Can someone tell me how these errors occur?
It's simply because of the way the block explorers are interpreting the transaction data... or improperly interpreting the data, as the case may be.
Block Explorers 1&2 are only looking for the
address (ie. the hash of the public key)... so they only see 2 transactions for that
address, 1 in, 1 out and attribute the address to have a balance of zero.
Block Explorers 3&4 are looking for any transaction that has the
public key... unfortunately, the public key (0243c17584f2a00d7c22429e1444e2a94cae0a657b9a58ee6a87c33ef7bca97245) was used in a "spending script that required multiple sigs spend" (it's a precursor to the more common "multisig" that we have today)... so these explorers are seeing all these "extra" transactions where the public key was used as part of this type of multisig... but the address does not have exclusive control over those coins, so they shouldn't be showing as being part of the balance of that address.
Lastly, Block Explorers 5&6 are seeing the transactions for
both the address and the public key... but
only where the coins were sent to just that address or that public key (ie. not the "multisig" transactions.), so they arguably give the "most correct" answer.
In short, some old transactions that are now effectively "non-standard"
today, confuse block explorers made to parse what are considered to be "standard" transactions. (NOTE: This is a huge oversimplification.)
