I have been toying around with a system of investing that I think actually is a “complex adaptive system.” I know this term is super in vogue rn, NGL it’s my clickbait. But luckily I’ve thought about this for awhile and wanted to make a short thread

First: I want everyone to notice that the funds you respect are super concentrated, but many I think organically drift there over time. akrex was not always a 10%+ weight towers, but let those companies compound there by *not selling losers*

Partially they captured the LEFT TAIL

The first core lesson that I’ve learned is that selling winners is THE CARDINAL SIN. I want a system that sells weeds and buys flowers. Averaging UP is hard to do as a human, but not as a system.

Second: I personally do not believe that I can tell you what stocks in a 25 stock portfolio will go up the most or the least if it’s diversified. Unless you’re all long the same thing, I think it’s very hard to know returns prospectively (dumb ikr!)

But there’s a big caveat

I DO THINK (whether thru hubris or otherwise) that I can identify stocks that are decent investments. Tell me to price target I’ll laugh. I would of never guessed TTD at 450, even when I was a hyper bull at 90s-110s. So that’s back to point 1. Never sell winners.

Third: I want something that adapts. We all know the lesson of valeant and endowment bias. I don’t want to be the guy who makes the decisions if I’m being honest, I want to be an analyst and a system churn away at optimal weights w/o human bias. Here’s the plan....

100 starter positions at 1%, maybe I’ll start with some 2% in stocks that I currently like (would of put them at 5% in my previous worldview of position sizing)

I can find 100. It’s really not that hard but honest it’s hard to do deep work, but luckily 100 diversifies.

Next, a simple momentum+++ system will allocate position sizing. I am a huge fan of ranking. I would rank in 3 ways. Valuation, growth, momentum (woo! I am a believer, this is a whole other thread) let’s break down ranking

Valuation: I am a FCF boy at the end of the day. Rank it all 1-100 by a 50/50 blend of fwd P/E & EV/EBITDA, because that’s what street does. You can do this mechanically w/ data plugins

I would then multiply by a FCF / NI conversion (NI-Capex). Rank it all 1-100

Growth: I don’t want to be playing % revenue games, please refer to the dollar revenue piece I wrote here

https://mule.substack.com/p/better-framing-s-curves

I want to reward consistent higher growth in revenue /EBIT, not insane high % on small bases. Some TBDs on how I rank but I got some in mind

Momentum: prob gotta dust off the academic studies, some more TBDs. I believe in momentum. Sorry haters. Also think that momentum is fundamental momentum, and as of late has been picking up lots of alt data

Also why would I punish a stock that goes up? Literally the left tail

But add it all together. I will have 3 ways to rank 100 stocks. avg that. Each Q my job is to prune the portfolio. sell bottom 10-20% mechanically and either raise cash / BUY WINNERS FORCEFULLY / redeploy into new starters

I top 30-50% ranked stocks get added to say 25 bps a Q

I am not comepltely sure about the specifics but the point here is thinking of this lil portfolio like a collection of random genes. I believe that by systematically rewarding good behavior (cheap, grows rev/EBIT, goes up) and by rooting out bad behavior (the inverse)....

I can make a portfolio thats super diverse, yet adaptive. As a stock slows down and gets expensive, it gets force sold into stocks that are doing the opposite. Not all at once but slowly to evolve

I think of it like a cell w/ tons of genes. Each Q is a generation of selection.

But isn’t this just momentum already? yes but with much quicker conclusions. I am mostly unsure how much shifting of weights I should be doing based on ranking, but I think it can and should push optimal solution quickly. Also none of this would work pre 0 trade fees

Anyways I have been a big fan of ranking (did quite a bit in the past!) portfolios. I don’t think I’m smart enough to know my best idea prospectively, how the F am I to know the future! But I do think I can create a diverse portfolio and reward good behavior, root out bad.

Anyways, I am definitely looking for thoughts on this. I’ve thought about this a lot for a long long time. It’s the first portfolio weighting system with a feedback loop I’m aware of!