The growth data of prediction markets looks promising, but peeling back the surface reveals issues: severe liquidity shortages, most trading volume concentrated in a few hot events, and long-tail markets effectively nonexistent. Kalshi and Polymarket are both relying on incentives to stay afloat, which essentially masks the flaws of natural demand with subsidies.

More fundamentally, there is a structural dilemma. Pure binary trading models cannot compete with perpetual contracts because they lack continuous price discovery mechanisms and native hedging tools. Market makers face enormous direct event risk, while retail traders are stuck in low leverage, high-cost traps. This is not an optimization problem; it is a design flaw.

I have observed an interesting phenomenon: traders who truly understand perpetual contracts are often criticized and dismissed as "not understanding the market" when they criticize prediction markets. But quite the opposite—these are the users that platforms should be most eager to attract. They can see through the fog of fake trading volume and identify deadly issues like jump risk.

The trap of local optima lies here. The stories of BlackBerry and Yahoo have already been told, and now prediction markets are replaying the same script. Early success led teams to believe that the current model is the ultimate form, but the vision of everything being predictable has yet to be realized. True breakthroughs require a complete overhaul of the underlying protocols—introducing leverage decay, layered liquidation, cross-market hedging mechanisms—all of which need to be redesigned from scratch.

The problem is, admitting that the current state is broken means giving up the achievements already made. That’s too difficult for any management team.