Did you give it the appropriate personality? It's pretty easy to fix it so it doesn't struggle against the falcon by taking off emags and trading a railgun for a LAG so you can afford resistant flux conduits
Actually I did one better - I issued the Full Assault command at the start of combat just to make sure that a Shield Shunt build would have the maximum aggression possible.
The actual reason your build struggled against Falcons is speed: during every trial involving the SIM Falcon, whenever the Falcon was pressured, it was fast enough to fall back behind another enemy to cause the autopilot AI to switch to the closer enemy, and thus the SIM Falcon was never successfully destroyed in any of the 10 trials. This is a significant issue with using SO on larger-than-frigate ships: the AI is very easily outsmarted and outmaneuvered in a fleet context against multiple enemies when it has neither range nor overwhelming speed.
For example, I don't agree that my SO build is specialized for weak ships, even though it does struggle against faster ships.
Both the ACG and Thumper are punch-down weapons due to their low hit strength. To illustrate, I ran your Eradicator (P) build 1vs1 against SIM Conquest, with 0 of 5 successes (average 71.6% hull damage to Conquest, highest 92.2% so it's quite possible to win with random luck). Even the Conquest's 1200 base armor with no skills was enough to slow down ACG/Thumper damage and prevent the Conquest from being overwhelmed. The Thumper fared especially poorly: in every instance it dealt well less than half the hull damage of the ACG despite the Thumper's huge burst frag DPS, due to the Conquest's residual armor.
Now the ACG's raw DPS is so high that it's feasible to simply spam enough ACGs using DP-efficient ships to overwhelm heavily armored targets - the Brawler (LP) is particularly good at this strategy. However, the Eradicator (especially the P variant) isn't really a good ship for this, and on its own the ACG isn't very efficient against heavy armor compared to how easily it annihilates lighter targets.
In contrast, the AI autopilot had no problems soloing the SIM Conquest using the HVD Breach build, because the Breach's irreducible scripted armor damage easily strips capital-grade armor, and HVD has high enough hit strength to punch through capital-grade residual armor and effectively chew through hull. This is an example of a build that isn't just punching down on weak targets.
It's pretty easy to fix it so it doesn't struggle against the falcon by taking off emags and trading a railgun for a LAG so you can afford resistant flux conduits, I just didn't bother because the example we were working with didn't have much EMP damage.
A bit off topic but shield shunt also works if you have an Officer with Elite Polarized Armor & Resistant Flux Conduits Hullmod. High tech EMP weapons wont be giving you any trouble with that combo, but its really only something I would consider on a much more maneuverable/high armor ship as you are still susceptible to a lot of other damage sources.
Unfortunately, Polarized Armor and RFC aren't panaceas for Shield Shunt against EMP damage. Polarized Armor is a permanent -25% EMP with Shield Shunt, so stacked (multiplicatively) with RFC that's -62.5%. However, if you're factoring in skills, Elite Target Analysis is +100% on its own. (We'll ignore other bonuses like CR bonus, sizes bonuses from Wolfpack/Target Analysis, and even systems like HEF.) A small weapon mount has 250 hp while a single shot from Ion Cannon does 400 base damage (and Ion Beam has 400 base EMP DPS as well), so 400*2*0.375 = 300 damage received, resulting in small weapons still being disabled by a single shot. This doesn't even factor in: multiple weapons, bursts from larger weapons like Ion Pulser or Tachyon Lance, EMP arcs, or simply being fired on repeatedly for longer than 1s.
There are a ton of other factors that go into this, such as hit location, weapon size, EMP arc targeting, and so on. However, it's pretty clear that while EMP resistance has a large effect against stray EMP damage, it's not going to hold up at all against sustained EMP attacks without using shields to relieve the pressure.
Even worse than EMP is large anti-armor damage from weapons like HIL, Tach Lance, Plasma Cannon, Hellbore, etc., against which Shield Shunt has no recourse. Armor tanking is very effective early on (arguably
too effective, especially against Pirates and Hegemony) so it's quite reasonable that Shield Shunt falls off very hard late game. Effective armor tanking against tougher fleets requires using shields to block EMP/anti-armor to preserve armor for soaking up weak hits.
On the other hand, converted hangar is a bit of a sly trick in the sim because even if it doesn't contribute much damage it diverts the attention of the enemy so your 1 ship doesn't get overwhelmed, while in a fleet there's already other ships pulling enemy attention.
That being said, I'm super curious what a peak performance eradicator looks like to you.
Were I to use Eradicator in one of my fleets, the specific build would depend heavily on composition of the entire fleet. However, I can't imagine a fleet where I'd use an SO Shield Shunt Eradicator, whether base or (P). It's simply not the right ship for that strategy: not tanky enough to survive, nor fast enough to outmaneuver enemies.
Calling CH decoys a "sly trick in the sim" implies that it doesn't work in live combat, but in fact I use CH extensively in my actual fleets - it's not a simulator-only gimmick. All builds have weaknesses, not just SO and Shield Shunt builds; for long-range kinetic builds, their main vulnerability is being swarmed and not being able to apply shield pressure fast enough to push enemies back and maintain range. Using CH as a defensive decoy is actually a perfectly logical strategy to compensate for this weakness, even during full fleet combat. (You'll notice that the Warthog's limited engagement range serves as a feature to tether itself to its home ship as a decoy, instead of wandering away to chase some random frigate it can't catch.)
In fact, one tendency I see in many players is to not compensate at all for build weaknesses, but instead merely excuse them as "this ship is designed to be a flanker" or "this ship is only designed to fight X types of enemies but avoid Y enemies" - and then blame the AI when the build fails because the AI isn't programmed to execute those parameters exactly. I believe that the AI performs best with builds that are more situation-agnostic and with some safeguards against the most common failure scenarios - and for long-range kinetic builds, CH is one example of such an AI-friendly safeguard.