I'm not talking about the standard Murder. I've been talking about the variant Murder with 2 lances front @ 60, 2 lances side at 45 and 4 WBs side at 45. This is the ship I have been referring to all along. I don't take the standard Murder because I like to stay at range and the WBs aren't as reliable at range. This is why your arguments havent made any sense. Youre arguing against a ship that I'm not talking about.
I don't have my gunnery table handy, but at 45 abeam a Carnage gets 3 dice doesn't it? Thats one hit on 5+ armor average for 16WBs. I think the Murder's 4WBs turn into zero and it only gets 2 lances, which is also one hit. So for 10pts cheaper you get equal damage. On LO its still pretty even. Against 6+ armor, the Murder's lances come in handy too. Firing from 45cm at closing 6+ armor the Carnage gets 8 I think while the Murder gets 2WBs and 2 lances. Again, about the same damage both regular and on LO for 10pts less. Getting closer doesn't put the Carnage way above the Murder for damage output either. They are pretty close in terms of damage.
The Carnage can hit abeam at 60 while the Murder can't but with only 2 dice. The Murder can hit fore at 60 with considerable firepower, but the Carnage can't. The Murder's off side is just as strong as its on side. The Carnage's off side is 6WBs weaker than its on. The Murders allow for lots of opportunistic fire, the Carnage does not. But in the same situation they will perform about the same, although the Murder is cheaper. I like the lance variant Murder than the Carnage because it is much more flexible than the Carnage.
That being said, Carnages are useful and I advise taking one for 4+ armor and defences.
If Murders would take a beating being in front of prow armor, Carnages would take the same beating.
Don't just say fail to the mathhammer. We aren't dealing with fractions of hits here. Since each die has no memory or knowledge of what the other die roll, each die has a 1/3 or 1/6 chance of hitting. Its easily possible that no die will score hits. You are more likely to fail with each individual die with WBs than with lances, and youre more likely to success with each lancebdie than WBs. You will tend to get more fails than above average rolls with WBs than Lances, and will get a more equal number of fails and above average rolls for lances. This is because we only get a success or failure so only certain points on the curve are relevant and complete failure is an option. The probability doesn't work the way you think here.