E286: How LPs Can Actually Find Alpha in Venture

24:57There is actually, there is no way to know. The way to know is to write a check to see that company come to life and actually see if what was in that deck was true or not, or if there was a business somewhere in what was described in the absolutely and totally inaccurate pre-seed deck. That I think is the way this world works. it's not something where like oh i found out like i have all this private data and i like managed to like create this model and this is like the thing that i want to invest in it's like that's not correct that's not a correct model of what's actually happening here said another way in the precede round the true traction is so sparse that it pales in comparison with what that founding team will do so whatever you're investing in whatever thesis whatever is in the deck is such a small percentage of what will become a company that you might as well have just invested in the founders without an idea, but it would essentially be the same exercise.

25:47I would say your perspective is a little bit closer to the less radical epistemic version of like, oh, if I just knew what would I, but I actually, I think, I think I've gone full von Neumann. I think I've gone full ontik. I think there's actually like, you know, the world in which you write a check or in the world you don't write a check, those are actually different universes. But doesn't that assume that it's the lead investor? In other words, if you were the lead investor, you'd... Sure. Although I think, yes, but I think the investment traction is something where like, oh, we got a commitment for this person, whatever.

26:17Like I do. And then I think you actually see, I mean, you observe this behavior in terms of seed investments as well. I mean, I've seen it for companies that I've started where before you have kind of more than half the round committed. And then at that point, VCs start looking at this as not a question of like, should this company exist or not? But rather like this company exists, do we want to kind of go on this ride or not? And that to me is like a really, like, I think that sort of illustrates the kind of the power that the earliest stage investors really have. That's not to say, are there pieces of data that would be incredibly valuable for seed stage investors?

26:54There are, but they're not attraction metrics for a company, right? If you had access to Alfred Lynn's email and you could know every single startup that's coming up for a partners meeting at Sequoia, that would be a very, very, very valuable thing to know. But that's very different than I think what – I've never seen the pitch of a data-driven VC where they say that they've hacked into top partners' emails of various VCs and are going to front-run them. That would, you know, not because they had to invest in that because it's probably legal, but like that's the kind of data-driven strategy that you'd want, not like, oh, we scraped LinkedIn.

27:35It's like. Going back to this hypothetical thousand portfolio, let's just make it easy and say that you invest a thousand dollars into a thousand credible startups. So they pass a certain threshold and you invest into each one of them. How would that rank versus a concentrated seed portfolio? Would that be top quartile? Would that be mean? Would that be bottom quartile? And how do you think about kind of that versus doing, say, a fund of fund or investing to a GP? Yeah. So it would be the mean almost by definition. And what you tend to see with venture capital is that the mean is around the 75th percentile of funds.

28:17And I think one of the challenges with this asset class and investing in this asset class is there's a lot of navel-gazing, ego-stroking around like, what do the best investors do? And when you look at the strategies of the best investors, they don't invest in the way that I talked about. What they do is they'll make extremely concentrated bets. They'll follow on huge. And what you see there, I think what you're observing, are just folks who took on risk and got lucky. Risk cuts both ways. It's not always like, oh, that was a really risky investment, so it was bad. Risk can be beneficial. And there's enough investors who do that kind of investing that if you look at the top decile of early-speed investors, they will all have that profile.

28:58They'll all have the profile of making concentrated bets, of doubling down on their winners. The challenge is that that's not – it's not a replicable piece of advice to tell someone of like, oh, yeah, the best way to do seed investing is just get really lucky. Like, it comes out in the data. How do fees play into that? So if you're investing in a two and 20 fund versus let's say you make those thousand investments direct, are you now above that top quartile? How should one think about that? I don't think fees play a significant part of the story.

29:33And that's because, well, interestingly, the fees that play the most part of the story are actually the management fees and not the carry. And the reason the management fees play such a large part of the story is that management fees cut into the amount of money that actually gets invested. It's so funny. This is all a consequence of the power law, right? The venture capital as a just historical is so driven by the vintage years that have been super successful. And those are, you know, and so you have this dynamic where, you know, the real returns of the asset class are being driven by the years in which it'll be the venture capital average will be, you know, 5, 8, 13x or whatever over the lifespan of these funds.

30:19the manager fees actually are super expensive in those cases because you're not getting you know oh that vintage year every dollar i put in 10x you know it actually becomes really expensive if you have a 20 percent manager fee because that's that's you know you're literally like losing several multiples of the fund just from manager fee and i think this is actually something where my perspective has changed in turning angel list is like when i join you know i have this my friends and i were doing this investing out of this fund called indicator it's just our money we don't have LPs. And, you know, it's like, ah, I don't need to invest in someone else.

30:49Like, I don't need to pay carry to someone else. Like, I know how to do seed investing. I'm doing great. Like, and I think one of the things that really changed my perspective was, was, you know, kind of seeing fund performance and research into fund performance. And, and frankly, I'll give him credit here, just directly meeting Ryan Hoover, product hub founder of this one. I started AngelList. he would have yeah he's an awesome guy and i just you know he he's like really good at consumer product investing right he founded product on he's so plugged into that world and like he's just you know he's he's got the like he's just the consumer facing investor like he's just really good at it you know if he like you know we tried an indicator you said with your stuff they were always terrible investments like by far like the very difficult space very very hard especially if you're like where indicator was successful or has been successful hopefully has been in kind of deeper hard tech investments and so you know when we were seeing like i thought we we tend to see like pretty good like companies that want to do something in space like we tend to see a lot of those opportunities that was like pretty good by the time a consumer company like gets to us like Like, man, everyone's passed on those guys.

32:03And like, those are bad, they were just bad investments. And I had this perspective, like, you know, I can't do this kind of investing. Ryan can. And like, if he's making me money, why shouldn't I pay him, Kerry? Like, and so that has been a real change in perspective of like, look, I don't adequate coverage for my own portfolio to this space. And so, you know, I did the investment in We Get Fun. Another one that I did was N49P run by Alex Norman, who's kind of the Angelist Canada guy in Toronto. It's like super plugged into everybody. Every Canadian doing any kind of startup or startup investing goes Alex Norman.

32:37You know, I don't, I'm not plugged into Canada. I was like, okay, this guy will give me differentiated deal flow into a narrow area of expertise that I'm not already exposed to. And if he makes me money, I don't mind paying him carry. And so that has been my real kind of perspective shift on this is that I don't really, I don't, you know, obviously fees are a zero sum game. $1 paid in fees is$1 less of returns. But it really seems like given the dynamics of the space and just like, frankly, the inability to replicate the quality of the portfolios, you know, I'm not saying like, hey, if you're a seed investor and you have a certain slice of deal for it, whether it's like, oh, I used to work at Uber.

33:20So everyone, ex Uber, I'm not saying you shouldn't like funds that overlap a lot with your existing investments might not make sense to do a fund investment. But like, if you have, if you have access to funds that are seeing that are high quality and seeing differentiated slices of space, you don't see, I have a really, really hard time. Oh yeah. Like, but I see, I see some consumer companies. It's like, dude, like Ryan Uber sees you're not on his level. Like there's very few people who are at a V, if he's giving you the opportunity to investor in his fund, then you don't have to do any more consumer, but you don't have to do any more shitty consumer investments because I got Ryan Uvers.

33:53So per your model, you believe that the market isn't efficient for the deals that get done, but there is still adverse selection. Oh yeah. So I think adverse selection is really maybe the entire story about venture capital. I think that's actually the most important concept in all venture capital. So we have some research that's going to be published pretty soon that actually tries to unpack adverse selection from the perspective seed stage check sizing. So think about it this way. So again, you're a seed stage GP, you're writing checks, and your check size that you typically write in a seed stage company is that you have your typical check size.

34:32Some of the time, actually it's generally about two-thirds of years, you'll kind of observe this behavior where you'll either have a big check, which is like more than twice as large as your typical check size, or you'll write a small check, which is less than half your typical checksets. And what's cool about the Angelus data is we are looking at a data set that's 15 ,000 seed investments. And so like there's no, it's very unlikely that you will pull out any other conclusions other than the ones we found because we have access to the, you know, such a huge slice of seed investing universe that like this is what's happening in the space.

35:06So it's not like, oh, we have this anecdote about this one small check we wrote. It's like there's thousands of investments that are going behind us. So what we found was actually twofold. I'll put it this way. In a normal asset class, if you approach this from a, like, this is what quantitative finance was defined around. I'm going to hire a bunch of PhDs. I'm going to come up with these signals. And then what I'm going to do is I'm going to more or less, you know, depending on if you actually precisely articulate this in the case of like some pretty decent, like reasonable assumptions, you pretty much allocate proportional to the quality of the signal of the company.

35:37So if a company is great, you want to put a bunch of your money in. if your company is less good, you put less of your money behind that. And that gives you the, if you want to interpret it as like the highest mean with the lowest variance, the best risk return profile for your portfolio will be from investing more money behind better signals and less money behind weaker signals. And in general, if you have people who have skill, which like we think, yeah, we think our GPs as a whole do, but you should see this phenomenon where big checks outperform small checks. Just because big checks are made with conviction and all the signal, small checks are not.

36:14What you actually see is kind of the opposite. So what we see is that consistently small checks are the highest performers. Now, it's not consistent in like every single small check is the best investment, but like it definitely shows in the data. Small checks are the best investments. Small checks relative to a typical check size for that GP. So that again, a small check is less than half of the typical seed investment check that you'd right that year. Those are the best performing investments. As an investor, I'm sure you know the thing there, right? Similar thing where the return of an asset class is inversely proportional to how close the investment is.

36:47So the VCs that are investing in Austin from San Francisco, those returns tend to be better than San Francisco, San Francisco. And the thesis is that if somebody's willing to jump on a plane and go to board meetings, they must love it that much more. Is this the case here where they love the investment so much, they're willing to take a smaller percentage of the company because they see the upside? Is that kind of narrative? Yes, but really it's because they're being squeezed down by better investors, right? That's what's happening. These become the very hot rounds where there's a lot of interest in participating.

37:17And we know what people's typical check size would be. And they're willing to write a, instead of a 250K check, they're willing to write a 100K check to participate. Now, that's one side of the market. The other side of the market is the big checks, right? Someone has a typical check size of 250K, they're going to write a 750K check in a seed round of a startup. So we did also check, oh, is this like an ownership percentage thing? It's not an ownership percentage thing. We have a lot of, it's something different. So you would only write, if you're a typical check size 250K, you're writing a 750K check with this company, you got high conviction, right?

37:51This thing's going to be a winner. What we see from the data is that big checks are like maybe a little bit better than typical size checks. And so what is happening here? What's happening here for the big checks is that you have this adverse selection problem where you are writing a big check, presumably in an area of your expertise. You know how to pick companies. You can always just opt for writing your typical size check, but you have opted to write more than two size, two X your typical check into this company. What's happening? Why are you doing that? You have so much conviction that this thing's a winner, but your conviction is almost totally balanced out by the fact that that capacity exists in the first place, that a better investor than you is not taking your capacity.

38:40And so that I think is one of the most fascinating parts is that what we observe in the seed asset class is that the individual idiosyncratic signal is worth almost nothing. And the common signal is worth everything. And so this is why I'm saying like, oh, you know, it's not a question of data about finding out weird facts about a company or whatever that other people don't know. Your idiosyncratic signal doesn't really matter. What matters is the common signal of everybody's kind of collective ability to evaluate the quality of a company. And so that I think is, you know, when you think about what the seed asset class looks like, that's kind of a beauty contest dynamic, right?

39:17That makes the asset class look a lot more like vintage Rolexes or sneakerhead Air Force Ones, where everyone kind of collectively agrees like, oh, yeah, this kind of Rolex is worth more than that kind of Rolex. Even though you're like, for someone who's outside that space, you're like, I don't know, their watches, they both work. They're both 50 years old. They seem the same. You're like, oh, no, that one's worth three times as much because everyone sort of agrees that it's worth three times as much. That's something I think is really fascinating about this asset class. is like you don't have the dynamic of having a bunch of like dedicated researchers finding all this super interesting information to come up with this idiosyncratic edge that you would in bond investing or in hedge funds.

40:00You have something that looks a lot more like trying to buy Air Force Ones and trying to get like a cool colorway for Air Force Ones. And that I think is like what is one of the aspects. So the adverse selection side of that is really, really huge, Right. The fact that like capacity is available for you to take is in and of itself a negative signal where you means anyone who's not his name again, anyone who's not Alfred Lind. The fact that capacity exists for you to take is a negative signal about the company. So unless you're a top 10 % investor in that domain, in that space, you having capacity, you should have the reflection to sit around and say, why am I getting this capacity if I'm not one of the top 10 % investors?

40:46It's probably a negative signal. And can my conviction overcome that negative signal of me having capacity? And you should, you know, if it's kind of outside your area or something like you probably your convictions probably wrong, like the correct the correct thing to do, even if it looks I mean, this is something we ran into an indicator with our consumer investments, like we didn't know what we were doing. And maybe we wrote a bunch of bad checks, like, but they all look great. They all look better to us than the checks we wrote into deep tech companies. But like the fact that this company was being offered to us and that, you know, we didn't know we should not have made those investments from the perspective like we don't have.

41:18I think it's actually a conscious signal because what it's saying is that the valuation is being driven by popular or generally acceptable narratives that are accepted outside of the industry. And that it's essentially almost the definition of dumb money. None that's dumb, but that it's not specifically smart in that domain. So it's like a conscious signal. It's not only should you not be investing, but you're probably investing at a higher valuation. Yeah, I think that's right. If you've done hard tech for 10 years, and let's say you've done space technology for 10 years, and you know everything, something that is obviously not good for you, it may still have a narrative that might work for everybody else.

41:57But if you're passing on that, and if people like you are passing on that, either the entire industry is wrong, which happens once every, you know, several decades, although that happens much less within venture, it usually happens, you know, NASA might be wrong, but a bunch of space VCs are unlikely to be wrong. All things being equal, I think it's a contra signal. It's worse valuation and worse signal. there is as well an aspect when we think about adverse selection i think where this can manifest itself most principally in terms of investing in gps is the concept of style drug where like you know lps don't like it when gps are doing something different or whatever and and i think a um a chunk of that that hasn't been articulated is you know if you if you're pivoting to a new space if you have a new thesis oh now i'm gonna like now i'm all about clean tech and I'm going to do, you know, carbon is the number one issue of our time or whatever.

42:46If you're not doing that pivot, like, you know, you were a SaaS investor. Now you're doing carbon investing. If you're not doing that pivot in such a way that you're going to see the same sort of rank order list quality of companies, it's not good, right? You don't, your LPs may be less upset about, like, I disagree with your thesis about where you're going to find returns. And it's like, you're just going to see worse clean tech companies than you saw SaaS companies. And so like that, that will be the issue with your style drift, not, oh, you're changing your perspective on where venture returns are going to come from for the next decade.

43:19Adverse selection explains so much actually in terms of the, of the behavior. And actually, you know, we had an experience like this when we first raised money for our first startup, which we had a VC do like very deep diligence, actually like correct diligence on us and like was like, oh, I think we're, you know, we're in a past for these reasons or whatever. And then like the round came together and a bunch of other, you know, VCs, good names were participating. And the VC who did this diligence came back and was like, oh, can we like write a check, you know, just for the optionality of participating.

43:47and at the time it was like completely baffling to me because of like well you did all this you they did the best diligence they could have done and it was really and they were accurate actually in terms of the success of the company but they still wanted to participate and i think it actually as as insane as it sounds this phenomenon actually kind of makes sense from a data-driven perspective which is that like the company has a pretty good common signal maybe you're maybe you trust your idiosyncratic signal less, maybe you still kind of participate just in case the common idiosyncratic signal actually turn out to be wrong.

44:20And that I think is actually, it's weird to me because that was an experience where I was like, boy, this doesn't make any sense why you'd ever do that much work and come to the correct conclusion and still like go against it. And now, you know, years and years and years later, 10 more than 10 years later, I have this perspective of like, actually what they did kind of make sense. If it is about adverse selection, if it is about the common signal that VCs share, you should trust yourself a little bit less. It's fascinating to see how my perspective has shifted on this. A lot of the practices that I thought were completely incoherent and irrational, honestly, sort of makes sense.

44:56And I think this goes back to the original point, right? I think VCs are actually doing a pretty good job. Almost pains me to say that. But yeah. So the good news is that VCs are investing. The bad news is Alpha is very difficult. I know you're a purist, so I know whatever you believe you are productizing and you do. So you're CIO of Strawberry Tree Management Company, which is an independent affiliate of AngelList. What strategy are you using with your own money and in your business? So Strawberry Tree is, we're independently operated in the sense that AngelList people don't dictate how we do investing.

45:30So we're an RIA. And so our kind of mandate is to do the investing on the AngelList platform that AngelList itself can't do as a ERA or as a non-RIA. And so one of these is a fund of funds. And so we started a, last year we started a fund of funds to invest in the best GPs on the platform by anticipated future performance. And I think we had to develop like quite a few. I go back and forth on defining whether any of these metrics are novel or not, right? What we try to do is mechanize insights about the way you pick GPs so that we just put them into Python. And every month we get a dump of, here's the funds that we should invest in.

46:08We go out and invest these funds. You know, we're a funder now. We've done more than 60 fund commits. So we've done more than one fund commit a week since our initial close. And I believe we are, I can't prove this. I don't know. but it is my belief that we're the most prolific venture fund investor over the past 15 months, anywhere, of anybody, from a team of two people and a Python script. And so, you know, we're trying to invest fairly broadly, but that's kind of the strategy that we have been pursuing because I do think to a certain extent, and okay, how did I get to this point, right? When I joined AngelList, I think I had the idea of like, hey, I'm going to use all this quant stuff.

46:45You know, I'm a quanti guy. I used all data. Let's use it to pick startups. And like, I don't think that works anymore. I just don't, for all the reasons we've talked about, I don't think it exists. But the most prominent reason that it doesn't exist is pricing. The issue is, you know, the company that's raising it 50 might be 11 or 12 times better than the company that's raising it five. But like, there's a lot of variability in the space. Like you're not going to harvest sustainable alpha by doing that kind of investing. And then their access and adverse selection becomes a huge problem as well.

47:12So if it's not startups, then I was like, hey, well, what else do we have? So we have funds. The interesting thing about funds is that funds more or less charge the same thing, right? You know, 1 % admin or management or 2 % kind of fees and 20 % carry. And they all kind of charge the same thing. So if you have, you know, positive quantitative signals, you can kind of directly convert those to alpha in a way that you can't with startups, right? I can do all this fancy modeling and say, like, oh, this is a great startup. But, like, everyone else sees those signals, too, and you're paying for it.

47:43With a GP, if you have some metrics that, oh, this person's really great, they're charging the same. And that's actually even different than I know, you know, top hedge funds will charge, you know, zero and 50 or three and 30 or whatever. You don't see that kind of price discrimination a lot of the time in venture. And so it's what we sort of, you know, in general, I think fund to funds are terrible financial products. I do think, you know, Fortisworth, I think we've built something that is not, I wouldn't be doing it if I thought it was a terrible financial product. I didn't start with the idea of making a fund of funds.

48:14It's where the data has guided me in terms of where the data-driven alpha in the space is. And it's not from picking startups. It's from picking the GPs who pick startups. That's because those GPs will see opportunities without being adversely selected and in theory have some kind of alpha. The core idea is that like, okay, can we find a signal that has some positive relation to the performance of the fund we're going to be investing? Okay, we have those signals. The magic is you don't have to pay for those signals, right the the the gp who has those signals is charging you the exactly the same as a gpu doesn't have those signals it's not priced in in the same way that the like yeah this repeat founder of a successful company you went to mit like that's priced in bro but like the gpu is doing their fun two or fun three and they had some decent success they're raising a little bit of a fun they're still charging you know one or two and 20 and the way to rephrase that is there's some signal, but either it's not generally accepted or it's some access alpha.

49:11In other words, there's something different about their picking, but it's not priced into the general market such that the price is higher because of this insight. Correct. I think it's really for maybe three reasons. So I think we have three reasons. Maybe start with one of them, which is that in general performance, it's pretty random. So when we do an investment for us, we have written these letters for people, for their data works about why we're making the investment. And GPs generally like that because like, you know, they get this, they get a very quantitative letter that show other LPs like, oh, these guys think we're good from the data.

49:44But what we say in the letters is that every fund we invest in has an above average chance of being an above average venture capital fund. That's the level at which we're willing to say, right? We're not writing checks to like, you know, oh, this person is the best person on the platform. They're so amazing. like they're I'm giving them all our money like it's like we think they have an above average chance of being an above average fund and so one of the reasons you don't see price and discrimination is like and and like we've invested a serious amount of time and money and research time and ingenuity into coming up with our signals which I'm happy to talk about it fine like I'm I'm delighted to share those if you want to go to those but like even with all that sort of data understanding the edge is pretty small and so it's like what that means practically is like people like someone's like it's very hard to have a level of consistent outperformance you see this for big funds too that publish it like that like you actually can can get three and 30 from the market that people are willing to pay that because like you know your next one might not like there's a decent shot that it won't be above average and then and then you just burned everything right so i think there's that gps themselves because the inherent vault variability of space don't have a lot of price and power in the same way that like a top quant hedge fund might, right?

50:54Hey, this is the kind of trade we're doing. These are realized returns. We've been crushing this for 11 years. We've never had a down quarter. We're going to charge you zero and 50. There's actually data to suggest that the larger funds charge a higher management fee of being equal. So as funds get larger, they have more, they become more de-risked. They're later vintages. They become also a different asset class. I would say first three vintages versus different vintages are a different asset class as well, neither here nor there, but they're certainly not priced in the earlier stage. You kind of have like the Lake Wabagon effect where all your funds are above average.

51:28Put it to dollars and cents. Like what kind of alpha are you looking to get from these above average? Maybe, maybe kind of our rank or less might have a 0.4 correlation with future kind of performance. It's not super strong. It's there. And how that manifests itself, if someone says like, hey, I have this signal, it has a 0.4 correlation to future performance. The right way to behave around that is to take a lot of coin flips, right? If you have a somewhat, it's not to say like, oh, 0.4 correlation, give me the top three names and I'm going to invest in all of those. It's to say like, give me the top 70 names and I'm going to invest in all of those because each of those becomes a weighted coin flip.

52:12And if you want to come out better on average, you want to take a bunch of those coin flips, not just flip three coins with a 0.4 coin. How do you think that translates into returns? So let's say you had a fund of fund, you had hundreds of these funds. How does that move the mean? We have told our LPs for the first fund that our target was to get the 75th, which is based on kind of covers, is the 75th percentile sort of blended vintage year IRR compounded for seven years. That's our target return, which I believe was, if you look at the different vintage years we looked at was somewhere between a 3.6x and a 6x net return to LPs.

52:47And so we've invested in the past 15 months. I didn't have any connection with these GPs whatsoever. They're on AngelList. The Python strip said to invest in them. So we invested in them. Outcome of the process is the good thing. So some companies die, some companies do really well, but it's the process that's actually working. We've gotten lucky, but I think we've also put ourselves in a situation to get lucky. And then the thinking on it is this kind of weighted coin flip idea is that actually it happens to be the case. Maybe this is related to the underlying volatility of the asset class, which is why the signal is so noisy, is that investing in a whole bunch of funds and getting exposure to 1 ,000 plus thousands of startups is what I was saying at the start is the right way to invest in seed stage venture.

53:29So we have this really nice alignment where the right way to invest with our signals is to write relatively small tracks and a whole bunch of funds, which will then go and write relatively small checks in a whole bunch of startups, which actually is the correct way, I think, to invest in startups. And so we're fortunate to be in an asset class where the fact that we don't have super strong signals, you know, kind of implies this chain of strategic behavior that is the right way to behave. Now, the reason that could be self-referential is like, you could say like, hey, if the space weren't so insanely volatile, maybe we could build signals that had higher fidelity.

54:09Yeah. I mean, I think that's, that's, it kind of fits together, right? We don't have a super, we don't have a super clean signal, but the behavior that's implied by not having a super clean signal is, is also like, in my opinion, the correct way to, to invest in the asset class. So it's a, it's really, it dovetails really nicely, I think. And that kind of, for me, was a huge motivator to even go and pursue this idea was that like, Hey, you know, the implications lead down a logical path that I had already kind of accepted. But I do think, so for the fund of funds, it's just really, so a chunk of it is the lack of the price.

54:44There's access that we get because we are an Angelist affiliate where every GP on the platform has made the affirmative decision to work with Angelist. So they all have a person that they deal with at Angelist almost every single day. So we have generally been very successful. And this is something I have real eyes on when we're a rest of this fund is like, how many of the funds we want to invest in do we actually get money into? And our rate has been over 90%. You know, we've missed out on a couple where people are just like, you know, when we had our first close, some people were already like fully committed or whatever, and we've gotten to their next funds instead.

55:16But our rate of success placement is very, very high, which is, if you do believe this is a reverse selection thing, like that's the feature you want. You want to maintain that. And then I think that the data side is so interesting as well. So one of the big signals that we use, that's actually the most prominent signal that we use when we make this decision to invest is called markups over baseline. So we look at all the investments that GP has made over the past. It's like smart graduation rate. There's not like a crazy, like, oh, wow, we threw AI, all this advanced machine learning. It's just like, it's straightforward, right?

55:47We look at all the investments someone's made over the past three years. And we know from our data, the baseline rate of the chance of those investments being marked up or marked up means price equity round at least 10 % higher. So if you made a seed investment 18 months ago, So that rate is probably 16%, right? So we go through, we see, okay, you made 20 investments. You've had three markups. The expected number of markups you should have is 5.8. One of the most absurd repeated things in all of venture capital is that you can't know whether somebody is skillful until there's DPI. In other words, taken to the extreme, you made 10 investments, eight of them are on their series F, none of them have done DPI, and this person's up 30x on paper.

56:29or you can't say that that person's a good investor. I think it's absurd. It's just one of these things that sounds smart that people repeat. I would take that a step further, which is I think, I would say that DPI is not even relevant, not because it's not relevant to investors, but it's not relevant as a future-facing signal because it takes so long to assemble enormous DPI. You know, you're talking about someone who is maybe great at investing, or maybe they got lucky 11 years ago. And like, are they retired? Are they still plugged in? Like, are they still like hustling and scrapping for like those same seed stage deals or whatever.

57:00That was, you know, that was more than a decade ago. Their lives are, it's probably completely different lives. Said another way, the DPI as a signal, even if it has some positive correlation, is outpaced negatively by what's happened to that person over the last 10 years, whether they've had style drift, whether their funds have gotten bigger. It's compressed or washed out. But I would actually go so far as to even be critical of that, which is that you might say, if you tell someone that, like, hey, you know, it's 10 years before these things get really real DPI. So like, it's probably not a great future single because like, who knows where they're at now versus 10 years ago.

57:34Someone will be like, well, yeah, but this guy had a super unicorn exit three or four years into investing. So they know how to pick the really, the really good winners, which are the big winners that happen super fast, then they know what to do. And I actually would go so far as to argue that that single is in fact negative, provided with fall caveat. The caveat is you need to be good enough at investing to have a giant early success, right? If you're terrible at investing, and there are some GPs on the Angels platform who we assess at being terrible at investing, that simply would not have the chance of having a giant, you know, humongous exit three years in after they made a seed investment.

58:12Provided that you are of the skill level to put yourself into that situation of having a giant exit, I believe it is actually a negative signal for future performance to have that giant exit versus not have that giant exit. And the reason why is not anything you do, but rather the behavior of other LPs. All it takes is a bunch of several LPs to be like, wow, and you had that giant high profile exit. So you know what you're doing. Here's a bunch of money. Immediate effect of that is for you to suddenly be, to never have to like fight for money ever again, at least for the next decade, right? At least don't people forget about the huge exit, right?

58:49It's upstream of style drift, of going into other domains. So it hurts your alpha as an investor. And then you start investing in, you know, now you have all this money to invest. You invest in whatever you want. You just lose the discipline that may have helped you. And so actually it is, if you can disentangle the like skill, there is a skill part of having a giant exit, right? You have to be good enough to at least put yourself in position. If you can disentangle the luck from the skill, I would every single day of the week invest in the GP who is skillful but didn't get lucky, then the GP who is both skillful and lucky.

59:22And that's not because, you know, all else being equal, their predicted future performance, you know, based on the skill signal might be the same, but all else is not going to be equal. The GP who had that giant early exit is going to get a bunch of money for a bunch of other LPs who think that they've, you know, cracked the code about how to do investing, and their future performance is just not going to look as good as the GP who didn't get lucky. There's a word for these conscious signals, these negative driving forces. It's actually called negative alpha. It's not a term that many people use.

59:52It's literally alpha signals and you have negative alpha signal. Abe, it's been too long. Where should people go if they'd like to follow your research, if they'd like to learn more about what you're working on and just keep up to your work? So the AngelList blog, AngelList.com slash blog. And if anyone wants to, you know, my email is Abe at AngelList.com to talk about research. And yeah, I'm always open to that. I get a lot of good ideas from people just emailing me about, you know, what's some interesting stuff that we could look at with our data set. And yeah, I'm open to hearing back from folks on, you know, kind of what this asset class looks like.

1:00:26Because I think it is, I was reflecting the other day that kind of I've been in the Angelus universe now for six and a half years, which is longer than I've done really anything, longer than I've worked at a job, longer than it took me to get a PhD. And I think there's just some really fundamentally just like intellectually interesting, weird, fascinating stuff about this asset class that like I maybe didn't expect when I started. I think when I started, I was like, I'm going to run this stuff, crunch it, you know, we're going to crush it with a machine learning model and then it'll be solved.

1:00:59And that's the journey that I've been on has been on its own kind of fascinating trajectory. Through the podcast, I get to meet some of the best investors, some of the best asset managers in the world. And one recurring theme that was first brought to my attention by Mike Maples, I think is one of the greatest early stage investors. And he really focuses on find your early believers, find the people that see the world like you see it. And don't try to convert the non-believers. Focus on your believers and focus all your energy on that. and I think certainly it makes sense why there's some people that believe venture, you know, their cousins, uncles, sisters, startup that they met at a bar is going to do as well as a well-veted, diversified portfolio.

1:01:42Those people should go invest in that. Some people should invest in fund of funds and everything in between, but appreciate what you're doing and appreciate you taking the time. Yeah, thank you so much for having me on again. It's been a pleasure and yeah, maybe in another few hundred shows I can come back on for round three. I would love that. That's it for today's episode of How to Invest. If you're a GP with over 1 billion in AUM and thinking about long-term strategic partners to support your growth, we'd love to connect. Please email me at david at weisbergcapital.com.