Check version
Check for duplicates
Check for support
What platform are you running Path of Building on?
Windows
What is the value from the calculation in-game?
Three different builds survive the Maven Memory Game in 3.28 gameplay videos, but PoB's Maven Memory Game preset predicts all three should die:
"Min Roll HBD" = PoB HBD ÷ 0.741, representing the absolute best-case damage roll. Build 3 at 0.56 HBD is still dead even at minimum roll (0.76) — there is no damage roll that produces survival under PoB's current values. Yet the character clearly survives in the video.
The builds are correctly rank-ordered by survival margin (worst HBD = closest to death), which indicates a systematic multiplicative overestimate rather than random variance or build-specific unmodeled mechanics.
What is the value from the calculation in Path of Building?
For build 1 (Consecrated Path Juggernaut), PoB shows:
- Maven Memory Game HBD: 0.76 (without guard skills), 0.91 (with Vaal Molten Shell)
- Both predict death
When we calculate using the pre-#9372 damage values (before MonsterPhysConversionMulti was applied to the Maven preset), the numbers match reality:
| Damage Model |
Per Burst (Build 1) |
4 Bursts Total |
vs 5,860 Life |
Matches Video? |
| Current PoB (with ×4 phys conv) |
3,471 |
13,884 |
DEAD (237%) |
No |
| Pre-#9372 values |
868 |
3,472 |
Survives (59%) |
Yes |
| Pre-#9372 + Molten Shell |
651 |
2,604 |
Survives (56%) |
Best match |
The old values predict ~56% life remaining with Molten Shell — matching the video observation of "about half life." Even accounting for the full ±20% damage roll range, the two models never overlap: current PoB kills at minimum roll (175%), old values survive at maximum roll (51% life).
Suspected Cause
PR #9372 applied the MonsterPhysConversionMulti (×4.0 at level 80+) to the Maven Memory Game and Sirus Meteor presets in BossSkills.lua. However, the pre-existing BossSkills.lua multiplier of 104.29 (= 3 × 34.76, originally from PR #5612 / #6806) was derived from game data that likely already included the old per-skill "% more" bonus that GGG replaced with MonsterPhysConversionMulti in 3.27.
The original hardcoded elemental values from PR #5612 (Feb 2023) were:
Lightning = { 34.76, ... }, Cold = { 34.76, ... }, Fire = { 34.76, ... }
These were extracted from the game's actual damage data, which would have included whatever conversion bonus existed at that time. PR #6806 (Dec 2023) refactored to physical-with-conversion (Physical = { 104.29, ... }) while preserving the same total damage. Neither PR verified whether the multiplier already contained the conversion bonus.
When PR #9372 applied the new ×4.0 MonsterPhysConversionMulti on top, the result was effectively applying the bonus twice — once baked into the 104.29 and once as the explicit ×4.0 multiplier.
This is the same class of bug as #9444, where the MonsterPhysConversionMulti was also incorrectly applied to player conversion effects (Shaper of Flames, Kaom's Binding). That was fixed in #9527 where the description notes:
"When I added the fix for scaling enemy damage values by player and enemy conversion mods (#9372), I forgot to include the fix where only enemy skill damage conversion is affected by the phys conversion multiplier"
The ×4.0 multiplier has been over-applied before in this same PR chain. The question is whether the BossSkills.lua base values already accounted for the old equivalent bonus.
The preset tooltip contradicts current values
The Maven Memory Game preset tooltip states:
"It is three separate hits, and has a large DoT effect. Neither is taken into account here. i.e. Hits before death should be more than 3 to survive"
This creates a confusing situation with current values. The tooltip says "more than 3" hits before death is needed — but PoB's displayed HBD is for the combined total of all bursts, not per individual hit. For builds with 0.56–0.89 combined HBD, the per-hit HBD would be:
| Build |
Combined HBD |
Per-Hit HBD (÷4 bursts) |
Tooltip says "> 3" |
| Conc Path Jugg |
0.76 |
3.04 |
Borderline |
| Ward Loop Sab |
0.89 |
3.56 |
Passes |
| EB Inquisitor |
0.56 |
2.24 |
Fails |
With current values, the tooltip threshold is nearly impossible to reach for most builds. The EB Inquisitor at 2.24 per-hit HBD clearly fails the tooltip's own "> 3" check — yet survives in-game. Either the tooltip threshold is wrong, or the damage values are wrong.
With pre-#9372 values, per-hit HBD would be ~4× higher (e.g., the Inquisitor would be ~8.96 per hit), and the "> 3" threshold would make sense again as a reasonable survival check — which is consistent with it being written for those original damage values in PR #5612 (Feb 2023).
How to reproduce the issue
- Import any of these builds:
- Set Boss Skill Preset to "Maven Memory Game"
- Set enemy type to "Guardian/Pinnacle Boss"
- Observe "Hits before death" < 1.0 (predicting death)
- Watch the corresponding video — character survives
PoB for PoE1 build code
| # | Build | PoB HBD (Maven Memory Game) | Even at Min Roll | Video | Outcome |
|:-:|-------|:---------------------------:|:----------------:|-------|---------|
| 1 | Consecrated Path Juggernaut (17 endurance charges) | 0.76 | 1.03 | https://youtu.be/0zunlKeyDo4 | Survives at ~50% life |
| 2 | Ward Loop Saboteur | 0.89 | 1.20 | https://youtu.be/QwBxsoFf-hg | Survives |
| 3 | Energy Blade Shock Nova Inquisitor | 0.56 | 0.76 | https://youtu.be/fnNSONBQHjc | Barely survives (~5-10% life) |
Screenshots
Can provide if needed — the discrepancy is reproducible with any of the three builds above.
Check version
Check for duplicates
Check for support
What platform are you running Path of Building on?
Windows
What is the value from the calculation in-game?
Three different builds survive the Maven Memory Game in 3.28 gameplay videos, but PoB's Maven Memory Game preset predicts all three should die:
"Min Roll HBD" = PoB HBD ÷ 0.741, representing the absolute best-case damage roll. Build 3 at 0.56 HBD is still dead even at minimum roll (0.76) — there is no damage roll that produces survival under PoB's current values. Yet the character clearly survives in the video.
The builds are correctly rank-ordered by survival margin (worst HBD = closest to death), which indicates a systematic multiplicative overestimate rather than random variance or build-specific unmodeled mechanics.
What is the value from the calculation in Path of Building?
For build 1 (Consecrated Path Juggernaut), PoB shows:
When we calculate using the pre-#9372 damage values (before
MonsterPhysConversionMultiwas applied to the Maven preset), the numbers match reality:The old values predict ~56% life remaining with Molten Shell — matching the video observation of "about half life." Even accounting for the full ±20% damage roll range, the two models never overlap: current PoB kills at minimum roll (175%), old values survive at maximum roll (51% life).
Suspected Cause
PR #9372 applied the
MonsterPhysConversionMulti(×4.0 at level 80+) to the Maven Memory Game and Sirus Meteor presets in BossSkills.lua. However, the pre-existing BossSkills.lua multiplier of104.29(= 3 × 34.76, originally from PR #5612 / #6806) was derived from game data that likely already included the old per-skill "% more" bonus that GGG replaced withMonsterPhysConversionMultiin 3.27.The original hardcoded elemental values from PR #5612 (Feb 2023) were:
These were extracted from the game's actual damage data, which would have included whatever conversion bonus existed at that time. PR #6806 (Dec 2023) refactored to physical-with-conversion (
Physical = { 104.29, ... }) while preserving the same total damage. Neither PR verified whether the multiplier already contained the conversion bonus.When PR #9372 applied the new ×4.0
MonsterPhysConversionMultion top, the result was effectively applying the bonus twice — once baked into the 104.29 and once as the explicit ×4.0 multiplier.Related: #9444 / #9527
This is the same class of bug as #9444, where the
MonsterPhysConversionMultiwas also incorrectly applied to player conversion effects (Shaper of Flames, Kaom's Binding). That was fixed in #9527 where the description notes:The ×4.0 multiplier has been over-applied before in this same PR chain. The question is whether the BossSkills.lua base values already accounted for the old equivalent bonus.
The preset tooltip contradicts current values
The Maven Memory Game preset tooltip states:
This creates a confusing situation with current values. The tooltip says "more than 3" hits before death is needed — but PoB's displayed HBD is for the combined total of all bursts, not per individual hit. For builds with 0.56–0.89 combined HBD, the per-hit HBD would be:
With current values, the tooltip threshold is nearly impossible to reach for most builds. The EB Inquisitor at 2.24 per-hit HBD clearly fails the tooltip's own "> 3" check — yet survives in-game. Either the tooltip threshold is wrong, or the damage values are wrong.
With pre-#9372 values, per-hit HBD would be ~4× higher (e.g., the Inquisitor would be ~8.96 per hit), and the "> 3" threshold would make sense again as a reasonable survival check — which is consistent with it being written for those original damage values in PR #5612 (Feb 2023).
How to reproduce the issue
PoB for PoE1 build code
Screenshots
Can provide if needed — the discrepancy is reproducible with any of the three builds above.