undefined | Better HN

0 pointslogicallee10y ago0 comments

I was watching this comment carefully as I thought I might have to delete it (due to people not understanding the arguments therein). It started rising to +2 or +3, so I stopped watching it. Now that I've checked again, I see that a few downvoters got it down to -1.

Since it is too late, I am happy to explain the thoughts in the above comment. I studied this area of economics from formal sources as well as having exposure to startups.

To summarize. One might think that if, say, a guy might be able to get a pizzeria off the ground, and it's 50/50, if it works he'll make a business with an enterprise value of $20,000 using his $5,000 investment, that's our assumption -- then if he wants to raise his sights and make the same investment, but this time is targeting $200,000 - his chances of success drop to 5%. And if he targets $2M then they drop to 0.5%. And if he targets $200M then they drop to $0.005%. If he targets $2B then they drop to $0.0005%. And if he targets $20B then it drops to $0.00005%. If he targtes $200B then it drops to 0.000005%

But 0.000005% is 1 in 20,000,000. There are 318 million people in America, so by those odds, the number of people in America who are capable of building a $200 billion business is exactly 15.

Yet there are more than 15 American companies that had those kinds of valuations in the past couple of years; Apple alone would account for 3 of them.

So with these kinds of large-scale successes, the risk-reward equation clearly doesn't actually work. And this ignores all of the home-run hits short of $200 billion. Clearly, risk does not increase at the same rate that reward does.

The risk-reward trade-off is just not a model that you can successfully employ when it comes to startups.

---

Secondly (this is not directly related to the above) the world is super-connected. There is 0 chance that if I find a pizzeria run by some guy in the city with 1 location as a small business, that 14 months from now it will have 50 locations that he personally operates. But if you find some guy with 2,000 downloads today from the Google Play store, there is no reason this same guy can't make something else and get 200,000 downloads. That is a factor of 100.

So the whole IDEA of a "small-business" just doesn't apply to online startup-like economies. The only question is whether the guy who made a niche app with 2,000 users is even interested in trying to use his same skills making something innovative and new, which he gets good press for, but which addresses the whole global smartphone audience.

The thing that makes it a startup is this hyper-growth component, like whether he's trying to roll out at that scale.

If someone doesn't understand either of the two (quite distinct) arguments above, or has any type of comments or rebuttal, I can follow up. Thanks for your time.

0 comments

3 comments · 1 top-level

roymurdock10y ago· 2 in thread

The risk/reward profile is the foundation of all of modern finance.

Risk is defined (in finance) as exposure to volatility.

Volatility is measured by calculating the standard deviation of annualized returns on an investment over a given period of time. The longer the time horizon of the investment, the more exposure to unknown and unaccountable variables you face.

Ultimately, this is because we do not have perfect information regarding the future. In fact, we have a very limited pool information that gets exponentially smaller the further out into the future we look.

In your example, the professor from 70 years in the future is much more valuable than the one from 2 years in the future precisely because the 70 professor has a much greater set of information about the future. Therefore he minimizes the risk that you will invest in the wrong technology because you already know what the world will need 10, 25, 50, 70 years down the line.

AstralStorm10y ago

Without any kind of actual valuation (tied to revenue) you cannot gauge risk. This is why startups in "pre-revenue" stage are essentially infinite risk.

Reward is also very guessed or estimated or both. So essentially the profile is made out of whole cloth as are valuations.

logicalleeOP10y ago

Obviously I didn't specify enough. The example considers current investment possibilities in possible startups. I asked you to imagine an entire faculty (actually two of them combined, Cal Tech and MIT) standing before you, asking for money. I then asked you to consider different versions of this same situation, corresponding to different levels of future-knowledge.

Let's further specify three groups, A, B, and a control C. A and B are asking you for $1 million to build their project. The control, C, is asking for $200K.

A is from 2 years in the future, B is from 70 years in the future, and C the control unlike A and B isn't a full cal tech + MIT faculty, it's just one guy with a youtube channel and no special knowledge, from our time. But they're all going to build their projects today if you give them funding.

A has proofs of concept and has just filed for patents but says they need funding to build prototypes, and for further patent filings. They say their patents and technology and the 2 years competitive advantage are worth $50M. But they will raise at $7M valuation, for a return of 7.14x over the investment timeframe between the current round ($7M valuation) and the next round in 2018 ($50M valuation).

B has nothing, but being from 70 years in the future, says they are presently worth $1 billion, but they understand it is difficult to raise money. So they are raising at a valuation of only $10M, selling 10% for $1M. They will use the money entirely on equipment and filing thousands of pages of patents, before their dazzling demonstrations allow them to raise their next round. It's the entire Cal Tech and MIT faculty from 2086! According to their story, after they show the dazzling new technology, they will proceed to raise a round in 2018 at a valuation of $1B ($10M valuation today, $1B valuation in 2018.)

C, the control, is some kid from today who is trying to raise some money so he can maybe raise at a valuation of $1.7M a year, or maybe 2 years, from now. Maybe. He already has some revenues from 5M youtube subscribers and he is raising a seed round to expand his channel into a whole media empire. His revenues are currently $5000 per month, or $60K per year, and he is trying to raise at a valuation of $600K or 10x earnings. He is only raising $200K. A year from now he hopes to raise a proper seed round at $1.7M, for 3x return. ($600K valuation today, $1.7M valuation in 1-2 years).

You are an investor. So, according to the traditional Risk-Reward profile, if A, B, and C are your possible investments, and A returns 7.14x (over 2 years), B returns 100x (over 2 years), and C returns 3x (over 1-2 years) and C already has an established product and established earnings, then the risk on B must be huge, whereas the risk on A is smaller and the risk on C is smallest of all.

But in fact, since 70 years is so far out, they have so much advanced knowledge, that the risk in that case is virtually 0 (there is no way that the entire faculty of both Cal Tech and MIT from 2086 with all their knowledge from them aren't worth $1B+ if they have $1M in financing and two years to spend on proofs of concepts, patent filings, etc! Yet they're raising at a valuation of only $10M. It's easy!)

Meanwhile, the group from only 2 years in the future is in a much riskier and more precarious position than the group from 70 years in the future. Their volatility is higher. This should be obvious.

While at the same time, even though the kid with the youtube followers already has a real product and real revenues, and is only trying to make a modest 3x return from a current valuation of $600k (10x earnings) to a valuation of just $1.7M -- in fact, his is the highest risk at all.

So this example shows that the least risk (essentially 0 risk) might be in a 100x return, the second-least risk might be in the 7x return. The control aiming for 3x return may be the highest risk of all.

I hope my thought experiments make it very clear to you that the risk/reward profile is flawed, and that there need be no particular risk associated with an outsize reward; likewise, the level of reward does not in any way need to match the exposure to volatility.

We are talking about the context of seed investments into startups, and I do not mean to generalize to other asset classes. The Risk/Reward profile is not an accurate statement of the situation investors are faced with in considering seed-stage startup investments.

Sorry about all the text, but it seems I didn't specify enough the first time around! Thanks for any thoughts and happy to hear your response.

j / k navigate · click thread line to collapse