Centralization tends to lead to this problem—product iteration can't keep up with the growth of user data. Based on a few prediction platforms I've used, the experience indeed varies.

In terms of depth and liquidity, a certain international prediction platform performs well, with many real participants and a proactive team communication. But another prediction platform's attitude is quite cold, and emails mostly go unanswered. Conversely, some platforms on the BSC chain have serious fake order issues, but interestingly, the team is very down-to-earth, and feedback is generally responded to quickly. However, in terms of depth, it is indeed limited; many complex trading ideas simply can't be implemented on these platforms.

Ultimately, it's a contradiction between scale and completeness—where there is liquidity, service awareness may not be in place; where the service is good, the product depth is lacking.