Constructing a contract protocol

 It is a directed graph, with a root.

The root identifies the block chain public key that decodes the contract on some time lock and count value. So, at the proper time out, anyone can go the the block chain and pick out current encrypted time, and unlock the contract. That is the root encrypting key

The, say, any person can supply a public key and secretly encrypt the contract taking the part of the defined party in the contract. Say, time depositor. It i defined in contract by public key. He encryption steps the contract down one step, and after that she will have exits that her public key will decode when valid. She may have time locked Btc funds, and now a third party can use those fund, in exchange for Swift cash by public key with a verifiable Swift account. 

The contract eventually exists, the Swift ledger exchange partially reversed. Public pits can be specified apriori by contract, so the contract specifies market conditions are execution time.  The signature tree need not contain all signatures on every outcome.  If my condition is met, my counter party can use my public key and advance the contract own that path.  But we can see pars of provability, the signatures must be specified apriori, the signatory returns is encrypted portion of the contract, and a public key. That constitute legal agreement.

What is different?

The public Spectre system requires the trader  or miner to push the contract, it does not have the momentum.  A bot in an unknowable processor would just act on behalf of all parties. In a network, all the bots would act in accordance without congesting. 

In the public contract ledger I can always thin my contract by having all parties agree to some common exchange, especially  local short chain. Then, in many cases the conract just shifts liability to Coinbase or some other well known exchange.  The large exchange will have their own well know Taproot trading API available for standard. The public contract chain gets the equivalent of a subroutine call, provable at the interface. So everything seems to be equivalent between cache Spectre and ledger Spectre.  There is a large instruction count difference as the cache Specter need not refer to public keys constantly, nor pause for more than one miner approval. And remember, a cache Spectre can maintain accounts in its own name for neutrality carry over.