In Defense of Bad Ideas | Hacker News Reader

In Defense of Bad Ideas

surjan.substack.com

101 pointsas8949 comments

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).

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.

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.

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.

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.

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.

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.

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.

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.

> 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?

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

Sure, but they don't have a call option. All they have is an obligation.

Short just means they have a negative amount of that, in this case a negative amount of options.

No, short means your position is predicated on the idea that the future price of something will be lower than the current price.

Owning a positive number of put options makes you short the underlying stock, which is presumably why the comment upthread described put options as "short put options". Selling put options gives you a long position, not a short position.

I wanted to analyze this in terms of the original comment, in which "short put options" have limited upside and unlimited downside, but put options do not have unlimited downside for either the buyer or the seller. The buyer can't lose more than the price of the option, and the seller can't lose more than the strike price. Put options are a case of "limited upside, limited downside" from any perspective.

Call options are the ones with unlimited upside (for the buyer) and unlimited downside (for the seller).

You have to think of the futures definition of long and short, not the equities definition of short. In every futures contract, the long is paying money and the short is receiving it for the contract. Every options contract has a short, the writer, and a long, the buyer.

In this case, the original commenter explicitly (erroneously) stated "options". Options cannot have unlimited downside; you can always choose not to exercise the option.

If you write a put option, it has downside limited only by the underlying zero price of the security minus the premium. If you write a call option, your downside is theoretically unlimited. If you buy a put or a call your downside is limited to the option premium.

edit: i should say writing a put option (short put) your potential downside is the difference between the strike price of the option and the premium, if the underlying security goes to zero.

If the underlying security goes below zero, that doesn't increase your potential downside. The option you sold will force you to buy the security, but it won't force you to sell it.

Thanks for clarifying

> short means your position is predicated on the idea that the future price of something will be lower than the current price.

Yes, but in this case the "something" is the option, not the underlying. This is completely standard use of the word "short" in the finance industry.

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.

Options on futures are mostly just a convenient way to specify an option on the underlying product. Instead of having the option to buy gold bars (which trade on a different exchange), you have the option to buy gold futures, which has 99% the same behaviour in terms of price.

Options on options are pretty esoteric, I've never heard of anyone actually using one. I guess you can get very leveraged exposure to volatility that way, if that's what you need.

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.

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.

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.

> 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

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.

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.

Orbit though?

Sub-orbital. Think getting the craft up to altitude of 100+ km, hang out for a few minutes, and descend back to Earth. SpaceShip One, the prototype inspiring the Virgin Galactic fleet, was capable of reaching speeds of 0.9 km/s. [1] Escape velocity to get into orbit from Earth's surface is 11.2 km/s. [2]

[1] https://en.wikipedia.org/wiki/SpaceShipOne

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

Escape velocity is a velocity to escape gravitational field. Orbit velocity is a velocity to orbit a thing inside its gravitational field.

For example, ISS has orbit speed of about 7.7 km/s at 350 km above the ground. Orbit speed "at the surface" of Earth is about 8km/s, I believe.

Thanks for the clarification.

My point, not to be overly disrespectful, is that suborbital flights are so different from what SpaceX has done that they seem like party tricks, not stepping stones.

In the big picture, it's not a huge amount wasted, I was just remarking on it being less significant in retrospect than one might have hoped.

What about it?

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...

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

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?

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.

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.

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).

"better built cars are more expensive" is not the same as "expensive cars are better built".

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.

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.

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.

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

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.

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.

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

Reply on news.ycombinator.com