Since I’m in Claude, I pointed Fable at the problem and here’s the output
The core diagnosis: the game’s specialization incentives operate at the wrong scope
The existing bonus stack is entirely planet-scoped. A CoGC advertising campaign grants 25% to one industry on that planet, experts are earned and activated per base (max 5 per industry, 6 active total), giving up to 28.4% for a full set, and the only company-scoped bonus is the HQ faction bonus, roughly 10% to industries determined by where your HQ sits, modified by the ratio of used to total base permits. CraftsmanThirteen’s numbers confirm how this stacks in practice: CoGC at 25%, a full expert set at 28.4%, and faction at 4–6% add up to a 60%+ production bonus that barely changes across profession.
Here’s the perverse consequence: because the big bonuses are planetary, they successfully forced bases to specialize — and in doing so made empires modular. A vertically integrated 30-base empire isn’t fighting the specialization incentives at all; it’s a portfolio of individually specialized bases, each sitting on its matching CoGC world with full experts. The only differential between “one company running the whole chain” and “five companies trading with each other” is the single-digit faction bonus. lowstrife’s framing is exactly right — once your company grows enough to vertically integrate everything on its own CoGC worlds, the negative incentives vanish and only the ~10% empire bonus remains — but I’d sharpen it: the fix has to operate at company scope, which is why Counterpoint’s “Experts 2.0” instinct — experts attached to companies rather than bases — is aimed at the right structural layer.
The second, deeper driver: integration is insurance, not yield
The thread’s most economically important post is CraftsmanThirteen’s #12, and it’s not about bonuses at all. He wants to be vertically integrated because the market can’t guarantee reasonably-priced inputs — 49 ZR on the book when he consumes more weekly, LFL bid below its ZR replacement cost — and because his supply chain collapsed when the game’s largest carbon producer quit, leaving him with money problems for weeks. Profit is secondary; there’s too much uncertainty in the market to rely on it.
That means the marginal veteran isn’t weighing “10% bonus vs. 0%.” They’re weighing “10% bonus vs. an insurance premium.” If a supply disruption idles a high-tier base for two weeks a couple of times a year — plus permanent counterparty-attrition risk in a game with a small population — the implied insurance value of integration is plausibly 10–20% of annual profit, with fat tails. No plausible production bonus outbids that, and one large enough to try creates worse problems (below). This is a two-equilibrium coordination problem, and lowstrife nailed the loop: the markets are not robust because nobody specializes; if specialization were encouraged, trade would flourish — as it visibly did for the new electronics intermediates after the MM changes. The economy is sitting in the autarky equilibrium, and it’s locally rational for every individual to stay there. Bonuses change relative payoffs; they don’t, by themselves, provide the shove between equilibria. A liquidity backstop does. So any real solution is a package: reward channel + risk-reduction channel.
The margin-vs-volume insight is the correct sorting criterion for proposals
AGM-114’s contribution is the sharpest design constraint in the thread, and it holds up quantitatively. Production modifiers don’t scale consumable and building-fab usage, so they improve margins — a strongly specialized player can drive a price below the generalist’s breakeven and hold it there indefinitely, which means the modifier must stay conservative, and conservative modifiers don’t do much (see the existing HQ bonus). Formally: if a good’s daily cost is inputs I·r + fixed time-costs C (consumables, degradation), a speed bonus b moves your breakeven unit cost from I + C/r to I + C/(r(1+b)) — the effect is small for input-heavy goods but large for time-cost-heavy ones (SD, AML, resource extraction), which is exactly the class SLKLS flagged as breaking under speed buffs too. Whereas an area/permit advantage scales throughput and time-costs together: unit economics unchanged, profit per permit up. Specialization becomes a reward for growth rather than a requirement for profitability. This also protects the autarky playstyle, which matters for retention — Filefolders and CraftsmanThirteen both said they’d disengage if end-to-end shipbuilding stopped being viable, and Laaxus’s warning that players quit when they lose freedom is the right constraint to honor. Carrot-only design: integration stays exactly as good as today in absolute terms, specialization gets better in relative terms.
Against that criterion, my ranking of the in-thread mechanisms:
Best bonus mechanism: SLKLS’s distributed pool, with the payout partially converted from efficiency to capacity. A fixed company-wide boost budget (say 75%) distributed according to your expert/activity distribution, capped per profession (say 25%), so full strength means three professions — and a new player is automatically at max boost with no margin disadvantage. This is elegant for four reasons: it’s automatic (no point-assignment UI, no gaming a declaration), it satisfies Weiiswurst’s hard constraint that the boost must not scale with HQ level or you make the new-player experience worse, it repurposes an existing system, and — if the distribution is computed over a trailing window of production activity — re-speccing is organically slow, which threads Laaxus’s needle between commitment and freedom. Its one weakness is that as written it’s a pure margin-channel bonus. I’d split the payout: a modest efficiency component (≤10% per specialized industry, visible and motivating) plus a capacity component (buildings of your specialized industries consume ~15–20% less area, or equivalently bases gain bonus area restricted to matching buildings and habitation). The capacity piece is where the real power goes, because it’s competition-safe and because area/permits are the actual binding constraint in the endgame — note lowstrife’s own instinct that realization should be slow since you can’t magically reformat your bases every time you unlock something; area bonuses have exactly that property.
One measurement detail: weight the specialization index by area or workforce cost, not building count — SLKLS correctly noted that building-count portions bias toward bases with 30 small pioneer buildings over bases with 10 big high-tier ones, which would perversely punish exactly the players you want moving up the tech tree.
Worth adopting alongside: CraftsmanThirteen’s permit-point idea, in reduced form. Full decoupling of HQ level from permits is a big rework with the CoGC-flipping incentive problem AGM-114 identified, but the narrow version — permits cost fewer points when the base matches your specialization and/or sits on the frontier — is another volume-channel reward and attacks planet-crowding simultaneously.
What to avoid: recipe input/output modification (startube’s original form) — lowstrife’s rebuttal is the right one: items are balanced across wildly different margin structures, from thin-margin high-throughput electronics to fat-margin extraction, and yield changes break that balance; production speed is the designed lever. Also avoid the two dangerous perks in the OP’s list: the Marketeer MM-price bonus (an inflationary faucet with obvious wash-trade abuse surface, in a game whose price bands have already been manipulated historically) and the plot-limit bypass (deletes planetary scarcity, one of the few genuine sources of geographic trade). And avoid tying the core bonus to HQ level or ARC level as the strength axis — Tonatsi’s corporation idea is fine as flavor, but jcheung’s HQ-level scaling runs straight into Weiiswurst’s objection. The rest of lowstrife’s perk list (specialized cargo capacity, reduced consumable variety, governor upgrades) is good Layer-3 material: identity, resource sinks, volume-flavored, safe.
The recommendation, as a package
Layer 1 — Company charters (the reward): activity-derived specialization index per industry, trailing 8–12 week window, area-weighted; benefits capped at ~1/3 share so three industries is full strength. Payout: small efficiency (cap ~8–10%) + significant matching-industry area discount (~15–20%) + specialization-gated permit discounts. New players are at full charter strength from day one by construction.
Layer 2 — Market de-risking (the shove): this is the part I’d argue is necessary, not optional. Expand MM coverage onto the chokepoint intermediates (the ZR/MHL/FLX class — single-recipe, ship-critical, currently one-player-deep), with deliberately wide bands that algorithmically tighten or withdraw as organic book depth develops — CraftsmanThirteen already proposed the bootstrap-then-scale-back pattern, and lowstrife’s post-MM-change electronics volume chart is direct evidence the mechanism works. Add recurring supply contracts (auto-renewing, deposit-backed penalty terms) so the “corp-mate who sells me cheap carbon” relationship survives as an institution rather than a friendship — this directly attacks the attrition risk that drove CraftsmanThirteen out of corporations. The linked commodity-futures discussion is the ambitious version of the same idea. Separately, liquidity fragmentation across four CXs with this population size makes every book four times thinner; anything that funnels flow (new-player routing toward Moria/Antares, cheaper cross-CX arbitrage) compounds Layer 2.
Layer 3 — perk trees as long-term sinks, cherry-picked from the OP, added only after 1+2 stabilize.
If forced to pick a single minimal change instead: decouple the HQ faction bonus from unused permits (everyone in the thread agrees it’s never worth leaving permits unused for the bonus, so that modifier is dead weight), make the industries player-allocatable per CoinCrafter’s idea but flat rather than HQ-scaled, and add the chokepoint MMs. That’s maybe 60% of the value at 20% of the dev cost.
