Even if a transaction chain appears to be "linear", all transaction chains are formed via a DAG by bitcoin's nature (assuming the transaction is valid).
Yes, I know, The distinction on the webpage is because "linear" chains are pretty common, at first I thought they were CPFP to the extreme, so I colored the transactions from red to green to distinguish this case. But I found a lot of chains of this kind with transactions having two outputs: one being a OP_RETURN and the other the utxo dependency of another transaction. Maybe you can get some of them just now, they are pretty common.
I would suggest you monitor any such transactions and capture the graph immediately prior to the transaction chain getting confirmed.
Yes, something more to the todo list. But not sure if worthy. Maybe just for fun.
If immediately prior to when the graph of transactions are confirmed, there is a transaction at the "tail" of the graph that has an outsized transaction fee, most likely someone was trying to minimize their tx fees, and was using a strategy similar to CPFP.
Indeed, Txs with red color has the most sat/Vbyte and the green the less.
The first thing that comes to mind to me is that someone might be trying to test the limits as to when certain entities are willing to accept unconfirmed transactions.
Thanks a lot for your guesswork. Not really sure for what could this be useful anyway.