Hacker News Reader

Comment by in_cahoots | Hacker News Reader

This is the sort of sophmoric 'conventional wisdom 2.0' that's no more effective than the conventional wisdom it aims to replace. The thesis is basically, "most people think / act / look like X, therefore the opposite is more valuable."

That guy who looks and sounds like a doctor may have paid his own way through college. Maybe he went to speech therapy to overcome a stutter. Maybe he works out regularly and follows a good diet because he truly believes in the advantages of a healthy lifestyle and he knows how hard it is for his patients to do the same.

The point is, the vast majority of the time we don't have any way of judging how much a person has had to overcome just by looking at their appearance. It's not that we're drawing the wrong conclusion from the metric, it's the metric itself that's broken.

View on HN

> it's the metric itself that's broken.

That's the point. When a metric is broken, you can exploit its defects.

Let's say you have two surgeons applying for a job. One has the image down, one does not. Both have a 90% success rate in their surgery. The one with the better image will get hired more often. The one with the bad image will have to be better at his job (95% success rate) to get hired as often.

When you have a highly overvalued attribute, that attribute becomes an indicator of overvaluation. Meaning, it's a negative indicator of true value. The article applies this rule to "image".

If I told you that startups with a name that started with an "S" got 10x the early valuation compared to other companies, would you ever invest in a company that started with an "S"? Hopefully not. The startup may be good, but the odds are lower it would ever live up to its hype.

I agree with you in theory, but in real life there is nothing that is that quantifiable. What you are proposing is similar to video game stats -- if a character has 10 points to allocate, putting some in charisma necessarily means taking some out of intelligence.

In real life, I have no way of knowing which doctor is better at his job. The only metric I have to go off of is the degree and board certifications, and yet this article discounts those entirely in favor of the other guy. To use bedside manner as a metric instead of appearance, most of the brusque, rude doctors I know are just jerks, not real-life versions of Dr. House.

Adjusting for the relaxed standards of computer programming, proportionally I know just as many mediocre unconventional ones as I do mediocre conventional ones. The comments in this very thread suggest that the one with the worse image will get closer attention by many (though obviously not all) hiring managers just because of his/her appearance. So why should I believe the unkempt programmer was hired because he was more talented, and not because some other person just assumed he was?

> What you are proposing is similar to video game stats -- if a character has 10 points to allocate, putting some in charisma necessarily means taking some out of intelligence.

It seems similar, but it's really not the same. There's a comment in this thread suggesting if a surgeon spends more time fitting in, that means they have less time and energy to spend on being a surgeon. I totally disagree with this.

I don't believe there's some finite amount of energy (that you could call willpower or character points) that's somewhat consistent across people and you only get to spend in certain ways.

To keep overusing the surgeon analogy: let's pretend New York surgeons suck compared to Boston surgeons. For every measurable metric, Boston surgeons are better on average. Everyone knows this. Accent becomes a reasonable indicator of skill.

You go to a top-tier hospital in a different state. Everyone there is great at their job. You meet two doctors. One has a Boston accent (indicator of skill) and one has a New York accent (indicator of lower skill).

The suggestion of this article is to look at the New York doctor and think "What's so great about this guy that even while walking around with something that every single person is going to judge him for, he still made it all the way to a top-tier hospital? He must be damn good at his job to make it this far."

An attribute can be an indicator of value. But, if it is overvalued, it becomes an indicator of overvaluation.

When people are discriminated against but still successful, the discrimination criteria becomes an indicator of skill.

Shit, call it the Jackie Robinson Effect. When no one wants you around, when you don't look the part, you have to be damn good at your job to keep it.

I'm guessing here, but I'd bet money that in the first few years of desegregating baseball, the average black player in the MLB was way better than the average white player. They had to be amazing to get a spot. Being black in the MLB was an indicator of value. Now that those pressures aren't the same, I'd bet they're equivalent and being black in the MLB isn't an indicator of skill.

Clearly, I cannot choose the wine in front of me.

But you would have counted on this, so I cannot choose the wine in front of you.

No, I think you're missing his point. If people succeed despite the fact that their looks/behavior make others biased against them, then it means that they must be really good.

You're trying to make this argument without once considering that the situation is statistical. You're demonstrating that you can't KNOW these things, which is true, but certainty is not the only criteria on which it is possible to make a decision.

While there may be overvaluation of appearance/social ability, it's likely exaggerated in this particular example.

Appearance and social ability in surgical teams are important because surgeons are also leaders and caregivers. These attributes will matter to the members of their team and their patients. Team members who believe in their leadership tend to be more effective. Patients who believe in their care have better recovery rates.

If you did find such an example, it would be worth finding out what they are doing in other areas to compensate for the lack of this attribute.

> Team members who believe in their leadership tend to be more effective. Patients who believe in their care have better recovery rates.

There may be other factors regarding recovery rates perhaps counter-intuitive with respect to the presence of leadership (or appearance of leadership). Somewhat related:

https://hms.harvard.edu/news/startling-benefit-cardiology-me...

"High-risk patients with certain acute heart conditions are more likely to survive than other similar patients if they are admitted to the hospital during national cardiology meetings, when many cardiologists are away from their regular practices."

That is blackly humourous to me. Wow. Thanks for the link.

"The researchers found that certain intensive procedures were performed less often on the high-risk patients in the study during meeting dates than outside meeting dates.

One explanation for these findings, the researchers said, is that physicians who don’t attend the conferences take a more conservative approach for high-risk patients; another is that the physicians who stayed behind were reluctant to perform intensive procedures on another physician’s patients while that doctor was out of town. Survival rates might be higher because, for high-risk patients with cardiovascular disease, the harms of intensive procedures may unexpectedly outweigh the benefits."

This sounds a little like the programming conference effect. Where someone comes home excited about a new way of doing things, and applies it without first proving it out in isolation. Except people die.

I haven't heard about the programming conference effect. That makes sense. Proving it out involves exposing the theory to a wide range of adversarial inputs to demonstrate robustness.

There seems to be a similar trend in mortality with respect to getting treated in a teaching hospital vs nonteaching hospital:

http://health.usnews.com/health-news/patient-advice/articles...

At teaching hospitals treatments, may be very conservative as well (perhaps even more so than at non-teaching hospitals) with residents whose teaching is freshly learned.

Yes, this is a paradox that applies in business.

A VC may very well believe that you are brilliant, and doing great work, and can make all the right decisions.

But it matters to his investment that he believes that others will buy into your agenda.

Investors are incentivized to prefer a 'fast talker' over someone who is not, if only because they believe that the 'fast talk' will benefit their future investment.

Success rates are themselves a flawed metric. Doctors who take on more difficult cases will have lower success rates, but their experience will be more valuable.

Well of course. You're giving us two identical (for this purpose) people and then saying one has something more desirable. Naturally, all else being equal, that person has a slight edge.

> The one with the bad image will have to be better at his job (95% success rate) to get hired as often.

Or he could fast talk his way into preference, or get a drinking buddy to hire him, or countless other ways that don't involve being better at his job.

> When you have a highly overvalued attribute, that attribute becomes an indicator of overvaluation.

That is a hell of a leap. It is actually a strict logical fallacy, called affirming the consequent. https://en.wikipedia.org/wiki/Affirming_the_consequent

In other words this is a huge gap in your reasoning, it is claim - not a conclusion - that you are making and it is one that requires evidence. To put it another way: your argument is wholly conclusory and the only way to settle this argument is with evidence. It cannot be derived from first principle. There may be some instances where surface-level indicia (I can tell about the quality of a fabric pretty quickly, and it is typically indicative of good tailoring) are strongly correlated with quality and others where there is a low (a fresh paint job on a house) or even a potenitally inverse correlation (REALLY NICE ads for attorneys - located on bus stops).

But your blanket statement is just unsupported. I am not saying it is false - I am saying it is unsupported and you are jumping to a huge conclusion that does not follow logically, nor can it be proved to follow on the basis of logic alone.

That logical fallacy is not equivalent to what I'm saying.

    Assume attribute X is overvalued.
    Object A has attribute X.
    Object A is indicated to be overvalued.

This may not meet the strict requirements of a college logic class as stated, but I still believe it is true.

I'm not making a statement here about exactly what attributes are overvalued (fabric, paint jobs, bus stops or otherwise).

An argument of the article is that "image" is overvalued. I don't think the author did the best job supporting this argument.

The fallacy he's intuiting is really the implied denial of the antecedent:

  If X, then Y
  !X
  Therefore, !Y

As in:

  Attribute X is overvalued
  Object B does NOT have attribute X.
  Therefore Object B is NOT overvalued.

So in other words, your statement may be true:

> When you have a highly overvalued attribute, that attribute becomes an indicator of overvaluation.

but the implication of the inverse: when you DON'T have a highly overvalued attribute, lack of that attribute becomes an indicator of a lack of overvaluation. is not.

I think the main problem with the article is that it's conclusions are covered by signalling theory [1][2] but the author seems blissfully unaware that this field exists.

A signal of quality is likely to be reliable if it's inherently expensive to produce and is correlated to the desired metric. A big peacock tail is a reliable indicator of the peacock's health, because to falsely grow a bigger tail than it should support would end up being detrimental for the peacock. Same with lion manes, college degrees, etc.

The author points out that there are metrics and signals that can be overvalued, and that's true, but the hospital example is a terrible one because fellow physicians have access to the true quality of a doctor. A surgeon with a good degree, expensive watch, and nice clothing will not fool his or her fellow doctors because they actually see how they operate in the theatre. If a doctor X tells me that surgeon Y is good (and I trust doctor X), then that's sufficient evidence because doctor X almost certainly knows better than me, and I can entirely skip the process of evaluating surgeon Y's sartorial choices myself.

Where the author's point makes more sense is addressing pop-culture indications of quality, which are more likely to be incorrect because the general public (or the media industry specifically) is not an expert in any given field.

You probably shouldn't hire an engineer simply based on ideas of what hackers look and talk like gleaned from Mr. Robot, but I think most people already knew that?

[1] https://en.wikipedia.org/wiki/Signalling_(economics)

[2] https://en.wikipedia.org/wiki/Signalling_theory

The vast majority of time we don't have all the information for making decisions. That's exactly why we use heuristics that are relying on some things being more or less statistically probable.

He's talking about iterative/cumulative asymmetric choice markets.

And signaling.

That's all.

EDIT: I almost didn't read the article because of your comment. That would have been a mistake.

Reply on news.ycombinator.com