A large reputable member threshold multisig operating as functionaries for a Bitcoin-pegged deterministic replicated state machine sidechain with as-compatible-as-possible-with-mainchain semantics is probably more reliable and secure that most alternative chains.

• Large: The number of functionaries should be large enough to ensure geographic, jurisdictional, and administrative distribution.

• Reputable members: Functionaries should be chosen who would suffer a reputational loss in the case of poor performance.

• Threshold multisig: A M-of-N multisig. M should be at least ⌊N/2+1⌋, to reduce the chance of equivocation.

• Deterministic, functionaries: Discretion is unpredictable and morally hazardous. The semantics enforced by the functionaries should be deterministic and predictable, not discretionary. Semantics should never change, and if they must, changes should be announced long enough in advance to make exit practical.

• Bitcoin-pegged: If the currency of the sidechain isn't bitcoin, users of the sidechain cannot meaningfully exit. Ability to exit incentives the functionaries to be good stewards of the sidechain.

• Replicated state machine: The state of all functionaries should be the same, and it should be able to recreate and run ones own copy of the state machine.

• Sidechain: Functionaries should publish a sequence of block headers where each block header includes the hash of the current state, as well as the hash of the previous state. Previous states should also be made available by functionaries, in order to make the system auditable.

• As-compatible-as-possible-with-mainchain semantics: The ability of users to exit should be maximized, and making the semantics of the sidechain as close as possible to those of the mainchain maximizes the ability to exit, by allowing users to destroy atomically destroy assets on the sidechain in exchange for mainchain assets which mimic the properties of the atomically destroyed assets.

• Probably more reliable and secure than most alternative chains: Alternative chains suffer from many issues. Proof-of-work suffer from a large global pool of potential hashrate that can attack. Proof-of-stake chains suffer from byzantine consensus mechanisms of ever increasing complexity which must operate in a fully adversarial environment, and have economics which allow and incentivize centralization. A multisig chain operates with a far simpler and more understandable security model: the functionaries periodically agree on a new set of transactions, run the transactions on the old state to produce the new state, and sign and publish the result. Such a system should be more reliable, predictable, and secure.

  And not only that, if the functionaries are chosen carefully, such that there is a huge number of them, perhaps greater than 50, and they are all reputable entities, they should be both incentivized to run the chain properly, and with limited latitude for malicious behavior.