Conversation

Random thought sparked by some dynamics I’m thinking through for a gig. When your top 2 rivals ally, it’s an sign of your own power, so in a weird way you want to see it happen, and see if you can take on both at once. Test of strength. If you win, your dominance is deepened.
4
61
It’s the flip side of “the enemy of my enemy is my friend” — which is only true if enemy #1 is too big to take on alone. Nobody *wants* alliances with all their inherent compromises and tradeoffs unless forced into them for survival.
1
16
If top 3 players in a contest A, B, and C have strengths of 9, 8, and 7, then B needs a little luck to beat A, and C needs a little luck to beat B, and a lot of luck to beat A. Low incentive to ally. But if the power levels are 9, 5, and 4, B and C *have* to ally to target #1
Replying to
The dynamics are governed by the disproportionate rewards of being #1. If you’re say 51% of the market, you enjoy weird premiums. Selling costs plummet. There’s a term for this I’m forgetting… not monopoly rents, but dominance premiums.
2
12
And it can work down the stack, with #2 being disproportionately more rewarding than #3. But the effects wear off sharply past about 3 or 4. 5+ rankings are commodity. Bronze is the last medal.
3
6
A “fair” unequal market would be a pure commodity market where everyone has similar margins and earnings are roughly proportional to volume. Not counting economies of scale effects, so assuming everyone is in same operating regime and on same cost curves.
1
3
Winner-take-most situations are more common than either winner-take-all or winner-take-fair-share. And value of cooperation falls as you get closer to the winner-take-all end of market. Factors that drive alliance effects: luck levels, rank-based unfairness in rewards, innovation
1
7
So for A, seeing B and C ally is a sign of peak-dominance… and potential impending takedown. If you don’t act fast, complacency could make you vulnerable to the Fellowship of Reluctant Allies.
1
8
Something something Afghanistan, Taliban vs ISIS situation. It’s interesting that they hate each other too much to ally despite hating the US more and being vastly weaker in direct military capability.
1
4
The reason A vs B+C is often an interesting and weirdly stable equilibrium is it’s the edge of divide-and-conquer. A can always stress the division, B+C can always bond tighter briefly. But they are unlikely to take down A since as they get closer to winning, the bond will weaken
1
6
So an A vs B+C dominance hierarchy is a “tell” of a winner-take-most asymmetric rank-based reward contest with low innovation and low levels of luck.
1
5
By luck I mean dumb luck effects not innovation or serendipity. The kind that can alter the outcome of a battle but not the war of iterated outcomes. If the leader A can simply keep the game going and not lose, the luck will average away.
1
5
It’s a generalization of the “guerrillas win when they don’t lose” principle. The military weakness is irrelevant. In terms of the dimension that matters (“depth of territorial roots”), they are the most powerful. The player for whom “don’t lose” equals winning is the top dog.
1
7
It’s even the preferred equilibrium for A. A vs B+C is the most profitable condition for A, and least profitable condition for B and C, so the one that accrues the most rewards for A. Logic: if you can dominate an alliance of your 2 rivals, people will pay more for your power.
1
1
Reverse the logic of Avis #2 ad: “We try harder because we’re number 2” If you’re Hertz, it means you’re phoning it in and still raking in the biggest bucks. Sweet deal.
1
5
Arguably you should prefer this dominance equilibrium condition to even monopoly, so you should NOT knock out B, C even if you can, because being a monopolist means more, and different kinds of threats, inelastic demand, etc. You want to be A in an A vs B+C state.
Quote Tweet
Replying to @vgr
It would be better to give a knock-out blow to both rivals before they realize you're a threat.
2
2
I think if you’re solving for the “maximize easiest money” above a minimum (like say you want the laziest way to make at least $10 million/year), your best bet is to be approximately twice as good as next two rivals.
2
4
So if A = 9, B = 5, C = 4, nominally, A=B+C so it looks evenly matched. Innovative new entrants avoid entering because it looks like an actively contested market with no lazy/corrupt monopoly incumbents. But actually A > B+C because of high alliance coordination costs.
1
2
You have all the real benefits of being a monopolist (rent-like easy money), but not the target on your back from POV of either regulators or disruptive startups. Startups often want to take on monopolies with the next great idea, but apparent oligopolies are less attractive.
1
4
Regulators see real competition and go away. Disruptors see that it’s not a winner-take-all, so they walk away. So you end up with sweet winner-takes-most. Only reason to even try to grow overall market (nonzero sum) or market share (zero sum) is if your minimum need isn’t met.
1
3
The default case is when all 3 are supplying the same thing, but can also be analyzed in scope ways. A sells good big and small widgets, B sells better small widgets, C sells better big widgets.
1
2
Customers who need only one size buy A if price sensitive, B or C otherwise. Customers who want both though face integration costs and will prefer A more, because B and C cost more both nominally and introduce extra hidden costs. Hence one-stop-shop platform economics.
1
3
If the economies of scale additionally have high sensitivity at larger scales, you can be low-cost leader AND have higher margins. This reinforces the A vs B+C equilibrium because remember A only has to not-lose.
1
1
Can’t share the actual examples I’m thinking through, and can’t think of perfectly clean examples, so would welcome examples for my case files. One good approximate example I have no stake in is Apple vs Google Android+Samsung in phones. Maybe Boeing vs Airbus+Bombardier?
1
1
I avoided quoting this because it’s closer to winner-take-all than I’m thinking of 😄 A vs B+C is more like first prize BMW, second prize Toyota Corolla, third prize, used Kia. If you just want a functional car third prize us actually fine.
Quote Tweet
Replying to @vgr
First prize: Cadillac Second prize: steak knives Third prize: you’re fired
3
1
I should name this whole thing. Something like the Wise Incumbency effect. Be decisively powerful but avoid winning 100%. Conversely, advice to startups: attack A vs B+C markets — they are less contested than they seem, and poor at innovation. It’s a lazy dominated oligopoly.
1
6
This pattern is very general btw, not limited to business competition or war. Any kind of resource contention among agents will do. For eg. projects fighting for attention share even within a single person’s head. Case to be made that happiness = A vs B+C contest in your head.
1
2
For eg for years I think I was in a writing vs consulting + community-curating equilibrium. Then I ditched community around 2019 and replaced it with paid newslettering, which is an unstable situation, and now makering is disrupting the attention equilibrium significantly.
1
I bet there’s a clean game theory model here. Iterated game with probabilistic payoffs for A, B, and C based on strengths, burn rate and extinction limits.
1
Kinda hard to communicate the fundamental elegance I see in this pattern because it seems to have a lot of moving parts. But it’s simpler than it looks, just not as simple as spherical-cow textbook asymptotes of perfect competition commodities or perfect monopolies.
1
Player A can be on a low-cost trim trajectory of “steady growth,” pace-setting the race with 2 variables — benchmark cost and tempo of incremental innovation. And do so with minimal maneuvering or strategizing. But B and C have to actively maneuver to follow, at high cost.
1
1
If A decides to lower prices, everybody’s unit economics shifts and there’s a cost pressure others can handle less. If A decides to introduce a feature it’s basically in, with near zero adoption risk, but for B, C, it only works if it’s a radical leap OR if A follows suit.
1
1
Show replies