In Defense of Bad Ideas

https://surjan.substack.com/p/12-in-defense-of-bad-ideas

by as89 • 6 years ago

101 48 6 years ago

surjan.substack.com

perl4ever3• 6 years ago

I'd forgotten about that stuff, but it seems like the bad idea was the prize itself.

I mean, the winner did just enough to get the prize, and stopped, and nothing further ever came of it, right?

Just because of how SpaceX overshadows everything in retrospect.

as89• 6 years ago

No, they're still going at it. Partnered with Virgin to make further iterations of the Spaceship One concept under The Spaceship Company. They're shooting to do private spaceflights under the name Virgin Galactic.

perl4ever3• 6 years ago

Orbit though?

jkhdigital• 6 years ago

Bad ideas are magical when they come in the form of long call options (bounded cost, unlimited upside).

Bad ideas destroy companies/nations/civilizations when they come in the form of short put options (bounded benefit, unlimited downside).

sudeepj• 6 years ago

This is indeed very well put. However sometimes due to nature of the idea it is hard to evaluate the extent of downside in advance and what may seem like a limited downside actually turns out to be worse.

082349872349872• 6 years ago

True. Also: when one is in a local maximum, bad ideas are the easiest way to get in the attraction basin of a better one.

thaumasiotes4• 6 years ago

long call options
short put options

I tried to think of what a "short call option" or a "long put option" would be, and I don't think the concepts exist. Do they?

ekelsen• 6 years ago

If you buy a call option, someone has to be on the other side of the contract. They are the short position.

ptr• 6 years ago

Generally a “long call” would mean that you bought a call option. If you sell/write the contract, you’re short. Your position is in the derivative instrument itself, it's not cascaded onto the underlying.

I also know multi-layered derivatives such as options on options and options on futures exist (as I’ve written code for them), but I’m curious what they’re used for.

pjc50• 6 years ago

Succinct and accurate. A couple of real examples: the famous LTCM https://en.wikipedia.org/wiki/Long-Term_Capital_Management blew up in one of the late 90s financial crises by doing something that was almost exactly that (writing a lot of puts)

The 2008 mortgage crisis didn't have unlimited downside, but an awful lot of parties were making tiny slivers of money off things they thought had zero chance of default whose default risk was both higher and correlated.

Nuclear power arguably goes in this category too: it's mostly great, which is why people quoting numbers of deaths claim it's safe, apart from the couple of catastrophes that have left huge areas of contaminated land.

If you're trying something that has no risk to others and the worst case for you is the loss of your investment, feel free to go for it. If the worst case affects others negatively or your liability might be more than your investment, you should definitely think twice.

Oh, and be very careful about margin investing and options trading for non-specialists.

whack• 6 years ago

There was nothing fundamentally wrong with LTCM's business model, or with writing puts. Where they failed was in their execution - they underestimated the risk of fat-tails, and the level of correlation between different securities.

There are lots of cases where we do things that could potentially be catastrophic, but we do them anyway because the risk-reward justifies it. For example, going on a road trip instead of sitting at home and watching TV. The upside of going on a road trip is bounded (fun holiday) but the downside is catastrophic (deadly car crash). If we avoided every single activity that involved catastrophic risk, our lives would be far worse.

high_derivative• 6 years ago

Not sure what this metaphor adds, especially because it's inaccurate.

A short put does does not have unlimited downside as the value of the underlying can only go to zero. A short call has unlimited downside as the value of the underlying can go arbitrarily high.

short_sells_poo• 6 years ago

For sure, but the return distribution of equities at least was historically highly negatively skewed. There can be upside volatility in stocks, but it has either no persistence or it happens just after a massive crash. Taking this skewness into account, it is not unreasonable to consider short puts more risky in a very handwavy and back-of-an-envelope calculation.

Of course, other assets with more symmetric distributions (FX?) might have no such bias.

marcosdumay• 6 years ago

Ideas should have a lot of "volatility", so yeah, this makes a lot of sense.

But of course, the usual environment of the financial world does not apply. "Bounded" can still be large enough to break a nation.

surjansingh2• 6 years ago

Great point. I wasn't even thinking of this in a larger context. I wonder if the downside is more important than the upside. So if it's just me trying to build an experimental aircraft, at worst I die. If I'm a city planner, I can make a bunch of lives miserable. Not sure what the upsides would be in either case, but the number of people affected by the downside is pretty easy to tell. So maybe the fewer people affected by a bad outcome, the more radical your ideas can be.

krm01• 6 years ago

Would have loved to read a longer writeup exploring that premise (which I agree with). The difference between Bad ideas and Good ideas, is determined by the outcome. Something that seems bad but ends up working (milliondollarhomepage, pet-rock, etc.) ends up being brilliant, in the eyes of the public.

cycomanic• 6 years ago

I would argue that Walter Benjamin's "History is written by the winners" (rephrased) is someone a longer write-up of the same premise, albeit not in a technical area.

PresidentChuck• 6 years ago

Yeah, all ideas are bad until they accepted. To create is to challenge the status quo

hliyan• 6 years ago

Reminds me of the strange aircraft designs of Burt Rutan, the closest thing the 80-90's had to an Elon Musk: https://www.flyingmag.com/photo-gallery/photos/awesome-airpl...

taneq• 6 years ago

If you're referring to SpaceShipOne, that's because Rutan's company built it: https://en.wikipedia.org/wiki/SpaceShipOne

adwn• 6 years ago

The article assumes that SpaceShipOne's design was actually a good design. It wasn't.

To be precise, it was a good design to satisfy the artificial requirements of the prize, but nothing else. The design doesn't scale. The hybrid rocket motor has been the source of countless headaches, lousy performance, and more than a decade of delays, the feathering mechanism has already cost the life of a test pilot, and the need for a pilot plus co-pilot severely reduces the payload capacity.

Now, the article's thesis – that good and bad ideas are sometimes hard to distinguish before they've been tried – is still true, but the author chose a bad example.

darkerside• 6 years ago

It's a great example. The only reason these factors clearly outweigh the positives is because the benefit of hindsight. Or do you imagine the brightest minds of NASA just weren't thinking clearly for decades?

The killer feature of the shuttle was supposed to be the cost savings from being able to make multiple trips using the same equipment (sound familiar?). It turned out not to be a feasible solution from a technical cost standpoint, but if people can't clearly estimate the cost of adding a new feature to a website, I totally believe would have been indistinguishable from a good idea at the time.

pdonis• 6 years ago

It turned out not to be a feasible solution from a technical cost standpoint

The shuttle turned out not to be a feasible solution because it was loaded down with other requirements that had nothing to do with the killer feature you describe. A key one was the requirement from the US Air Force that the shuttle have enough cross range capability to be launched into a polar orbit and then return to the launch site after one orbit [1]. That was what drove costs up to the point where the shuttle wasn't feasible any more in terms of the killer feature.

[1] https://history.nasa.gov/sts1/pages/scota.html

surjansingh2• 6 years ago

As the author of the article, maybe I can explain why I think it’s a good example. I’ll just say - I absolutely could have chosen something that was more of a slam dunk great idea (in hindsight) with tons of follow-on success. I mostly chose it because the Rotary Rocket and SpaceShipOne sit next to each other in Mojave and got me thinking about this in the first place.

As you said yourself, it was a good design to win the prize. That to me is the key. It did everything it set out to accomplish. Your comments all relate to the scaled up SpaceShipTwo - a completely different vehicle - which I agree that the jury might still be out on. But to me, that’s like saying that a hang glider is a bad idea if someone tries to make a hang glider for ten people. Is it the hang glider that’s the bad idea or the fact that someone’s trying to scale something that maybe doesn’t scale well?

That it was the first private crewed spaceflight shouldn’t be underestimated either. I think this list of human spaceflights really puts it in context: https://en.wikipedia.org/wiki/List_of_human_spaceflights

Thanks for reading and for your perspective.

bluedino• 6 years ago

I've always wondered how terrible automotive designs make it to production. Engines or transmissions that quickly fail, cars that are just garbage in quality...nobody along the years of development stopped to say "maybe we shouldn't build this..", or perhaps someone did?

Did the company building this get their moneys worth when it comes to lessons learned?

Cthulhu_• 6 years ago

When it comes to mechanical failures like that, it's never due to incompetence; it's due to cost cutting concerns (e.g. using less material, cheaper / inferior material) and / or planned obsolescence. Automotive engineers are highly competent, and most will have had a lot of education when it comes to materials, strengths, stresses, wear and tear, etcetera. And car companies build up that knowledge over the decades.

Definitely not a problem with all cars, but the better built cars are more expensive.

I want to say "you get what you pay for", but at the same time I know nowadays there are a lot of products (e.g. electrical tools) where the brand is bought up and the new owner makes big cost cutting measures to maximize their profit margin, the money coming from "brand inertia", that is, a brand is known for being good for a couple years after they start reducing quality and durability.

reactor4• 6 years ago

Nissan's CVT seem to be a counterexample of everything in your first paragraph.

I could also bring a lot of counterexamples to your second paragrah (e.g.: many german cars).

dognotdog• 6 years ago

I'd say it is due to incompetence, but not as much the individual engineer's as failures of the organization, though I have seen plenty of people not doing the job they're qualified for. These failures stem from plain mis-management, political turf wars, and misaligned incentives -- cost-cutting being the most obnoxious driver, causing a landscape of balkanized teams to fight for resources and engineering concerns becoming an aftertought.

jotm2• 6 years ago

At least for complicated mechanical devices like transmissions, or really anything that has a lot of friction between metals, getting the right alloy is essential.

An example from the world of home appliances, Bosch/Siemens have been using the same drum design for a decade, before production was moved from Poland/Spain to China, likely for cost cutting measures.

The Chinese made machines (same internal design, new model - they always do that, with a facelift and maybe new electronics) had a very high bearing and spider failure rate.

They switched the alloy supplier. Makes sense, why import it from the EU. Either they knew it was lower quality, or were duped, but anyway, their product quality took a massive dive because of a seemingly simple to produce part.

jotm2• 6 years ago

That rotary rocket looks like a design that could work, tbh

Instead of rocket propulsion, first stages could in theory deploy helicopter wings and get a decent slow down even before spinning them up to full speed (with autorotation the blades act like a parachute/wings).

Seems like it would be a harder implementation than the high precision control with a rocket, though.

wojtczyk• 6 years ago

Does anybody know, why the rotary rocket body did not start spinning itself once it lifted off the ground? There’s only one propeller.

pathhandwaving• 6 years ago

Some ideas are just bad Example: Should I take this unmarked pill a stranger just handed me? No you probably shouldn't Example 2: Should I jump off this cliff with rocks at the bottom? Is it higher than 2-5 feet. Probably not.

Example 3: Should I eat this raw meat product that has been sitting out on my counter for more than a day. Probably no. I understand the point here, but some ideas are just bad ideas.

In Defense of Bad Ideas

Related Stories