That's fair. But precision isn't
really required, because it is a huge mistake to think that investors use precision.
It's possible to think that they do, and that they make a choice based on adding up the dollar-value of alternatives, multiplying each by its probability, and summing the results.
If that were the case then two things would be true (among many others):
----------------------------
I
First:
Debt that can be repaid would greatly increase the chances of a landed investment (investment being made in response to pitches), since there are many, many cases in which a small group of people formed into a company can repay an early debt, even if the company ultimately folds. All these cases would detract from the down-side. This would, in theory, tip the investment in the favor of being made. For example, if a company has a story that you might make 15x your money in 2 years, but you believe there is a 1 in 15 chance of this happening, then it should objectively make a big difference whether (or how many) out of the other 14 chances repay your money. If all 15 repay your debt, and 1 in 15 makes 15x, then that's a good investment, slightly beating the alternative places you could park your money. On the other hand if most of the cases end up losing 100% of your investment, that "should" make the investment quite a bit worse. For example if 14 in 15 lose the investmnet and the fifteenth makes 10x in 3 years, then that is not great.
But htis is not how investors actually make their decisions.
No investor will give a seed that has a 8 in 10 chance of being repaid (as debt). Period. They just don't.
Firstly, they are looking for that big, big win. And secondly, they want to skew the probability distribution toward that win.
An investor far prefers: (package 1)
0.0125 probability of massive, huge, breakaway hit: 500x
0.05 probability of HUGE growth: 50x
0.9375 probability of total loss (goes to 0 in 3-7 yrs)
------
0.0125 * 500 + 0.05 * 50 = 8.75x average.
Meaning: if you invest in 67 companies you will get one 500x growth story to woo your next set of LP's with. (1/0.0125 = 67), 1 in 20 of your companies will at least show a 50x growth story, roughly paying for the rest which silently go away within 10 years. If you have a $500M fund you can make 1250 seed-stage investments of $500K. Hopefully you can invest enough Googles with them.
to: (package 2)
0.05 probabibility of total write-off
0.15 probability of small win (e.g. 10% p.a., i.e. debt repaid)
0.459 probability of "failed" i.e. moderate growth (5x)
0.34 probability of successful growth (20x)
0.001 probability of 500x
-----------
0.15 * 1.1 + 0.459 * 5 + 0.34* 20 + 0.001 * 500 = 9.76.
Even though the first sums to an expected value of less than the second one.
Investors aren't trying to go for the second package - i.e. making a shot at a very sure 20x and maybe missing it and making 5x. That's not what they make seeds into, $2.5M companies they can believe in at $125K (20x).
It's just not what they're about.
That is not the story that is interesting to these investors. They would far prefer the first package (as far as I understand) that comes with a bigger shot at the moon, even though this greatly diminishes the overall return. Tons of companies are very close to being worth $2.5M, and not really more, and don't need much money. They're just not that interesting.
----------------------
II
Second:
A second way you can know that this isn't a case is that nobody would even look at an investment like (package 3, hypothetical)
Hypothetical:
0.0001 probability of 75000x
0.9999 unknown result
---------
This sums to at least 0.0001 * 75000 = 7.5x. But what does it mean?
Well, it means investing in something that will grow to 75,000 times its size with a 1/10,000 chance of doing so.
It means somehow onboarding 10,000 people - investing $200K into 10000 companies - that are all saying they will turn it into into let's say a $60B company.
Do you see anyone investing $200K at a $800K valuation into 10,000 nascent companies that are saying they will be $60B companies? And showing very, very little chance of doing so - 1/10,000?
Of course not. The amount of money required to invest in, say, 10,000 such companies is $2B. Taking $50K of it would cost $500M. Some VC's have this.
But do you see any VC's with massive, massive onboarding programs where they are pouring $50K into 10,000 companies, that each give you only 1/10 000th confidence that htye're actually the next Apple or Google?
No. It's just not being done. You can play with the numbers but they show that nobody is summing in this way.
They want a different distribution that they can believe in. They're fearing on missing out on something big. Not trying to play the numbers to make something absolutely extraorbital.