You're still confusing a lot of things here. The only real issue you have identified with Hive Engine is that no one else runs nodes. This is definitely a major issue, but it isn't a problem with the system itself.

First of all, you claimed that the transactions were not distributed across a network. I explained that they were, and then you countered with the fact that invalid custom_json transactions can be published to the Hive blockchain which has nothing to do with the transactions being distributed across a network, which they clearly are.

To be 100% sure of a transaction, you need to query a Hive Engine RPC node.

This is the case with any blockchain platform I am aware of. To confirm a transaction you need to run the node software and query it local database, or query the node of a trusted 3rd party.

On chain, there is absolutely no difference between the two.

On the Hive blockchain there is no difference between the two. Again, you're confusing things here. On the Hive Engine blockchain there would be a difference which can be independently verified. The fact that Hive Engine uses transactions published on a separate blockchain to populate its own blockchain is definitely confusing and different than most other platforms, but that doesn't mean that it's not a blockchain or doesn't have the same properties as you are claiming.

On the contrary, it's actually a very clever solution to be able to utilize already built features of the Hive blockchain to more easily create a new blockchain with additional features.

Can you tell me which one of these transactions are real, one of them is forged.

First of all, both of the transactions are "real". One of them is just invalid based on the rules of the Hive Engine protocol, and it is very easy to tell which.

Transaction 40ca892010c880dd6bb10777934df6a3b732a116 is valid whereas Transaction 72301f9df4c896da648f2f4729308ef488dc806b is invalid because you didn't have enough tokens in your account balance at the time.

I could "forge" a Hive transaction too, but anyone can query a Hive blockchain node to see that it is invalid.

No one does. 99% query only Engine's node. The only way to protect from this is to run your own node or hope someone else does and identifies it.

Yes, this is the problem. But the key is that you can run your own node and verify everything. The fact that people don't care enough to do that is a different issue. Also, I would venture to guess that somewhere around 99% of Ethereum users and applications only query Infura's node.

the public/private keys is to issue a untrusted custom json transaction, that's it.

That is correct, and again is how all blockchain platforms work that I am aware of. "protected by cryptography" does not mean that the content of the message is valid based on the blockchain's protocol. It means simply that you can verify that the message was sent by someone who has access to the associated private key, which is certainly the case for all Hive Engine transactions.

I can't see any other way for Hive to be relevant enough that people want to use it for large projects.

In my opinion the best path for Hive would be to focus on the niche of tokenized communities. I've looked at and researched a LOT about successful startups, and often the best ones with the best chances of success are ones that focus on one particular thing and work to be the absolute best at it. Even many of the big companies that now do everything started out this way, including Amazon, Google, etc.

As the developer behind Splinterlands, you got to admit using custom jsons for every thing is wonky.

Not at all. The purpose of the custom_json transactions is to provide immutable, cryptographically-signed, ordered transactions, and custom_json transactions on the Hive blockchain provide all of these things exactly as we need them. There's absolutely nothing "wonky" about it.