5 Oct. Two game-wide dials built as exam post steps (tools/bonusshare.js), each 0 = today to the bit (checked on all 16 armies). Discovery only: nothing adopted, no ledger file or value changed. Measured on the ledger snapshot dcd477d8a78c (the tip after T-715's leader fixes, 34b6285814), Track A (step 34b, wCarry 0).
In plain English
The idea. An army's rule, the detachments its players run and its units' auras hand out bonuses. Some units can receive most of them (they carry the right keywords and their attacks qualify); others can't. The Thousand Sons research said reviewers rate the Psychic-keyword infantry up and the generic daemon engines down for exactly that reason. And the characters who make the hard rituals succeed (Magnus, Ahriman) are valued for that, not for filling a caster slot.
What we built.
- Bonus share (
bonusShare). For each army, a package: the army rule's effects, the effects of the detachments its 11th-edition tournament lists run (grimstat-corpus, CC BY 4.0, weighted by their share of lists) and its units' auras. Each effect is priced by the engine on each unit's own guns against the five standard targets. A unit's share is the part of the package's dice it can receive. Value′ = Value × (1 + w × (share − the army's median share)). - Bonus casters (
bonusCaster). Where the army rule is a dice-tested resource (Thousand Sons' rituals, World Eaters' blessings), a unit is credited with the extra damage, in points a game, that its bonus adds over free casters. Credit = w × that × the army's median Value ÷ the unit's points.
What it says.
- Bonus share helps Thousand Sons but not the game. On Thousand Sons it adds +1.3, +2.9 and +8.4 at 0.25, 0.5 and 1 (38.2 → 46.6). Its placebo, the same shares shuffled among the army's units, gives −0.3, −0.3 and +1.3.
- Game-wide it loses: pooled exam −0.1, −0.7 and −1.2 (placebo −0.2, −0.8 and −2.0).
- Held out by army: −1.22 ± 1.22. Two splits of three pick 0; the third picks 1 and loses 3.7.
- It costs most in Emperor's Children (−9.9 at 1), Agents (−7.6), Death Guard and World Eaters (−6.5, −6.2). It helps Drukhari (+6.3), Blood Angels (+2.4) and Adeptus Mechanicus (+1.6).
- Thousand Sons gains for the wrong half of the reason. The dial pushes down the units that receive nothing:
- the Maulerfiend falls from 1st to 7th, the Helbrute from 9th to 18th and the Spawn from 13th to 26th (C, D and D);
- the Forgefiend rises to 1st (B), since the army's own rituals, Destiny's Ruin and Twist of Fate, help any shooting unit;
- Rubric Marines move only a little (7th → 5th, S);
- Scarab Occult Terminators fall (0.29 share, A).
Most Rubric and Scarab damage comes from inferno boltguns, combi-bolters and khopeshes, which aren't Psychic in the datasheets. Their Psychic weapons are the Aspiring Sorcerer's few shots. So the Psychic-gated detachment bonuses reach a small part of their damage. The keyword story holds for who gets nothing, not for who gets the most.
- Sekhetar Robots don't score high for the reason the research gave. Their detachment makes their attacks Psychic. But no list in the corpus ran Sekhetar Cohort beside a detachment with Psychic bonuses (the two Sekhetar lists paired it with Warpforged Cabal), so the grant meets nothing. They rise (6th → 2nd at 1) on the rituals alone, like any gun platform.
- Bonus casters barely move anything.
- A plain caster who may Channel the Warp already makes Destiny's Ruin 98%, Doombolt 90% and Twist of Fate 72% of the time. Magnus's +2 lifts that to 100%, 98% and 90%, and saves the mortal wounds of channelling. That is about 4 points of damage a game. Ahriman's +1 earns nothing: his own mortal wounds cost more than a Rubric squad's, which outweighs the +1.
- The Exalted Sorcerer's once-a-battle second Doombolt (Grand Coven's Incandaeum, half the lists) is worth 11 to 12.
- Thousand Sons −0.3 at 0.5 and 1; pooled ±0.0; held out 0 (every split picks 0).
- The research's "Magnus and Ahriman make the hard rituals succeed" isn't what the dice say once channelling is free to try. The reviewers' high Magnus and Ahriman must come from something else: the rituals' range, Magnus's other abilities, or a reading of the rituals without channelling. Without channelling, +2 takes Twist of Fate from 28% to 58%: that is the story the research told.
- Together, on the stack (armyEnabler 1, sureFire 0.5, exclusiveCarry 0.5, objPresence 1 with extras and enabler):
- On Thousand Sons, the stack alone gives 49.5. That is the research's "existing dials take Thousand Sons 38.2 → 49.5", reproduced on the new snapshot.
- Both dials on top give 52.1, 52.8 and 54.0 at 0.25, 0.5 and 1. The placebo gives 48.7, 48.1 and 46.7.
- Pooled, the stack is 62.2 (+2.9 over Track A). Both dials on top take it down to 61.9, 61.2 and 60.5, level with their placebo (61.7, 61.3, 60.2). Held out on the stack, every split picks 0.
Verdict. Neither dial holds out game-wide. Bonus share is a Thousand Sons-only gain (+8.4 at 1, placebo +1.3), and it comes through the units that receive nothing rather than the Psychic infantry. It is a candidate only as an army-specific cascade axis at 0, 0.5 and 1, beside T-711's split and stickiness. Bonus casters, as built, carry too little to matter.
Thousand Sons, before and after
The package. Detachments are read from 9 of its 11 11th-edition lists (top-quarter lists counted twice):
- Grand Coven 50%, Warpforged Cabal 25%;
- Sekhetar Cohort, Changehost of Deceit, Ritual of Regeneration, Rubricae Phalanx and Hexwarp Thrallband 8% each.
Effect weights (w) and values (v, the median gain per unit; "1" is the army's median, used for a rule with no dice):
| From | Effect | w | v |
|---|---|---|---|
| Cabal: Destiny's Ruin (odds of a plain caster) | re-roll hit 1s, every Thousand Sons unit's ranged attacks into the target | 0.98 | 0.17 |
| Cabal: Twist of Fate | +1 AP, the same | 0.74 | 0.22 |
| Grand Coven: Psychic Maelstrom | Psychic weapons +1 to wound | 0.17 | 0.29 |
| Grand Coven: Wrath of the Immaterium | Psychic weapons Devastating Wounds | 0.17 | 0.30 |
| Warpforged Cabal (no dice) | vehicles (our reading) | 0.25 | 1 (0.23) |
| Sekhetar Cohort | Sekhetar Robots +1 to hit in melee | 0.08 | 0.25 |
| Changehost of Deceit | psykers 4+ invulnerable against shooting | 0.08 | 0.20 |
| Ritual of Regeneration (no dice) | psykers, not monsters: healing | 0.08 | 1 (0.23) |
| Rubricae Phalanx | Rubricae +1 save against Damage 1 | 0.08 | 0.23 |
| Hexwarp Thrallband (two options) | Psychic re-roll wound 1s / +1 to wound | 0.04 each | 0.17 / 0.29 |
Shares (median 0.42), highest first:
- 0.62 to 0.40: Vindicator 0.62, Predator Destructor 0.62, Forgefiend 0.62, Predator Annihilator 0.61, Land Raider 0.58, Sekhetar Robots 0.57, Tzaangor Shaman 0.56, Rhino 0.52, Lord of Change 0.47, Rubric Marines 0.47, Ahriman 0.46, Sorcerer 0.45, Infernal Master 0.45, Pink and Blue Horrors 0.44, Tzaangor Enlightened with greatbows 0.44, Sorcerer in Terminator Armour 0.43, Kairos 0.42, Defiler 0.41, Magnus 0.40.
- 0.34 to 0.27: Daemon Princes 0.34, Exalted Sorcerer 0.34, Scarab Occult 0.29, Mutalith 0.29, Exalted Sorcerer on Disc 0.27, Heldrake 0.27.
- 0.20 to 0: Flamers 0.20, Tzaangors 0.13, Helbrute 0.12, Maulerfiend 0.10; Screamers, Tzaangor Enlightened and both Spawn 0.
The rituals: our odds and prices.
- Each success is priced on the damage 500 points of the army's own guns put into the five standard targets, in the targets' points.
- Destiny's Ruin is worth 20 points (40 on 10+), Twist of Fate 23 (42 on 12+) and Doombolt 21 (53 on 11+).
- A plain caster's success with channelling, at +0, +1 and +2:
- Destiny's Ruin: 98%, 99.5% and 100%;
- Doombolt: 90%, 94% and 98%;
- Twist of Fate: 72%, 82% and 90%.
- Channelling costs D3 mortal wounds on a double, priced at the caster's points a wound: a Rubric squad's 13 points a mortal-wound pair, Magnus's 57.
Ranks among the army's 35 ranked units (exam: 38.2 Track A):
| Unit | Reviewer | Share | Caster credit (pts a game) | Rank: Track A → share 0.5 → share 1 → caster 1 | Stack → stack + both 1 | Value: Track A → share 1 → caster 1 |
|---|---|---|---|---|---|---|
| Rubric Marines | S | 0.47 | – | 7 → 5 → 5 → 7 | 1 → 2 | 1055 → 1112 → 1055 |
| Scarab Occult Terminators | A | 0.29 | – | 27 → 27 → 27 → 27 | 26 → 23 | 503 → 441 → 503 |
| Sekhetar Robots | A | 0.57 | – | 6 → 3 → 2 → 6 | 6 → 3 | 1130 → 1306 → 1130 |
| Tzaangors | A | 0.13 | – | 14 → 20 → 23 → 14 | 21 → 26 | 784 → 558 → 784 |
| Mutalith Vortex Beast | S | 0.29 | – | 25 → 25 → 25 → 26 | 27 → 27 | 574 → 498 → 574 |
| Defiler | S | 0.41 | – | 4 → 4 → 4 → 4 | 4 → 5 | 1150 → 1145 → 1150 |
| Maulerfiend | C | 0.10 | – | 1 → 6 → 7 → 1 | 8 → 16 | 1284 → 876 → 1284 |
| Forgefiend | B | 0.62 | – | 3 → 1 → 1 → 3 | 3 → 1 | 1218 → 1458 → 1218 |
| Helbrute | D | 0.12 | – | 9 → 12 → 18 → 9 | 19 → 22 | 945 → 663 → 945 |
| Chaos Spawn | D | 0.00 | – | 13 → 22 → 26 → 13 | 24 → 29 | 802 → 466 → 802 |
| Magnus the Red | S | 0.40 | 4 | 23 → 24 → 21 → 24 | 22 → 18 | 580 → 571 → 586 |
| Ahriman | A | 0.46 | – | 22 → 19 → 16 → 22 | 10 → 9 | 657 → 687 → 657 |
| Exalted Sorcerer | S | 0.34 | 11 | 26 → 26 → 24 → 23 | 25 → 20 | 552 → 510 → 637 |
| Exalted Sorcerer on Disc of Tzeentch | C | 0.27 | 12 | 29 → 29 → 29 → 28 | 30 → 28 | 310 → 265 → 408 |
| Sorcerer | S | 0.45 | – | 33 → 32 → 32 → 33 | 33 → 33 | 203 → 209 → 203 |
| Infernal Master | S | 0.45 | – | 28 → 28 → 28 → 29 | 29 → 30 | 336 → 345 → 336 |
| Daemon Prince of Tzeentch | B | 0.34 | – | 10 → 8 → 9 → 10 | 9 → 13 | 930 → 860 → 930 |
Ahriman has no credit: at +1, channelling on a 100-point character costs more than the +1 gains over a free caster.
Thousand Sons' exam score at each setting (the placebo is the mean of 3 seeds):
| Setting | 0.25 | 0.5 | 1 |
|---|---|---|---|
| bonusShare (placebo) | 39.5 (37.9) | 41.1 (37.9) | 46.6 (39.5) |
| bonusCaster (placebo) | 38.2 (38.2) | 37.9 (38.2) | 37.9 (38.4) |
| the stack (49.5) + both dials (placebo) | 52.1 (48.7) | 52.8 (48.1) | 54.0 (46.7) |
World Eaters' casters:
- Angron's re-roll lifts the chance of both dice blessings in a round from 0.58 to 0.867: 76 points a game.
- The Lord of Skulls' ninth die lifts it to 0.772: 52 points a game.
- World Eaters stay at 67.2 at every caster strength.
The tables
Held out by army (the strength chosen on 10 armies, 0 included, scored on the other 5; 3 seeded splits; mean ± SE; the stack's row is measured against the stack):
| Dial | Score | Real (picks) | ± SE | Placebo (picks) | ± SE |
|---|---|---|---|---|---|
| bonusShare | exam (mean over 11e lists) | −1.22 (0, 1, 0) | 1.22 | +0.00 (0, 0, 0) | 0.00 |
| bonusShare | consensus pairs | −1.22 (0, 1, 0) | 1.22 | +0.00 (0, 0, 0) | 0.00 |
| bonusCaster | exam | +0.00 (0, 0, 0) | 0.00 | +0.00 (0, 0, 0) | 0.00 |
| bonusCaster | consensus pairs | +0.00 (0, 0, 0) | 0.00 | +0.00 (0, 0, 0) | 0.00 |
| both dials on the stack | exam | +0.00 (0, 0, 0) | 0.00 | +0.00 (0, 0, 0) | 0.00 |
| both dials on the stack | consensus pairs | +0.00 (0, 0, 0) | 0.00 | +0.00 (0, 0, 0) | 0.00 |
In sample, each strength. Pooled over the 15 armies' 11th-edition lists (consensus pairs where an army has 2 or more lists). Thousand Sons and the Tyranids' three 11e headline lists are shown beside, never pooled.
| Setting | Exam, 15 | Δ | Placebo Δ | Consensus, 15 | Δ | Placebo Δ | Armies better | Thousand Sons | Δ | Placebo Δ | Tyranids, 3 lists | Δ |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Track A | 59.3 | 60.0 | 38.2 | 81.6 | ||||||||
| bonusShare 0.25 | 59.2 | −0.1 | −0.2 | 59.9 | −0.1 | −0.2 | 7/15 | 39.5 | +1.3 | −0.3 | 81.6 | −0.0 |
| bonusCaster 0.25 | 59.3 | +0.0 | −0.0 | 60.0 | +0.0 | −0.1 | 0/15 | 38.2 | +0.0 | −0.0 | 81.6 | +0.0 |
| bonusShare 0.5 | 58.6 | −0.7 | −0.8 | 59.2 | −0.8 | −0.8 | 7/15 | 41.1 | +2.9 | −0.3 | 81.3 | −0.3 |
| bonusCaster 0.5 | 59.3 | −0.0 | −0.0 | 60.0 | −0.0 | −0.1 | 0/15 | 37.9 | −0.3 | −0.0 | 81.6 | +0.0 |
| bonusShare 1 | 58.2 | −1.2 | −2.0 | 58.8 | −1.2 | −2.1 | 6/15 | 46.6 | +8.4 | +1.3 | 80.8 | −0.8 |
| bonusCaster 1 | 59.3 | −0.0 | −0.1 | 60.0 | −0.0 | −0.1 | 0/15 | 37.9 | −0.3 | +0.2 | 81.6 | +0.0 |
| the stack (armyEnabler 1 + the Votann three) | 62.2 | +2.9 | 63.0 | +3.0 | 11/15 | 49.5 | +11.3 | 81.3 | −0.3 | |||
| the stack + both dials at 0.25 | 61.9 | +2.6 | +2.4 | 62.7 | +2.7 | +2.5 | 9/15 | 52.1 | +13.9 | +10.5 | 81.2 | −0.4 |
| the stack + both dials at 0.5 | 61.2 | +1.9 | +2.0 | 62.0 | +2.0 | +2.1 | 9/15 | 52.8 | +14.6 | +9.9 | 80.8 | −0.8 |
| the stack + both dials at 1 | 60.5 | +1.2 | +0.9 | 61.2 | +1.2 | +0.9 | 9/15 | 54.0 | +15.9 | +8.5 | 80.5 | −1.1 |
Per army (exam score on its 11e lists; the stack, and both dials at 1 on top of it; the share's spread):
| Army (lists) | Track A | bonusShare 0.5 | Δ | bonusShare 1 | Δ | bonusCaster 1 | Δ | Stack | +both 1 | Δ vs stack | share median (quartiles) | lists read |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Thousand Sons (1) | 38.2 | 41.1 | +2.9 | 46.6 | +8.4 | 37.9 | −0.3 | 49.5 | 54.0 | +4.5 | 0.42 (0.27–0.47) | 9 |
| Space Marines (2) | 60.1 | 60.1 | +0.0 | 60.0 | −0.0 | 60.1 | +0.0 | 59.9 | 60.2 | +0.4 | 0.87 (0.86–0.88) | 41 |
| Adeptus Mechanicus (1) | 71.4 | 72.7 | +1.2 | 73.0 | +1.6 | 71.4 | +0.0 | 70.2 | 72.0 | +1.9 | 0.33 (0.33–0.42) | 20 |
| Agents of the Imperium (1) | 56.1 | 54.7 | −1.3 | 48.4 | −7.6 | 56.1 | +0.0 | 48.9 | 41.7 | −7.2 | 0.83 (0.83–1.00) | 1 |
| Astra Militarum (1) | 67.6 | 67.7 | +0.2 | 67.2 | −0.4 | 67.6 | +0.0 | 67.3 | 67.2 | −0.2 | 0.37 (0.29–0.40) | 25 |
| Blood Angels (1) | 61.3 | 61.3 | +0.0 | 63.7 | +2.4 | 61.3 | +0.0 | 67.7 | 70.2 | +2.4 | 0.79 (0.76–0.86) | 12 |
| Dark Angels (1) | 53.7 | 53.9 | +0.2 | 54.4 | +0.6 | 53.7 | +0.0 | 56.3 | 56.4 | +0.1 | 0.84 (0.73–0.84) | 28 |
| Death Guard (1) | 55.0 | 49.8 | −5.3 | 48.5 | −6.5 | 55.0 | +0.0 | 56.7 | 50.0 | −6.7 | 0.72 (0.72–0.74) | 24 |
| Drukhari (1) | 58.0 | 60.9 | +2.9 | 64.3 | +6.3 | 58.0 | +0.0 | 59.9 | 64.7 | +4.8 | 0.45 (0.00–0.65) | 12 |
| Emperor's Children (1) | 69.6 | 63.4 | −6.3 | 59.7 | −9.9 | 69.6 | +0.0 | 73.8 | 61.3 | −12.6 | 0.81 (0.68–0.94) | 21 |
| Genestealer Cults (2) | 56.4 | 57.2 | +0.8 | 56.6 | +0.2 | 56.4 | +0.0 | 56.9 | 57.6 | +0.7 | 0.00 (0.00–0.10) | 7 |
| Leagues of Votann (1) | 61.0 | 61.0 | +0.0 | 60.0 | −1.0 | 61.0 | +0.0 | 75.2 | 78.1 | +2.9 | 0.81 (0.79–0.81) | 16 |
| Orks (1) | 63.5 | 61.7 | −1.8 | 60.6 | −3.0 | 63.5 | +0.0 | 64.2 | 58.7 | −5.4 | 0.54 (0.41–0.67) | 19 |
| Space Wolves (1) | 50.7 | 50.0 | −0.7 | 48.6 | −2.1 | 50.7 | +0.0 | 54.3 | 52.9 | −1.4 | 0.78 (0.76–0.80) | 9 |
| World Eaters (3) | 67.2 | 63.4 | −3.9 | 61.0 | −6.2 | 67.2 | +0.0 | 72.5 | 62.1 | −10.4 | 0.67 (0.16–0.82) | 17 |
| Tyranids (fitted on, not pooled) (3) | 81.6 | 81.3 | −0.3 | 80.8 | −0.8 | 81.6 | +0.0 | 81.3 | 80.5 | −0.7 | 0.58 (0.51–0.77) | 31 |
Method notes and choices
The package.
- Lists. Each army's 11th-edition lists (events from 20 June) in grimstat-corpus (CC BY 4.0; its listhammer/ folder is not read).
- A detachment is read from the list's detachment line, from its name anywhere in the list, from the first five letters of each of its words in the header (French spellings such as "Phalange Rubricae"), or from an enhancement it alone offers.
- Top-quarter lists count twice.
- A Space Marine chapter we don't build (Ultramarines and the rest) counts as Space Marines.
- tools/bonusshare.json keeps the counts and shares only.
- Effects left out. An effect switched off by default (a rare condition, a 10+ or 12+ upgrade) is left out. A choice's options are weighted by their odds (the Cabal's rituals, World Eaters' blessing pairs) or equally.
- Who receives what.
- A detachment's effects, and a mark on an enemy (the rituals, Oath of Moment), reach only units with the army's faction keyword (Thousand Sons' marks also Scintillating Legions daemons). An effect that names its units (Khorne Daemonkin's Blood Legions) reaches those.
- Other army rules reach the units whose datasheet lists them.
- Allies in a pool (Brood Brothers in Genestealer Cults, Legiones Daemonica) therefore collect little. That is why Genestealer Cults' median share is 0.
- Gates we read from an effect's words: "Psychic ... only" needs a Psychic weapon, "Rapid Fire weapons only", "FLY units only", "INFANTRY; ..." and "not a TITANIC unit".
- Rules with no dice. A rule with no dice counts as one effect of the army's median value. Kinds that are list building or command points are left out. Warpforged Cabal's rule is read as vehicles only (NOTED_GATES).
- Engine pricing.
- Attacking effects use the frozen engine's per-profile damage (profileVs) on the unit's best form's weapons against the five standard targets.
- Defensive effects use the hits it takes to kill the unit from the five standard attackers (the elite one counted as melee).
- A unit's eligibility is the gain it receives ÷ the gain if every weapon and the unit qualified.
The casters.
- Free casters. Assumed to cover all four rituals every turn: every winning list carries Rubric squads and plain Sorcerers.
- Not counted:
- Temporal Surge (a move);
- stratagems (Infernal Fusillade makes inferno bolters Psychic for a phase; it would raise Rubrics' share);
- the Daemon Prince's Spirit Snare;
- the Warpmeld Dagger and Diabolic Savant;
- Scintillating Legions' own auras.
Doubts
- The share mostly measures "shoots, or not". Thousand Sons' package is three-quarters the two Cabal marks, which reach any shooting unit. The Psychic story turns on stratagems and enhancements our data doesn't price.
- The Thousand Sons gain has one list behind it. It is the redundantRolling list; a second reviewer would be needed to trust it.
- Few lists. Sekhetar Cohort's Psychic grant is read from co-occurrence in 9 Thousand Sons lists: it never meets a Psychic bonus.
- The casters rest on two assumptions.
- We assume channelling is always open to a free caster. If players rarely channel with key squads, Magnus's +2 is worth far more: Twist of Fate goes from 28% to 58% on 2D6.
- FOCUS_PTS, 500 points of fire into a ritual's target, sets the scale of every credit.
- Wahapedia's page is a later version. Its 11th-edition Thousand Sons page carries the faction pack v1.2 (August). The ritual numbers above agree with BSData 374f505's (pre-30 Sep) rule as armyrules.json words it.