So far as I can inform, P2TR is healthier than P2WSH in virtually each case. The main assumption within the design of P2TR is after all, that you’ll favor to make use of the keypath every time attainable, and that many extra advanced spending situations might be shortcut by getting everybody to agree on signing off on the end result straight in type of a keypath spend when the scriptpath spend would be capable of implement the identical consequence.
Particularly within the context of multisig and script timber with a number of leaves, the P2TR development is superior, however there may be one besides wherein it’s best to all the time favor P2WSH. Everytime you solely have a single script within the tree, anticipate that you’ll by no means use the keypath, and can’t profit from public key aggregation, it’s best to use a P2WSH output. The P2WSH output effectively commits to a single script, whereas a P2TR scriptpath requires the controlblock to disclose the leafscript within the tree that provides 34 weight models to the enter’s witness stack. One other benefit is after all that some senders will be unable to create P2TR outputs at first whereas sending help for P2WSH is already broadly applied.
Privateness
P2TR outputs will stand out at first since there might be few of them. In the long run, the anonymity set of P2TR is more likely to be bigger since each single-sig makes use of and extra advanced situations look the identical when they’re spent by way of the keypath.
Engineering effort to undertake
Sending to P2TR addresses would require wallets to implement help for Bech32m. Wallets that need to obtain and spend P2TR outputs might want to implement the Schnorr signing algorithm, and reengineer any OP_CHECKMULTISIG primarily based schemas with the op-codes accessible in Tapscript.
H/T to Vojtěch Strnad for correcting my unique reply by mentioning that P2WSH is extra environment friendly in a single-script use-case.