Polymarket's mirror-style order book sharing design is indeed ingenious—achieving a clever liquidity balance through the Yes+No=1 mechanism. However, this approach is more suitable for leading tracks and mainstream prediction categories, as it requires sufficient market maker participation. For long-tail markets, especially social predictions, an independent order book model might be more efficient. Imagine separating Yes and No into two completely independent pools, with the eventual winning participants sharing all the rewards of the opposite pool proportional to their share—could such a design better incentivize initial liquidity for long-tail assets? Both mechanisms have their strengths; the key is to match the market depth and participant structure.