person/entity.” This whole bottom-line approach and “what’s in it for me” is unfortunate to say the least. Behind every corporate establishment is a cadre of people. People with (possibly) spouses, children, non-deceased parents, neighbors, and friends. Possibly at some abstract level similar to the abstract “us.” Treat people like people, not some kind of legal liability.

Reply on news.ycombinator.com