but i say this because nothing stops pools from doing this. by this you will also find that no pool is going to be forced, made to, or coerced into including a LN close channel session onchain.
Would you believe that that would be the main reason why Antpool and Bitcoin.com are excluding Segwit transactions?
What about doing it as a means to collect higher fee rewards per block?
each pools has their own intentions. EG btcc used to let in tx with zero fee but only when the tx originating from their exchange
another pool guaranteed first-in (priority) if they done a API pushtx through a website portal*
another pool guaranteed first-in (priority) if they done a API pushtx through a website portal* when paying the pool indirectly
*then treating the normal network relay/mempool as second class
another pool wouldnt include transactions where the utxo is under 6 confirms(honestly, most pools should do this, as it would help avoid spam and also reduce orphan risk)
but back to the question you asked
well if you had 2 tx's both say 300bytes but one earning you 25cents and the other treated as $1 which one would you grab
if core removed the wishy washy witness scale factor *4, guess what
1. both legacy and segwit transactions can in full non stripped format all happily utilise the 4mb 'weight'
2. both legacy and segwit transactions would both be 25 cents. yep cores code makes legacy 4x more expensive. not discount segwit by 75%(thus code counters their PR adverts)
core can easily also limit the sigops/tx. that way it allows more tx per block during spam attacks. .. did you know the way core put in sigop limits that one person can make just 5 transactions and use up the block sigop limit because the tx sigop limit is so high (facepalm)
..
also by reducing the tx sigop limit, thus allowing more tx per block sigop limit also as a side effect reduces the chance of the whole linear validation delay issue.. but if core were to actually do some efficiency stuff and actually fix things, they would have no PR to say other non-blockchain networks are needed.
did you know that core could code many things. including a fee priority formula that charged people that re-spend funds more than once a day with a higher fee, thus a tx that has more aged utxos getting a lower fee.
imagine it a utxo spamming(re-spending) every block paying 144x more than someone spending once a day. which can be done with just a few lines of code.
yep. bitcoin is code and MANY things can be done. but core pretend only 2 things can be done... hense their false narative of
"gigabytes by midnight or LN"