Kinetic Vanguard v8.1.9 — 2024-only from here on, plus the audit results (real fixes vs. false alarms)

Homebrew Fighter subclass (mental-stat psionic striker). Analysis and rules-auditing done with AI assistance. No damage numbers changed in this patch — it's a rules-clarity and current-edition pass. If you were mid-campaign on v8.1.8, nothing about your character's output moves.

Two things happened at once, so this is a slightly bigger patch than usual: I drew a line on edition support, and I ran the ruleset through a second adversarial audit. Here's both, honestly, including the parts where the audit was wrong.

The line in the sand: current 5e only

The subclass now targets current 5e — 2024 rules plus errata — and nothing else. I'd been carrying 2014 fallbacks in several features: "under 2014 the feat is Crossbow Expert instead," "Weapon Mastery is a 2024 feature, ignore this on 2014," and a scatter of "under the 2024/2014 rules" qualifiers.

All of that is gone, and the docs are better for it — not because I dropped anything anyone was using, but because several features can now just state the rule instead of hedging two editions in one sentence. Kinetic Mastery doesn't need a "if your table uses 2024" caveat anymore; it's just how the feature works. The melee-disadvantage note names Sharpshooter and stops there. It reads cleaner because it is cleaner.

If you're still running 2014 and want to use this, most of it ports fine by hand — but I'm no longer maintaining that path, and the text now assumes 2024.

The audit: what was actually broken

I had the full v8.1.8 ruleset audited for RAW failures. Most of the findings were real, and they're fixed:

• Steeled Mind did literally nothing. It granted "+your Proficiency Bonus to concentration saves" — but Fighters are already proficient in Constitution saves, and you can't add PB twice. The feature was void. It now grants Expertise (double PB) on concentration saves, which is what it was always supposed to do.
• The critical-hit rule was too greedy on AoE. v8.1.8 said a crit doubles rider dice — fine for single-target riders, but Fiery Blast, Arc Nova, and Static Discharge are area riders that damage other creatures via saves. A literal reading doubled the whole burst against the entire room because the primary hit critted. Now: a crit doubles rider dice dealt to the creature you actually hit; it does not turn a save-based area attack into a crit for everyone.
• Ember Lance's doubling primer was underspecified. "Double the next strike's damage" — double what, exactly, and what happens on a crit? It now says precisely: double the Manifested Strike's dice and modifier (strike only, not riders), and on a crit you crit first, then double — the dice never roll more than twice, so no accidental ×4.
• The half-Proficiency-Bonus to attack could be read as adding PB twice. It's now written as a flat Psionic Focus bonus (+1 / +2 / +3 by level) — an independent bonus, not a second helping of Proficiency. Identical numbers; it just can't be rules-lawyered anymore.
• Smaller real ones: Cryokinesis's Slow mastery was missing the official "doesn't stack past 10 feet" cap (four strikes could read as −40 ft speed — fixed). Graze technically contradicted the "nothing happens on a miss" rule, since it's defined as an on-miss effect (carve-out added). Overload Mastery had no declaration timing (now: before your first Overloaded roll). Deflection Screen's counter-push assumed an attacker exists, which breaks against traps and environmental damage (now conditional). Surge's Tier 2 was missing "on a failed save." Leaping Strike could catch you in its own burst. And the level-3 feature table was quietly omitting two features it grants.

The audit: what was a false alarm

This part matters as much as the fixes, because a tool being confidently wrong is exactly how bad "corrections" get made:

• The audit flagged Manifested Strike's "counts as a ranged weapon attack but isn't a weapon" as a contradiction. It isn't — that's the exact construction the official Soulknife Rogue uses for Psychic Blades. Coherent, RAW, and deliberately left as-is.
• It flagged Dazzle as "casting spells while claiming not to be a spell." That was reading an old version — v8.1.8 already reframed Dazzle as a non-spell. (I did tighten the last of the "you cast" wording to "you produce the effects of," so it's now airtight, but the contradiction it described was already gone.)
• It wanted Blood Tax to grow a full damage-instance procedure — resistance rounding per instance, temp-HP interaction, ward bypass rules. I did the minimal version (each Overload is a separate instance, round each down) and deliberately skipped the rest. Over-specifying invites more arguments than it settles.

The tell on a couple of these: the audit was working from a copy of the doc that predated v8.1.8, so it "found" things I'd already fixed. Worth knowing if you ever run your own homebrew through an AI auditor — check that it's reading the current version before you act on anything.

The one real design decision it surfaced

2024 Action Surge lets your extra action be anything except a Magic action — and my standalone nova features (Firestorm, Arctic Tempest, etc.) aren't tagged as Magic actions. So you could Action Surge into a second nova, or Attack-plus-nova, which dents the damage parity I've tuned.

The clean fix would've been to tag those features as Magic actions — but that drags in Counterspell, antimagic, and a pile of "is this magic?" edge cases that fight the whole "psionics aren't spells" identity. So instead: you can take only one standalone psionic Action per turn, even with Action Surge (a second Attack action is still fine). Closes the combo without making the class more spell-like. Psionics stay magical-for-Magic-Resistance but explicitly not-spells and not-Magic-actions.

Files (v8.1.9)

1. Main rules — https://docs.google.com/document/d/1oyHShmKKp4fSfbyEbCXnCYk3DcZEFZi3EGr9HCVP3OA/edit
2. Pyrokinesis — https://docs.google.com/document/d/148QWEx5G8vmU9uDN9UUUCEq1VwYFVeOlJqmKoDhNO2g/edit
3. Cryokinesis — https://docs.google.com/document/d/15IUoGxu17pbAx1gDo-zVlMYkgbGZJBRFV7IKpwtJ9XU/edit
4. Psychokinesis — https://docs.google.com/document/d/1XKHTCpVhQiWqMAsLPqes09N2EJF1CYby_QnMkwr-JZk/edit
5. Electrokinesis — https://docs.google.com/document/d/1a96HoNJL5L_ykfbsFWCFA91iQMCiAlx4cRDbSQNr0y8/edit
6. DM Quick Reference — https://docs.google.com/document/d/1b3l7kHDg9XCaoaPiWoOOorc-faZaKZV5YeY3lhl6ook/edit

Still on the list

The audit flagged that my area features ("15-foot radius," "within 15 feet") don't use the 2024 templated area shapes (Sphere / Cylinder / Emanation), and that a couple of zones don't specify once-per-turn re-entry. Both are legitimate for a 2024-only doc, and both are their own project — eight features' worth of geometry decisions. That's a dedicated v8.2 pass, not something to rush into a clarity patch.

As always: if you find an edge case that isn't in the "fixed" list or the "deliberately left alone" list, that's a real gap and I want to hear it. This is the second full audit that's made the docs meaningfully tighter. Keep them coming.

Kinetic Vanguard v9.0 — the honest pass. I nerfed the whole class on purpose, and here's why.

Homebrew Fighter subclass (mental-stat psionic striker/controller). Analysis done with AI assistance; the updated Monte Carlo harness is linked below and reproduces every number here. This patch makes existing characters weaker. That's the point. Read on before you're annoyed with me.

Every balance patch I've shipped until now held one assumption I never questioned: that a Kinetic Vanguard should sit at roughly parity with a pure martial. I tuned the disciplines against each other and against a Champion fighter, and called ~100% "correct."

I now think that was wrong, and I want to be honest about it rather than quietly shave numbers and pretend it's a bug fix.

The design realization

A Kinetic Vanguard is a ranged striker/controller hybrid. It deals damage and it restrains, blinds, slows, shoves, and locks down. A Champion fighter does one of those things — it hits stuff. Hard.

So when my sims showed the Kinetic Vanguard matching or beating a Champion on damage, that wasn't parity. That was the hybrid getting striker damage AND controller utility — the whole kit, no trade. A class that does more should not also hit as hard as the class that only hits. The control is supposed to cost something, and it wasn't costing anything.

Pyrokinesis at 20th was the clearest tell: 164% of a Champion's single-target damage. The damage-specialist martial, out-damaged by 64% — by a subclass that also blinds you and sets zones. That's not a king. That's a problem wearing a crown.

What I did

Three levers, all pulling the same direction: down.

1. The Manifested Strike die now caps at level 11. It stops at 1d8 base / 1d10 Tier 1 / 1d12 Tier 2, instead of climbing to 1d12 / 1d20 / 1d20+1d12 by level 20. Levels 3–10 are completely untouched — the cut lands entirely at the high end, where the class was hottest. Think of it like a Monk's martial die: a hybrid pays for its versatility with a smaller hit.
2. Ember Lance drops from 3× to 2× your Proficiency Bonus.
3. The Tier 2 primer is now a half-strike, not a double. Prime-and-detonate still works — the next strike still hits harder — it's just an extra half a strike instead of a full doubled one. And Voltaic Bolt goes from 2/3 dice to 1/2.

The numbers, single-target vs. a Champion

Pyrokinesis is still the damage king — clearly the top discipline, and it still runs a little above a martial in the mid-game, because that's its identity and I'm not taking that away. But at the level cap it now sits just below a pure striker, which is exactly where a hybrid should be. The other three land in the 55–75% band: controllers first, damage-dealers second. They lock the room; they don't also win the damage race.

Blood Tax is unchanged. The prime-and-detonate pattern is intact. I made the primer smaller, not gone — a class's signature play should survive a nerf, or the nerf is just vandalism.

Why I'm not hiding this

I could have folded these cuts into a "clarity patch" and hoped nobody ran the math. I'm not going to, for two reasons. First, you'd notice — the sim is public, and someone would post the before/after within a day. Second, and more importantly: this is a philosophy change, not a bug. v8.1.5's Ember Lance nerf was fixing a genuine error (an uncapped rider scaling to 206%). This one is me saying "I had the target wrong the whole time." Those deserve different framing, and dressing a deliberate rebalance up as a fix is how you lose people's trust.

If you're running a Pyrokinetic built around that 164% ceiling, this hurts, and I'm sorry — but a subclass that outdamages the damage specialist while also controlling the battlefield was never going to survive contact with a thoughtful DM. Better it comes down now, honestly, than that your table quietly bans it later.

Files (v9.0)

1. Main rules — https://docs.google.com/document/d/1VXx0vxAEjTwdg_9cN6E6wDxjMitvuXJekClHvwo6mrg/edit
2. Pyrokinesis — https://docs.google.com/document/d/1grs-hvA80eu17S1EjpCwr0W02wxs31PzoeUGV6oaT8k/edit
3. Cryokinesis — https://docs.google.com/document/d/18c47H5H0d4wwaZPnbyNMm7p9r2TKpHfxH5nqhgsRfHw/edit
4. Psychokinesis — https://docs.google.com/document/d/15JLtJCU77eHVrVtjluf2IGAeCRkoBLxX2RkxDSCWRnY/edit
5. Electrokinesis — https://docs.google.com/document/d/1O8IECLBjOrR2cNCiLAVcqjYWr6VTs-4CxGa6-PDPUto/edit
6. DM Quick Reference — https://docs.google.com/document/d/1oXjzrHq-9TVK1k7GkE7b-y-92HCBPhBXktAdvfJVtoc/edit

Monte Carlo harness (v9.0): https://drive.google.com/file/d/1DIzeqFEc0fSWrri4_eZpeaREd729III9/view — reproduces the table above. If you think a hybrid should match a pure striker, change the Champion baseline and argue with me. I've been wrong about the target before; that's this whole patch.

What I want to hear

1. Does the class still feel good to play, or did I cut too deep? There's a real difference between "honestly costed" and "why would I play this over a Battle Master." If v9.0 crosses that line, tell me — I'd rather over-correct back up with data than leave it feeling limp.
2. Pyrokinesis players specifically: does 96%-at-20 with the crown still feel like the damage king, or does the die cap make it feel like everyone else with a fire coat of paint?
3. Is the 55–75% band too cold for the control trio? They're supposed to trade damage for lockdown, but if the trade feels bad at the table, that's the signal that matters more than my spreadsheet.

This is the biggest single change I've made to this class, and it's the one I'm least certain about — not the math (the math is solid), but the target. Playtest reports on this one are worth more than usual. Bring them.

Kinetic Vanguard v9.1 — my sim was wrong, the class was secretly a solo god-emperor, and here's the honest correction

Homebrew Fighter subclass (mental-stat psionic striker/controller). Analysis and balance-auditing done with AI assistance across several tools. This is the biggest single change I've made to this class, and it starts with an apology for numbers I published that weren't true. The corrected Monte Carlo harness is linked at the bottom and reproduces everything here.

I have to open with the uncomfortable part, because burying it would be exactly the wrong move.

The sim was undercounting. For versions.

Every "% of a Champion fighter" number I've published was calculated by a Monte Carlo harness I wrote. That harness had a bug: it stepped the Manifested Strike die up when you Overloaded, but it never added the Overload's flat damage bonus (+PB at Tier 1, +2×PB at Tier 2) that the rules require. So every Overloaded strike was scored low — by a lot, on a multi-attack nova turn.

When that got fixed, a second round of auditing found more problems in the harness: it made Flare's and Voltaic Bolt's Overload dice unreachable, it forced the single Tier 2 strike onto the first attack instead of the optimal one, and it handed Psychokinesis a free capstone bonus it never paid for.

Corrected, the real baseline was not the ~96% my v9.0 changelog claimed. It was roughly 170% of a Champion at high level — while also controlling the battlefield. Every discipline was at or above 100%. In plain terms: a Kinetic Vanguard wasn't a party member. It was the main character, and the other four people at the table were watching. Echo Knight on steroids. Solo god-emperor.

That's on me. The fix for a bad number isn't a quiet nudge — it's showing the corrected math and re-tuning against it in the open. So that's what v9.1 is.

What the 9.x arc is actually for

The purpose of this whole arc is simple: make the Kinetic Vanguard a teammate again. A hybrid that does damage and control should sit below a pure striker on damage — it's paying for the control half of its kit. When it matches or beats the damage specialist, the trade is broken and everyone else's turn matters less.

v9.1 is the first real step. It won't be the last.

The rebuild

The Overload economy, crushed. Tier 1 Overload no longer adds a flat damage bonus (it still steps the die). Tier 2 adds 1×PB instead of 2×PB. Your Blood Tax — the self-damage you pay to Overload — is unchanged, so you pay the same and get less. That's the nerf, stated plainly.

Tier 2 Overload, once per action. A Tier 2 is a big, near-guaranteed hit; you shouldn't chain it across every strike. One per Attack action, placed on whichever strike you want. Action Surge still grants a second Attack action and therefore a second Tier 2 — because Action Surge is a core Fighter feature this class respects.

Two damage riders trimmed. Flare's Tier 1 no longer doubles its die (it stays one die and instead makes its Blind unsavable). Voltaic Bolt's Tier 2 no longer adds a third die (it stays two and keeps its Restrain).

Ember Lance became a control rider. This is the biggest identity shift. Its old prime-and-detonate damage spike is gone. Now its damage is flat (your Proficiency Bonus at every tier), and Overload adds control instead of damage: Tier 1 gives the target disadvantage on its next save against a Pyrokinesis feature — priming it for your Flare or Fiery Blast — and Tier 2 adds that plus a Dexterity save or the target loses its reactions. No markers, no tracking, no state on the board.

Where it lands (optimized play, corrected harness)

These are optimized-play numbers — what a skilled player who places Tier 2 well and sequences riders correctly actually gets, not a flattering best-case-for-me figure. Pyrokinesis is now at parity-or-just-under a pure striker: still the clear damage king among the four, still a little hot in the mid-game (that's its hook), but no longer outshining the class whose whole job is damage. The other three land in the 65–80% band: controllers first, strikers second.

Down from ~170%. That's the story.

I'm not going to pretend this is finished

By the "make it a teammate" standard, Pyrokinesis at 96% is at parity, not below it — and I'd argue a hybrid should ultimately sit under a pure striker even at its best discipline. Getting the last stretch requires touching the shared chassis (the strike die, the mastery, or the per-hit modifier) in ways that hurt low-level play or the parts that make the class feel good. I chose not to swing that hammer in one patch. v9.2 will keep tuning toward true party-parity as playtest data comes in. This is a waypoint, an honest one, not a finish line.

Files (v9.1)

1. Main rules — https://docs.google.com/document/d/1yAD55kDiTiYPR79ZLqhAdm08R9BnXeB98xWQK173z0Y/edit
2. Pyrokinesis — https://docs.google.com/document/d/1eYBAflpMmG9rFmwFCUwVrADYyseCdlZkwoPUPlUZ2yg/edit
3. Cryokinesis — https://docs.google.com/document/d/1knO2xYjLjd-XSEOd-tgnnyUzRwUJJN2pFQfpGoeXSto/edit
4. Psychokinesis — https://docs.google.com/document/d/1cJYtW3RgY7wFSyuvN_hWm7BJstVJ7xZVfMHrQLSyfrU/edit
5. Electrokinesis — https://docs.google.com/document/d/1hDESPD6yc8D1vZw5MXOBWatY0cwwxHR44ndoyIGX0uc/edit
6. DM Quick Reference — https://docs.google.com/document/d/1XFlhV828f4WyTIpTNeraxC2xGsF1Ap-KAvqM9gygOHo/edit

Corrected Monte Carlo harness (v9.1): https://drive.google.com/file/d/1XUTdklOEwnl3Oen08kc_PREFnA_Ltg6C/view — it prints three models (published-compat, optimized, and an unconstrained ceiling) so you can see the sensitivity yourself. If you think a hybrid should match a pure striker, change the Champion baseline and argue with me. I've now been wrong about both the target and the math, so I'm not precious about it.

What I want to hear

1. Does it still feel good to play? There's a real line between "honestly costed" and "why would I take this over a Battle Master." If v9.1 crosses it, that's the most important thing you can tell me.
2. Pyrokinesis players: Ember Lance is a control rider now, not a damage spike. Does the prime-a-save-then-Flare rhythm land, or do you miss the detonation?
3. Control trio at 65–80%: too cold, or finally the right trade for the lockdown you bring?

Big thanks to the people running this through their own tools and finding the holes — the sim bug, the unreachable rider branches, the dead Overload tiers all got caught by careful outside eyes. That's the process working, even when what it catches is embarrassing. Keep it coming.

Kinetic Vanguard v9.2 — Pyrokinesis gets a party-support trick (and a dead low-level tier gets fixed)

Homebrew Fighter subclass (mental-stat psionic striker/controller), second step of the 9.x "make it a teammate" arc. Balance-checked with AI assistance; the relevant simulation is described below. Damage numbers are unchanged from v9.1 — this is a control-only update.

After v9.1's heavy correction (the one where I admitted my sim had been undercounting for versions), this one gets to be the fun kind of patch. It started as a bug fix and turned into an identity.

The short version

Ember Lance's Tier 1 Overload now gives its target disadvantage on its next saving throw against anything, from any source — not just your own Pyrokinesis features. It fires once per Attack action, it doesn't stack on the same creature, and because it's per-action, Action Surge lets you light up a second, different target.

That last part is the good bit. A Pyrokinetic on a nova turn can now prime two enemies' saves — which means your wizard's Hold Person lands on one and your teammate's control lands on the other. You're not the star of those plays. You're the setup. And that is exactly what this class is supposed to be growing into.

Why this needed fixing at all

In v9.1, Ember Lance became a control rider: flat damage, with Overload adding a save-disadvantage that primes the target for your Flare or Fiery Blast. Good idea, one problem — it did nothing at levels 3 through 6. Those Pyrokinesis features that force saves? Flare shows up at 7th. Before that, "disadvantage on the target's next save against a Pyrokinesis feature" primed a save that nothing in your kit was going to call for. A 3rd-level rider whose Overload was inert for four levels — while still charging you Blood Tax to use it — is a trap, and traps have no business in a class that already asks its players to manage resources carefully.

Broadening the disadvantage to any save fixes that instantly. At level 3, Overloading Ember Lance now does something real: it sets up whatever your party is about to throw. The dead levels are gone, and the mechanic that fills them happens to be the most teammate-flavored thing Pyrokinesis has ever had.

The guardrails (and the sim behind them)

A free-ish, party-wide "disadvantage on the next save" could get out of hand fast — three Vanguards stacking it on one boss could strip its saves against everyone's control every round. I flagged that risk, then actually checked it: I ran the stacked-save-or-lose scenario against a legendary boss with Legendary Resistance, with and without the debuff.

The ungated version — disadvantage on every attempt — roughly tripled the party's odds of landing a save-or-lose on the boss (about 7% to 36% over a sustained fight). That's exactly the "one Vanguard trivializes legendary creatures" failure the whole 9.x arc has been fighting. So the fix ships with two guards:

• Once per Attack action (not once per strike), so it can't blanket every attack.
• One Ember Lance disadvantage per creature — a second Vanguard's debuff doesn't stack on the same target.

Together those bring the boss case back near baseline. And note the Action Surge interaction is clean because of the no-stack rule: against a solo boss there's only one target, so priming "a second creature" does nothing there — the extra reach only matters in multi-enemy fights, where lighting up two targets is party support, not boss-melting. The simulation case we cared about is unchanged.

Files (v9.2)

1. Main rules — https://docs.google.com/document/d/1bvosm3Bbo9BUGFA1gCzS8dr91HheU3XL1y9wRsBRHmM/edit
2. Pyrokinesis — https://docs.google.com/document/d/1RWWphrnXTk3KSFHKxT2osdNlDmEkuj1d2ok1pAw8QBM/edit
3. Cryokinesis — https://docs.google.com/document/d/1wJanWpmd75siJl9EdsBI01uyjSQYjDsPMhhSnqoESkk/edit
4. Psychokinesis — https://docs.google.com/document/d/1utvJDx9-qU2mygaWhILV0lwq1lnfytcY2JtNED_yrcI/edit
5. Electrokinesis — https://docs.google.com/document/d/1ZT9s2S4mrs4Kyx3Ki5rQeZg9WE6SJ0cd-T3bHcGxH9o/edit
6. DM Quick Reference — https://docs.google.com/document/d/1fAmKEVQArBw6nlDsE9uLsR_hTvkV62tkPMq20_mbSdU/edit

Official harness (v9.2): https://drive.google.com/file/d/13si5SR7h0vM3ezC1iNorv2sOA6qJHT0C/view — balance numbers are identical to v9.1 (Pyrokinesis ~96% of the Champion benchmark at 20th, the other three disciplines 65–80%). This update is control-only and deals no damage, so the model is unchanged; the harness is version-bumped and annotated to say exactly that, but the math and the numbers it prints are the same as v9.1.

What I want to know from the table

1. Pyrokinesis players: does the prime-a-save rhythm feel good in practice? Setting up your own Flare is one thing; setting up an ally's spell is the new thing — does it actually come up, and does it feel like a role or a footnote?
2. Low-level Pyro (3–6): is Ember Lance's Overload worth the Blood Tax now, or is T0 still the only sane option before Flare?
3. Anyone who's stacked control with this: did the once-per-action + no-stack guards feel right, or did they get in the way of something reasonable?

This is a small patch, but it's the first one in the arc that adds something instead of taking it away — and it points where I want the class to keep going: a Vanguard who makes the rest of the party hit harder, not one who makes them redundant. More to come.

Kinetic Vanguard v8.1.9 → v9.2: What Changed, and Why Should You Care?

Player- and DM-facing migration guide

This post compares the final v8.1.9 rules directly against the final v9.2 rules. It summarizes the net result rather than dwelling on temporary mechanics that appeared briefly during v9.0 or v9.1 development.

The Short Version

• Manifested Strike now stops scaling at level 11. Its permanent ceiling is 1d8 normally, 1d10 at Tier 1, and 1d12 at Tier 2.
• Overload deals less bonus damage. Tier 1 no longer adds PB to damage; Tier 2 adds only 1× PB, down from 2× PB.
• Tier 2 Manifested Strike is limited to once per Attack action. Action Surge grants a second Attack action and therefore a second Tier 2 strike.
• Blood Tax is intended to remain unchanged: Tier 1 costs PB, and Tier 2 costs 3× PB, on a hit.
• Ember Lance has been rebuilt. Its high-damage doubling combo is gone. It now applies save disadvantage and, at Tier 2, can deny reactions.
• Flare trades extra damage for certainty. Tier 1 no longer adds a second damage die; instead, its Blind lands without a save.
• Voltaic Bolt loses one damage die at every tier. Its Tier 2 Restrained effect remains.
• Telekinetic Shove gets stronger forced movement. Tier 1 now pushes 15 feet and Tier 2 pushes 20 feet before applying Restrained.
• Cryokinesis and the universal/Advanced Training toolkits are functionally unchanged.

The practical result is a subclass that contributes through damage and control rather than dominating both categories at once.

1. Manifested Strike's High-Level Damage Curve Was Cut

The largest universal change is the new Manifested Strike ceiling.

Why players should care

At levels 3–10, the die itself is mostly familiar, but every Overloaded strike loses flat damage. At levels 15–20, the reduction becomes much larger because the base die no longer grows beyond 1d8.

A high-level Vanguard can still attack reliably, apply riders, exploit Weapon Mastery, and use its control toolkit—but it is no longer built to win the party's damage race by itself.

Why DMs should care

High-tier encounter design should become considerably less volatile. Bosses are less likely to lose enormous chunks of health to routine multiattack Overload chains, and other martial characters have more room to occupy the dedicated-striker role.

2. Overload Is Now a Precision Tool, Not an Every-Hit Damage Multiplier

Tier 1 Manifested Strike

• v8.1.9: Increase the damage die by one step and add PB to damage.
• v9.2: Increase the damage die by one step. No flat damage bonus.

Tier 2 Manifested Strike

• v8.1.9: Increase the die by two steps and add 2× PB to damage.
• v9.2: Increase the die by two steps and add 1× PB to damage.
• New limit: Tier 2 may be applied to only one Manifested Strike per Attack action.

Action Surge still creates another Attack action, so it can produce another Tier 2 strike. This preserves the Fighter chassis while stopping a normal multiattack sequence from applying Tier 2 to every shot.

Why players should care

You pay the same intended Blood Tax for less raw damage. That means Overload selection matters more:

• Put Tier 2 on the attack most likely to hit.
• Pair it with the rider whose control effect matters most.
• Avoid Overloading merely because the option is available.
• Remember that Action Surge creates a second Tier 2 window.

The decision is now closer to, “Which attack needs to be decisive?” than, “How many attacks can I afford to juice?”

Why DMs should care

The main tracking rule is simple: one Tier 2 Manifested Strike per Attack action, not per turn. A normal Attack action gets one; an Action Surge Attack action gets another.

The document also now explicitly states the general rule that Overload tiers are cumulative unless a feature says that a higher-tier effect replaces a lower-tier one.

3. Pyrokinesis Was Rebuilt from Solo Nova into Damage plus Team Setup

Pyrokinesis receives the most consequential discipline-specific changes.

Ember Lance

v8.1.9

• T0: PB fire damage.
• T1: 3× PB fire damage.
• T2: Retained the 3× PB damage and primed a later Manifested Strike to deal double strike damage.

v9.2

• T0: PB fire damage.
• T1: Damage remains PB. The first Tier 1-or-higher Ember Lance each Attack action gives the target disadvantage on its next saving throw against any effect from any source before the end of your next turn.
• T2: Includes the Tier 1 save-disadvantage effect and forces a Dexterity save; on a failure, the target cannot take reactions until the start of its next turn.

The disadvantage has two safeguards:

1. It triggers only the first time each Attack action that Ember Lance is Overloaded.
2. A creature can carry only one Ember Lance disadvantage at a time; repeated applications do not stack on the same target.

Action Surge can therefore prime a second creature, but it cannot pile a second disadvantage onto the same boss.

Why players should care

The old “prime and detonate” damage loop is gone. In exchange, Ember Lance becomes useful party setup from level 3 onward. You can soften a target's next save for your own later feature or for an ally's spell, maneuver, poison, monster ability, or other control effect.

That changes Pyrokinesis from a largely self-contained damage engine into the subclass's strongest team-combo discipline.

It is still the damage-focused discipline, but its best turns now reward communication with the party rather than simply multiplying its own next hit.

Why DMs should care

There is no longer a doubled-strike marker or special critical-hit interaction to adjudicate. Instead, track one straightforward debuff:

• It applies once per Attack action.
• It affects the target's next save against any source.
• It does not stack on one creature.
• Tier 2 separately forces a Dexterity save against losing reactions.

This is easier to run, but coordinated parties can make the save-disadvantage effect valuable. It should be allowed to matter; that party synergy is now part of Pyrokinesis's intended identity.

4. Flare Deals Less Damage but Blinds More Reliably

Tier 1 Flare

• v8.1.9: Increased the bonus damage from one Manifested Strike die to two dice; Blind still allowed a Dexterity save.
• v9.2: Remains at one Manifested Strike die, but the target is Blinded automatically on a hit, with no save.

Tier 2 still upgrades the condition to Incapacitated on a failed save.

Why players should care

This is a clean damage-for-reliability trade. Tier 1 no longer spikes as hard, but its control cannot be lost to a good Dexterity roll or Legendary Resistance.

Why DMs should care

A Tier 1 Flare hit now guarantees Blind through the end of the Vanguard's next turn. Plan for the condition, not merely the possibility of it.

5. Electrokinesis Loses Single-Target Damage, Not Its Control Ladder

Voltaic Bolt

• v8.1.9: 2 dice at T0, 3 dice at T1, and 3 dice plus a Charisma save against Restrained at T2.
• v9.2: 1 die at T0, 2 dice at T1, and 2 dice plus the same Charisma save against Restrained at T2.

Why players should care

Voltaic Bolt remains unavoidable bonus damage on a hit and retains its Tier 2 control, but it is no longer an oversized single-target damage rider. Electrokinesis is pushed more firmly toward its intended identity: arcing damage, disruption, and crowd pressure.

Why DMs should care

The discipline's battlefield role is largely unchanged. The adjustment is mostly numerical, so existing tactics and encounter interactions still work; the Vanguard simply contributes less boss damage while doing them.

6. Psychokinesis Gets Better Forced Movement

Telekinetic Shove's base effect remains a 10-foot horizontal push on a failed Strength save, replacing the normal Push mastery movement for that hit.

Its Overload distances change as follows:

• Tier 1: 10 feet → 15 feet.
• Tier 2: 15 feet → 20 feet, plus Restrained until the end of your next turn on a failed save.

The old Tier 1 wording did not actually increase the distance beyond the 10-foot base effect, so this is both a correction and a modest Psychokinesis buff.

Why players should care

Psychokinesis gains more meaningful positioning power precisely where it spends Blood Tax. Fifteen or twenty feet is enough to push creatures out of auras, off objectives, into hazards, away from allies, or through damaging zones.

Why DMs should care

Map geometry matters more. Review ledges, hazards, chokepoints, and persistent areas when a Psychokinetic Vanguard is present—but remember that the target still receives its Strength save.

7. What Did Not Change?

The update is large, but it is focused. The following core structures remain intact:

• The mental-stat Fighter chassis.
• The Psi Reservoir formula and short-rest recovery.
• Manifested Strike's 60-foot range, attack bonus, feat interactions, and Psionic Focus bonus.
• The basic declaration loop: declare the strike, one rider, and all Overloads before rolling.
• Psionic Instinct's free 3rd-level rider once per Attack action.
• Kinetic Mastery.
• Deflection Screen, Phase Step, Steeled Mind, Psionic Apex, and the Advanced Training pool.
• Cryokinesis's discipline features.
• The intended Blood Tax cost structure.

A character does not need a rebuild, new ability scores, or a replacement subclass sheet from scratch. They do need to update the Manifested Strike progression, Overload rules, and the affected discipline riders.

A Concrete Example of the New Power Level

The documents' level-11 Pyrokinesis example uses the same three attacks, the same 3 Psi, and the same 16 points of Blood Tax in both versions.

• v8.1.9: approximately 59 fire damage, including the doubled-strike primer.
• v9.2: approximately 38 fire damage, plus Blind and a chance to deny reactions.

That is the update in miniature: substantially less self-contained burst, but cleaner and more reliable control.

Player Migration Checklist

• Replace the old Manifested Strike scaling table.
• Remove the Tier 1 flat PB damage bonus from Manifested Strike.
• Reduce the Tier 2 flat bonus to 1× PB.
• Mark Tier 2 Manifested Strike as once per Attack action.
• Pyrokinesis: replace Ember Lance's damage escalation and doubling primer with the new save-disadvantage/no-reactions effects.
• Pyrokinesis: update Tier 1 Flare to automatic Blind with one damage die.
• Electrokinesis: reduce Voltaic Bolt to 1/2/2 dice.
• Psychokinesis: update Telekinetic Shove to 15 feet at T1 and 20 feet at T2.
• Keep Action Surge's separate Attack-action windows in mind.

DM Migration Checklist

• Enforce one Tier 2 Manifested Strike per Attack action.
• Continue charging the intended full Blood Tax on successful Overloads.
• Track Ember Lance's save disadvantage as a non-stacking “next save” debuff.
• Treat Tier 1 Flare's Blind as automatic on a hit.
• Expect lower high-level DPR, but remember that the Vanguard still brings substantial control.
• Revisit any encounters previously inflated specifically to survive v8.1.9 nova damage.

Editorial Heads-Up: Two v9.2 Text Inconsistencies

The final v9.2 document appears to contain two stale or contradictory lines that a table should resolve before play:

1. Tier 2 Blood Tax: The blue Blood Tax summary says the added and total Tier 2 Blood Tax are 1× PB. This conflicts with the attack-declaration table, scaling table, worked examples, and the v9.1 changelog, all of which say Blood Tax was unchanged and use 3× PB total. The intended rule is therefore clearly Tier 1 = PB; Tier 2 = 3× PB.
2. The level-11 Pyrokinesis example: One line still says Ember Lance gives disadvantage only against a Pyrokinesis feature. The actual v9.2 Ember Lance rule says the next saving throw against any effect from any source.

These look like documentation leftovers rather than deliberate alternate rules.

Final Take

v9.2 is a deliberate correction to the subclass's place in a party.

For players: expect less raw damage, especially at high levels, but more meaningful decisions about when to Overload and how to set up allies. Pyrokinesis changes the most; Psychokinesis quietly improves; Electrokinesis loses some single-target punch; Cryokinesis mostly carries on as before.

For DMs: the subclass should be easier to challenge without building every encounter around surviving one character's nova turn. Its control still deserves respect, but its damage is no longer supposed to eclipse a dedicated martial striker while that control is happening.

The intended fantasy survives: you are still a mental-stat Fighter throwing elemental and telekinetic force around the battlefield. You are simply doing it as one powerful member of an adventuring party—not as the entire party compressed into one character sheet.

For everyone side-eyeing the v9.2 nerfs: here's exactly how broken v8.1.9 was

I've gotten a fair amount of "did you really need to hit it this hard?" since the v9.x rebalance. Fair question. Instead of asking you to trust me, here are the numbers — from the same exact-math harness, same Champion benchmark, same three-round frame, run against the raw v8.1.9 rules (the version most people were actually playing) and compared to v9.2.

Fair warning: the v8.1.9 numbers are worse than you think. That's the whole point of this post.

The benchmark, so we're all honest

Everything below is "% of a Champion Fighter's damage" over three rounds, single target, with Action Surge on round one. The Champion here is a deliberately strong yardstick — legacy Great Weapon Master (-5/+10), Great Weapon Fighting, a magic weapon, expanded crit range. That's not a weak comparator I'm dunking on; it's a hard target, chosen so the Vanguard doesn't get to look good against a strawman. 100% means "matches a dedicated damage specialist swinging a magic greatsword with the most broken feat in 5e."

Hold that in mind: 100% is already the ceiling a hybrid controller should almost never reach, because it's also controlling the battlefield on the same turns.

v8.1.9 — what you were actually playing

Same-cadence three-round damage, raw v8.1.9 rules:

That is not a typo. At level 20, Pyrokinesis was doing nearly three times a power-attacking Champion's damage — while also blinding, priming, and scattering the battlefield. Electrokinesis, the crowd discipline, was at 193%. The "weakest" discipline, Cryo, which deals no zone damage at all, still beat the dedicated striker by 63% at 20th.

And that's the conservative read. If you play toward the damage ceiling (which optimizers do), it gets worse:

A fully-optimized level-20 v8.1.9 Vanguard of any discipline was doing roughly 3x to nearly 4x a Champion's damage. Every discipline. That's not a spicy subclass. That's a character who makes the other four people at the table into spectators.

v9.2 — the "nerf"

Same harness, same benchmark, optimized play:

Pyrokinesis went from 290% to 96%. The control trio went from 160-190% down to 65-77%. That's the "nerf." Put the two side by side and it stops looking like a nerf and starts looking like the class finally being a class.

Why it was invisible for so long

Because my simulator was lying to me — and I need to own that. The old Monte Carlo harness had a bug: it stepped the Manifested Strike die up on Overload but never added the flat Overload damage bonus the rules require. So every published "% of Champion" number was low, and I was tuning against a floor that didn't exist. When outside eyes (thank you, everyone who ran the docs through their own tools) caught the bug and we rebuilt the harness with exact math, the real picture showed up: the class had been running at 2-4x a striker for its entire life. The v9.x arc is the correction.

So, to the specific complaints

"Pyro feels so much weaker." It went from 290% to 96% of a power-attacking Champion. It is still the damage king of the four disciplines. It is still at parity with a dedicated striker at level 20, and above it in the mid game. What it lost was the part where it did that damage and controlled the field and doubled everyone else's contribution into irrelevance. You didn't lose your class. You lost the part that was a bug.

"The nerfs were too aggressive." The numbers say they were, if anything, slightly conservative — Pyro at 96% is still arguably a touch hot for a controller-hybrid, which is why the changelogs say the arc continues. There was no version of "gentle" that got a 290% class into teammate range. Gentle would have been dishonest.

"Why should I trust the new numbers?" You shouldn't take them on faith — the harness is public, it prints three models (published-compat, optimized, and an unconstrained ceiling), and the v8.1.9 comparison harness that produced everything above is the same code adapted to the old rules. Run it. Break it. If you find a bug, that's happened three times now and every time it made the class more honest, not less.

The actual point

If v9.2 feels weaker, that feeling is real — but "weaker than a character doing 3x a striker's damage" is the correct direction. The Kinetic Vanguard is now a strong striker/controller hybrid that sits at or just below a pure damage dealer while bringing control the striker can't. That's a good place. It was never supposed to be the entire party compressed into one sheet.

Play a few sessions at the new numbers before you judge. I think you'll find it's still one of the most fun things you can do with a Fighter chassis — it just isn't secretly running the whole table anymore.

Questions, pushback, and bug reports all welcome, same as always.

Harness code for reworked 8.1.9 numbers: https://drive.google.com/file/d/1JWupcvUfytCZp8IZEIIToOOGFmbggDnv/view?usp=sharing

# For everyone side-eyeing the v9.2 nerfs: here's exactly how broken v8.1.9 was

I've gotten a fair amount of "did you really need to hit it this hard?" since the v9.x rebalance. Fair question. Instead of asking you to trust me, here are the numbers — from the same exact-math harness, same Champion benchmark, same three-round frame, run against the **raw v8.1.9 rules** (the version most people were actually playing) and compared to **v9.2**.

Fair warning: the v8.1.9 numbers are worse than you think. That's the whole point of this post.

## The benchmark, so we're all honest

Everything below is "% of a Champion Fighter's damage" over three rounds, single target, with Action Surge on round one. The Champion here is a deliberately strong yardstick — legacy Great Weapon Master (-5/+10), Great Weapon Fighting, a magic weapon, expanded crit range. That's not a weak comparator I'm dunking on; it's a *hard* target, chosen so the Vanguard doesn't get to look good against a strawman. 100% means "matches a dedicated damage specialist swinging a magic greatsword with the most broken feat in 5e."

Hold that in mind: **100% is already the ceiling a hybrid controller should almost never reach**, because it's also controlling the battlefield on the same turns.

## v8.1.9 — what you were actually playing

Same-cadence three-round damage, raw v8.1.9 rules:

| Level | Pyro | Cryo | Psycho | Electro |
|---|---:|---:|---:|---:|
| 5 | 169% | 79% | 79% | 79% |
| 10 | 230% | 99% | 117% | 139% |
| 15 | 267% | 131% | 142% | 167% |
| 20 | **290%** | 163% | 170% | **193%** |

That is not a typo. At level 20, **Pyrokinesis was doing nearly three times a power-attacking Champion's damage** — while also blinding, priming, and scattering the battlefield. *Electrokinesis*, the crowd discipline, was at 193%. The "weakest" discipline, Cryo, which deals no zone damage at all, still beat the dedicated striker by 63% at 20th.

And that's the *conservative* read. If you play toward the damage ceiling (which optimizers do), it gets worse:

| Level-20 model | Pyro | Cryo | Psycho | Electro |
|---|---:|---:|---:|---:|
| Same-cadence (above) | 290% | 163% | 170% | 193% |
| Discipline damage ceiling | 342% | 200% | 207% | 229% |
| Full build, optimal Advanced Training | **384%** | 293% | 293% | 295% |

A fully-optimized level-20 v8.1.9 Vanguard of *any* discipline was doing roughly **3x to nearly 4x** a Champion's damage. Every discipline. That's not a spicy subclass. That's a character who makes the other four people at the table into spectators.

## v9.2 — the "nerf"

Same harness, same benchmark, optimized play:

| Level | Pyro | Cryo | Psycho | Electro |
|---|---:|---:|---:|---:|
| 5 | 101% | 65% | 65% | 65% |
| 10 | 114% | 76% | 85% | 96% |
| 15 | 111% | 72% | 77% | 87% |
| 20 | **96%** | 67% | 70% | 77% |

Pyrokinesis went from **290% to 96%.** The control trio went from 160-190% down to 65-77%. That's the "nerf." Put the two side by side and it stops looking like a nerf and starts looking like **the class finally being a class.**

## Why it was invisible for so long

Because my simulator was lying to me — and I need to own that. The old Monte Carlo harness had a bug: it stepped the Manifested Strike die up on Overload but **never added the flat Overload damage bonus** the rules require. So every published "% of Champion" number was low, and I was tuning against a floor that didn't exist. When outside eyes (thank you, everyone who ran the docs through their own tools) caught the bug and we rebuilt the harness with exact math, the real picture showed up: the class had been running at 2-4x a striker for its entire life. The v9.x arc is the correction.

## So, to the specific complaints

**"Pyro feels so much weaker."** It went from 290% to 96% of a *power-attacking Champion*. It is still the damage king of the four disciplines. It is still at parity with a dedicated striker at level 20, and *above* it in the mid game. What it lost was the part where it did that damage **and** controlled the field **and** doubled everyone else's contribution into irrelevance. You didn't lose your class. You lost the part that was a bug.

**"The nerfs were too aggressive."** The numbers say they were, if anything, slightly conservative — Pyro at 96% is still arguably a touch hot for a controller-hybrid, which is why the changelogs say the arc continues. There was no version of "gentle" that got a 290% class into teammate range. Gentle would have been dishonest.

**"Why should I trust the new numbers?"** You shouldn't take them on faith — the harness is public, it prints three models (published-compat, optimized, and an unconstrained ceiling), and the v8.1.9 comparison harness that produced everything above is the same code adapted to the old rules. Run it. Break it. If you find a bug, that's happened three times now and every time it made the class *more* honest, not less.

## The actual point

If v9.2 feels weaker, that feeling is real — but "weaker than a character doing 3x a striker's damage" is the correct direction. The Kinetic Vanguard is now a strong striker/controller hybrid that sits at or just below a pure damage dealer while bringing control the striker can't. That's a *good* place. It was never supposed to be the entire party compressed into one sheet.

Play a few sessions at the new numbers before you judge. I think you'll find it's still one of the most fun things you can do with a Fighter chassis — it just isn't secretly running the whole table anymore.

Questions, pushback, and bug reports all welcome, same as always.

v8.1.9 vs v9.2, played sanely: the full data, graphs, and CSV

My last reality-check post used a damage ceiling — a player Overloading recklessly with no regard for the self-damage. Fair criticism came back: nobody actually plays that way, so those 3x–4x numbers overstate real tables. Good point. So I redid the whole comparison with a sanity constraint, and I'm posting the data, the graphs, the raw CSV, and the code so you can check all of it.

The new rule: a 25% HP Blood Tax gate

Both versions are now modeled the same way — the player commits at most 25% of their max HP to fueling Overload through Blood Tax. No suicidal nova turns. This is "a careful player pushing hard but not dying," applied identically to v8.1.9 and v9.2 so it's a true apples-to-apples comparison. The self-damage budget scales with level (11 HP at level 5, up to 41 HP at level 20).

This is a much more honest model than the ceiling, and it matters: it caps exactly the reckless Overloading that produced the eye-watering old numbers. If v8.1.9 was only broken when played recklessly, this gate would show it. It doesn't. It shows v8.1.9 was broken even when played carefully.

v8.1.9 — careful play, % of a power-attacking Champion

Reminder on the benchmark: 100% means matching a dedicated Champion striker swinging a magic greatsword with legacy Great Weapon Master (-5/+10). That's a hard target — it's the damage a character gets when damage is their entire job. A controller-hybrid reaching 100% is already suspect.

v8.1.9, playing carefully, sat at 116%–170%. Every discipline beat the pure striker. Pyrokinesis by 60–70%. And it did that while controlling the battlefield. That's the definition of a class that makes the other seats at the table optional.

v9.2 — same gate, same benchmark

Pyrokinesis lands at parity — 98% at 20th, a little hot in the mid game (that's its identity). The control trio sits at 65–80%: below the striker, which is the point, because they bring lockdown the striker can't.

The reduction, level by level

How much each discipline came down from v8.1.9 to v9.2, under the same gate:

The cut lands hardest exactly where the problem was worst — high levels — and stays gentle at low levels. At level 20 the whole class came down roughly 40%. In raw numbers at 20th, careful-play nova damage against a 126-damage Champion baseline: Pyro went 201→123, Electro 171→97, Psycho 146→86, Cryo 146→85.

The graphs (one per discipline)

I plotted v8.1.9 vs v9.2 against the Champion line for each discipline across levels. Linking rather than embedding — the forum and inline images don't get along.

• Pyrokinesis: https://drive.google.com/file/d/1Rlf0pnu7Yx2R1UVSoKvhdo_K7rUvSuyB/view — the biggest drop; watch it fall from ~160–170% to the 98–120% band, hugging the Champion line instead of towering over it.
• Cryokinesis: https://drive.google.com/file/d/1yWQDYeC0MpKYM7G48XEou_Xjj4EvBNLm/view — the pure-control discipline; drops from ~105–116% to the mid-70s, cleanly under the striker where a no-damage-zone discipline belongs.
• Psychokinesis: https://drive.google.com/file/d/1H8Z5hSzNnJpzsyUw1zPHmPZMGvNSg3M4/view — tracks Cryo closely; settles into the low 70s.
• Electrokinesis: https://drive.google.com/file/d/15snPxolassAcq1F-Xb1l0BQRVd4S2afx/view — the crowd discipline; comes down from ~130–136% to the high 70s–low 80s, still the heaviest hitter of the control trio.

The raw data and the code

• Full CSV (every level, every discipline, HP budgets, expected Blood Tax, both versions): https://drive.google.com/file/d/1MdmEp6T3JFhTnWfdRRvI7jTF87nuESPZ/view
• Python that generated all of it (comparison, gate, graphs): https://drive.google.com/file/d/1azm7g3MyWz4eC-Z-3_GIWujy4JvXZ2V4/view

Run it, change the gate, swap the Champion baseline, argue with my assumptions. The whole point of posting the code is that you don't have to take my word for any of this.

What this changes about the earlier "3x–4x" post

Those numbers were a true ceiling — reckless Overloading, no self-damage limit. They're real, but they describe a player torching their own HP bar, not a normal table. These gated numbers are the ones to trust for how the class actually plays. They're less dramatic — v8.1.9 at 116–170% instead of 290–384% — and that's exactly why they're more convincing. Even played sanely, the old version beat a dedicated striker at every level and every discipline. The recklessness wasn't the bug. The recklessness just made an already-broken baseline look cartoonish.

The takeaway

Played carefully, v8.1.9 was a 116–170% class pretending to be a 100% class, while also running the battlefield. v9.2 is a 65–120% class that sits at or below a striker and brings control instead of eclipsing everyone. If the new version feels weaker, that feeling is honest — but "weaker than a careful-play 160%" is the correct direction, and the data says the correction was, if anything, slightly restrained.

Play a few sessions at the new numbers before judging. Bug reports and pushback welcome as always — the code's right there, and every time someone's found a hole in it, the class has come out more honest on the other side.

The other half of the rebalance: v8.1.9's damage was literally suicidal, and here's the Blood Tax math

Everyone's been focused on the damage nerfs. But there's a second story in the data that actually makes the class better to play, not just more balanced — and it's about survivability. This one's about Blood Tax: the self-damage you pay to Overload.

Same exact-math harness, same 25% HP gate, but now I'm reporting the self-damage side of the ledger — both what a careful player pays and what the theoretical damage ceiling would cost you in your own hit points. The numbers explain why capping Tier 2 Overload isn't just a damage nerf; it's a survivability buff.

Two ways to read Blood Tax

• Expected Blood Tax accounts for misses — you only pay it on a hit, so the expected cost is lower than the worst case.
• Maximum Blood Tax is the all-hit branch: every Overloaded strike connects, and you pay every point.

The 25% gate is branch-safe: no possible sequence of hits can push your self-damage past 25% of an average CON-14 Fighter's HP. That's the "careful player" line. Above it is what I'm calling the insane ceiling — maximize damage, ignore whether you stay conscious.

Level 20: the headline

Average Fighter HP at 20th is 164. Here's what the full damage ceiling costs you in self-damage, both versions:

Read the v8.1.9 Pyro row again. The full nova ceiling cost 204 points of self-damage against a 164 HP pool. You would be dead — at zero, unconscious — before you finished the combo. Even the expected cost, accounting for misses, is 113% of your max HP. That "342% of a Champion" ceiling from my earlier post? This is the fine print: it was only reachable by a character actively killing themselves to reach it.

Every v8.1.9 discipline's ceiling was in the same territory — 79% to 124% of your entire health bar, spent on your own turn, to hit those numbers.

Why the throttle is a survivability buff, not just a nerf

Here's the cleanest way to see what capping Tier 2 Overload actually did. Ask: when you push past the safe gate toward the damage ceiling, how much HP do you pay per extra point of damage?

Under v8.1.9, going for the ceiling cost you less than 1 HP per point of extra damage. That's an incentive to nova yourself to death — the exchange rate was so good that the optimal play was to burn your whole health bar. The system was rewarding suicide.

Under v9.2, the same push costs 4 to 6 HP per point of extra damage. That's a terrible exchange — which is exactly the point. The health throttle means there's no longer any reason to torch your HP bar, because the marginal damage isn't worth it. At level 20, v9.2 Pyro gains only 5.1 expected damage by abandoning the safe gate entirely, while nearly doubling its self-damage. Nobody's going to make that trade. The class now self-regulates: careful play is optimal play.

Insane-ceiling self-damage across all levels

For completeness — what the damage ceiling costs in expected self-damage at every tier:

The safe all-hit gate at levels 5/10/15/20 lands at 9, 20, 30, and 36 HP respectively — because Blood Tax comes in chunks of PB and 3×PB, the largest legal gated sequence at 20th actually tops out around 36 HP, about 22% of max. v9.2's ceilings sit close to or within survivable range at every level. v8.1.9's spiral out past your total HP by the mid-teens.

The data and the code

• Full Blood Tax CSV (every level/discipline/version, gated and insane, expected and max): https://drive.google.com/file/d/19wE_4vA2v8f3zKqg9kDBmrZ7WHwJYo9X/view
• Readable audit report (the plain-text table): https://drive.google.com/file/d/1_NlO6fkuO2KRwV64s94PlxFEpEnHjnnd/view
• The exact simulator that produced it: https://drive.google.com/file/d/1ldLRgDQD-iaNZFHfcKXLB0i8Gi9-u2z7/view

As always: run it, change the CON assumption, change the gate, and check my work.

One forward-looking note on the harness: starting with the next revisions, the 25% HP Blood Tax ceiling is getting baked into the official simulator as a default constraint, rather than being an optional gate I apply after the fact. The "insane" ceiling numbers are useful for showing how broken the old exchange rate was, but they describe a character who doesn't survive their own turn — so they're not a realistic balance target. Going forward, the harness treats "how much damage can you do without committing more than a quarter of your health to Blood Tax" as the honest question, because that's the one that reflects how the class is actually played. The uncapped ceiling stays available as a diagnostic, but the survival-gated number is the one the tuning targets.

The takeaway

The damage rebalance gets all the attention, but the Blood Tax story is the one that makes the class feel better to play. v8.1.9 didn't just let you out-damage a striker — it did so by handing you an exchange rate that made self-destruction the correct move. The optimal v8.1.9 turn was one that could kill you.

v9.2's Tier 2 throttle changes that math completely. You still Overload, you still pay Blood Tax, you still make the hard "is this hit worth the HP" call — but the system no longer rewards you for emptying your health bar, because the marginal payoff isn't there. That's not a nerf to how the class survives. It's a fix to it. The Kinetic Vanguard is meant to be a voluntary self-destructive striker who plays on a knife's edge — not one whose best turn ends with them face-down on the floor.

Same offer as every post: the code's public, the assumptions are all visible, and if you find a hole, the class comes out more honest for it.