Let's stick to the original point of contention. This is what you said:
I pointed out how that doesn't apply to missiles with optical guidance. Whether and/or when optical or radio is better is besides the point. It's clear to everyone that the NSM/JSM/LRASM/Penguin/SLAM-ER(-ATA) will not aim for the highest RCS spot, nor the ECM. I think it should also be clear that they are capable and dangerous missiles.
NSM and LRASM can be programmed to impact pretty much any point on a ship, the funnel included. I wouldn't be surprised if SLAM-ER has a similar capability.
What does "almost all" and "modern and capable" mean? How many exceptions am I allowed to find for that to still be true?
Here's a video of a NSM (a modern and capable missile?) test, where the missile impacts the superstructure:
You got an old ship where the superstructure is full of squared corners that will generate the highest RCS return. It can also target the superstructure because the sunlight has left the metal structure hot, making it look bright to a thermal camera.
NSM and LRASM are only programmed to hit a ship, not parts of. They use IIR to recognize the outline of a ship, not necessarily parts of a ship. If you want to program to hit parts of a ship, your missile database would exponentially explode in the sheer number of entries, and the time to query all these items would also take much much longer. Even if you did, the funnel is a bad choice for a target. An armor piercing missile can potentially just pass through the funnel when you want to sink the ship, not slow it down. Opening up the funnel isn't going to stop the ship.
There is a difference between IR guidance and IIR. IR guidance is heat seeking. These are missiles that will go after the hottest spot of the ship. IIR will identify the outline of a target by infrared camera, and match that to a database. The problem of both is the same countermeasures you use with aircraft. Flares can blind the camera and smoke will distort the ship outline.
Its not as if the 055 doesn't have layers of defense. Passives have long been used by navies for OTH targeting, only this time, with LRASM you are trying to use passives as a midcourse correction, which is also what some Russian missiles like Granit has used before. Traditionally, EMCON makes fleets difficult to target via OTH passive, and to add to that, if you have AESAs and solid state radars, you have LPI along with frequency agiles combined with pseudorandoms. To pull these signals apart from noise, you would need a server farm. Against thermal, you have management of your thermal contrast, hence the use of cooled funnels. Thermals are also more weather sensitive, water vapor does a good job of absorbing infrared, and you can get into a weather situation where both optical and thermal ranges are drastically cut. These two literally depend on their fair weather friendship, just as bad as microwave atmospheric ducting that OTH passives depend on.
The use of passives, IR and optics only serve to prevent the missile from being detected by ESM, and to be countered by ECM. It doesn't stop them from being targeted by an active radar, and have the missile blown away by gun or missile CIWS.
If your radar has detected and tracked the missile, but your ESM failed to receive any radar waveform from it, you can conclude that the antiship missile is a passive using thermal or optical in terminal, and your defenses can act on that assumption too.