← Home

Bullet scoring lab

Compare bullet quality bands (Great / OK / Weak), JD relevance, and the Composer selection score after recency and format adjustments. Raw % stays available for debugging — prefer bands when judging strength.

JD context

Comma-separated canonical skills used by the relevance scorer.

Selection weights

Quality and JD relevance are averaged, then recency/format multiply the result. Current total: 1.00.

Bullets

Set target-skill context independently for every bullet.

1
2
3
4
5
6

Ranked results

Sorted by selection score. Bullet quality does not change when the JD changes.

6 bullets
1

Built production .NET APIs handling 10,000 requests per day.

Quality: OKComposer eligibleScale

Metrics: 10,000 requests (scale)

OKBullet quality (39/100)
85/100JD relevance
62/100Selection score

Primary scores

Bullet quality39/100
Band: OK (Great ≥40, OK ≥10, else Weak). How strong this statement is independently of the JD.
JD relevance85/100
100.0% importance × 85.0% evidence
Selection score62/100
Base 61.7% × recency 1.00 × format 1.00

Bullet quality dimensions

Action100/100
Detected action verb "Built" → fixed 100%.
Impact50/100
No impact verb, but action + strong numeric scale → partial impact floor 50.0%.
Leadership0/100
No leadership or mentoring phrase → 0%.
Ownership0/100
No ownership phrase → 0%.
Architecture0/100
No architecture phrase → 0%.
Scale90/100
Detected scale metric "10,000 requests" (value 10000) → graded 90.0%.
Show calculation

1. Bullet quality

Strength of the bullet before JD relevance and recency.

  • Action

    Detected action verb "Built" → fixed 100%.

    100.0% × 10% = 10.0%
  • Measured impact

    No impact verb, but action + strong numeric scale → partial impact floor 50.0%.

    50.0% × 30% = 15.0%
  • Leadership / ownership (best)

    No leadership or mentoring phrase → 0%.

    0.0% × 20% = 0.0%
  • Technical depth (best)

    No architecture, migration, reliability, or performance phrase → 0%. Uses the best of those four.

    0.0% × 25% = 0.0%
  • Scale

    Detected scale metric "10,000 requests" (value 10000) → graded 90.0%.

    90.0% × 15% = 13.5%
  • Bullet quality

    Does not change when the JD changes.

    38.5%

2. JD relevance

JD importance × how strongly this bullet proves the skill.

  • JD importance

    Required missing skill ".NET" → 100%.

    100.0%
  • Target-skill evidence

    Skill shares a clause with action and production/scale evidence → 85%.

    85.0%
  • JD relevance

    Importance × evidence strength.

    100.0% × 85.0% = 85.0%

3. Selection score

How useful this bullet is for the tailored resume.

  • Bullet quality

    From section 1.

    38.5% × 50.0% = 19.3%
  • JD relevance

    From section 2.

    85.0% × 50.0% = 42.5%
  • Base selection score

    Weighted average of quality and JD relevance.

    61.7%
  • Recency multiplier

    Most recent job → ×1.00.

    × 1.00
  • Format multiplier

    Normal achievement bullet → no format penalty (×1.00).

    × 1.00
  • Selection score

    Base × recency × format.

    61.7% × 1.00 × 1.00 = 61.7%
2

Reduced processing time by 40% across a platform used by 3,000 engineers.

Quality: GreatComposer eligibleExceptional general evidenceMeasured impactPerformanceScale

Metrics: 40% (percentage) · 3,000 engineers (scale)

GreatBullet quality (71/100)
0/100JD relevance
61/100Selection score

Primary scores

Bullet quality71/100
Band: Great (Great ≥40, OK ≥10, else Weak). How strong this statement is independently of the JD.
JD relevance0/100
0.0% importance × 0.0% evidence
Selection score61/100
Base 61.4% × recency 1.00 × format 1.00

Bullet quality dimensions

Action100/100
Detected action verb "Reduced" → fixed 100%.
Impact100/100
Impact verb "Reduced" plus outcome metric "40%" (percentage), "3,000 engineers" (scale) → fixed 100%.
Leadership0/100
No leadership or mentoring phrase → 0%.
Ownership0/100
No ownership phrase → 0%.
Architecture0/100
No architecture phrase → 0%.
Scale90/100
Detected scale metric "3,000 engineers" (value 3000) → graded 90.0%.
Show calculation

1. Bullet quality

Strength of the bullet before JD relevance and recency.

  • Action

    Detected action verb "Reduced" → fixed 100%.

    100.0% × 10% = 10.0%
  • Measured impact

    Impact verb "Reduced" plus outcome metric "40%" (percentage), "3,000 engineers" (scale) → fixed 100%.

    100.0% × 30% = 30.0%
  • Leadership / ownership (best)

    No leadership or mentoring phrase → 0%.

    0.0% × 20% = 0.0%
  • Technical depth (best)

    Uses the best of: Performance 70.0% from "processing time".

    70.0% × 25% = 17.5%
  • Scale

    Detected scale metric "3,000 engineers" (value 3000) → graded 90.0%.

    90.0% × 15% = 13.5%
  • Bullet quality

    Does not change when the JD changes.

    71.0%

2. JD relevance

JD importance × how strongly this bullet proves the skill.

  • JD importance

    No target skill → JD importance 0%.

    0.0%
  • Target-skill evidence

    No target skill → evidence strength 0%.

    0.0%
  • JD relevance

    Importance × evidence strength.

    0.0% × 0.0% = 0.0%

3. Selection score

How useful this bullet is for the tailored resume.

  • Bullet quality

    From section 1.

    71.0% × 50.0% = 35.5%
  • JD relevance

    From section 2.

    0.0% × 50.0% = 0.0%
  • Base selection score

    Weighted average of quality and JD relevance.

    61.4%
  • Recency multiplier

    Most recent job → ×1.00.

    × 1.00
  • Format multiplier

    Normal achievement bullet → no format penalty (×1.00).

    × 1.00
  • Selection score

    Base × recency × format.

    61.4% × 1.00 × 1.00 = 61.4%
3

Led a team of 5 engineers through a critical platform migration.

Quality: GreatComposer eligibleExceptional general evidenceLeadershipMigrationScale

Metrics: 5 engineers (scale)

GreatBullet quality (55/100)
0/100JD relevance
48/100Selection score

Primary scores

Bullet quality55/100
Band: Great (Great ≥40, OK ≥10, else Weak). How strong this statement is independently of the JD.
JD relevance0/100
0.0% importance × 0.0% evidence
Selection score48/100
Base 47.8% × recency 1.00 × format 1.00

Bullet quality dimensions

Action100/100
Detected action verb "Led" → fixed 100%.
Impact0/100
No impact verb or outcome metric → 0%.
Leadership100/100
Leadership phrase "Led" with team size 5 → 100.0%.
Ownership0/100
No ownership phrase → 0%.
Architecture0/100
No architecture phrase → 0%.
Scale60/100
Detected scale metric "5 engineers" (value 5) → graded 60.0%.
Show calculation

1. Bullet quality

Strength of the bullet before JD relevance and recency.

  • Action

    Detected action verb "Led" → fixed 100%.

    100.0% × 10% = 10.0%
  • Measured impact

    No impact verb or outcome metric → 0%.

    0.0% × 30% = 0.0%
  • Leadership / ownership (best)

    Leadership phrase "Led" with team size 5 → 100.0%.

    100.0% × 20% = 20.0%
  • Technical depth (best)

    Uses the best of: Migration 65.0% from "migration".

    65.0% × 25% = 16.3%
  • Scale

    Detected scale metric "5 engineers" (value 5) → graded 60.0%.

    60.0% × 15% = 9.0%
  • Bullet quality

    Does not change when the JD changes.

    55.3%

2. JD relevance

JD importance × how strongly this bullet proves the skill.

  • JD importance

    No target skill → JD importance 0%.

    0.0%
  • Target-skill evidence

    No target skill → evidence strength 0%.

    0.0%
  • JD relevance

    Importance × evidence strength.

    0.0% × 0.0% = 0.0%

3. Selection score

How useful this bullet is for the tailored resume.

  • Bullet quality

    From section 1.

    55.3% × 50.0% = 27.7%
  • JD relevance

    From section 2.

    0.0% × 50.0% = 0.0%
  • Base selection score

    Weighted average of quality and JD relevance.

    47.8%
  • Recency multiplier

    Most recent job → ×1.00.

    × 1.00
  • Format multiplier

    Normal achievement bullet → no format penalty (×1.00).

    × 1.00
  • Selection score

    Base × recency × format.

    47.8% × 1.00 × 1.00 = 47.8%
4

Owned the technical design of a distributed event processing platform.

Quality: GreatComposer eligibleExceptional general evidenceOwnershipArchitecture
GreatBullet quality (51/100)
0/100JD relevance
40/100Selection score

Primary scores

Bullet quality51/100
Band: Great (Great ≥40, OK ≥10, else Weak). How strong this statement is independently of the JD.
JD relevance0/100
0.0% importance × 0.0% evidence
Selection score40/100
Base 43.7% × recency 0.92 × format 1.00

Bullet quality dimensions

Action100/100
Detected action verb "Owned" → fixed 100%.
Impact0/100
No impact verb or outcome metric → 0%.
Leadership0/100
No leadership or mentoring phrase → 0%.
Ownership90/100
Ownership phrase "Owned" → 90.0%.
Architecture90/100
Architecture phrase "technical design" → 90.0%.
Scale0/100
No users/requests/services/engineers scale metric → 0%.
Show calculation

1. Bullet quality

Strength of the bullet before JD relevance and recency.

  • Action

    Detected action verb "Owned" → fixed 100%.

    100.0% × 10% = 10.0%
  • Measured impact

    No impact verb or outcome metric → 0%.

    0.0% × 30% = 0.0%
  • Leadership / ownership (best)

    Ownership phrase "Owned" → 90.0%.

    90.0% × 20% = 18.0%
  • Technical depth (best)

    Uses the best of: Architecture 90.0% from "technical design".

    90.0% × 25% = 22.5%
  • Scale

    No users/requests/services/engineers scale metric → 0%.

    0.0% × 15% = 0.0%
  • Bullet quality

    Does not change when the JD changes.

    50.5%

2. JD relevance

JD importance × how strongly this bullet proves the skill.

  • JD importance

    No target skill → JD importance 0%.

    0.0%
  • Target-skill evidence

    No target skill → evidence strength 0%.

    0.0%
  • JD relevance

    Importance × evidence strength.

    0.0% × 0.0% = 0.0%

3. Selection score

How useful this bullet is for the tailored resume.

  • Bullet quality

    From section 1.

    50.5% × 50.0% = 25.3%
  • JD relevance

    From section 2.

    0.0% × 50.0% = 0.0%
  • Base selection score

    Weighted average of quality and JD relevance.

    43.7%
  • Recency multiplier

    Previous job → ×0.92.

    × 0.92
  • Format multiplier

    Normal achievement bullet → no format penalty (×1.00).

    × 1.00
  • Selection score

    Base × recency × format.

    43.7% × 0.92 × 1.00 = 40.2%
5

Tech: React, TypeScript, REST APIs, SQL, Chart.js, Cloud Deployments

Quality: WeakComposer eligibleFormat penalty
WeakBullet quality (0/100)
9/100JD relevance
1/100Selection score

Primary scores

Bullet quality0/100
Band: Weak (Great ≥40, OK ≥10, else Weak). How strong this statement is independently of the JD.
JD relevance9/100
45.0% importance × 20.0% evidence
Selection score1/100
Base 4.5% × recency 0.92 × format 0.35

Bullet quality dimensions

Action0/100
No action verb detected → 0%.
Impact0/100
No impact verb or outcome metric → 0%.
Leadership0/100
No leadership or mentoring phrase → 0%.
Ownership0/100
No ownership phrase → 0%.
Architecture0/100
No architecture phrase → 0%.
Scale0/100
No users/requests/services/engineers scale metric → 0%.
Show calculation

1. Bullet quality

Strength of the bullet before JD relevance and recency.

  • Action

    No action verb detected → 0%.

    0.0% × 10% = 0.0%
  • Measured impact

    No impact verb or outcome metric → 0%.

    0.0% × 30% = 0.0%
  • Leadership / ownership (best)

    No leadership or mentoring phrase → 0%.

    0.0% × 20% = 0.0%
  • Technical depth (best)

    No architecture, migration, reliability, or performance phrase → 0%. Uses the best of those four.

    0.0% × 25% = 0.0%
  • Scale

    No users/requests/services/engineers scale metric → 0%.

    0.0% × 15% = 0.0%
  • Bullet quality

    Does not change when the JD changes.

    0.0%

2. JD relevance

JD importance × how strongly this bullet proves the skill.

  • JD importance

    Preferred matched skill "TypeScript" → 45%.

    45.0%
  • Target-skill evidence

    Skill appears only in a technology list → 20%.

    20.0%
  • JD relevance

    Importance × evidence strength.

    45.0% × 20.0% = 9.0%

3. Selection score

How useful this bullet is for the tailored resume.

  • Bullet quality

    From section 1.

    0.0% × 50.0% = 0.0%
  • JD relevance

    From section 2.

    9.0% × 50.0% = 4.5%
  • Base selection score

    Weighted average of quality and JD relevance.

    4.5%
  • Recency multiplier

    Previous job → ×0.92.

    × 0.92
  • Format multiplier

    Looks like a technology list (Tech: React, TypeScript,…) → ×0.35.

    × 0.35
  • Selection score

    Base × recency × format.

    4.5% × 0.92 × 0.35 = 1.4%
6

Worked on an internal platform.

Quality: WeakWould be filteredFormat penalty
WeakBullet quality (0/100)
0/100JD relevance
0/100Selection score

Primary scores

Bullet quality0/100
Band: Weak (Great ≥40, OK ≥10, else Weak). How strong this statement is independently of the JD.
JD relevance0/100
0.0% importance × 0.0% evidence
Selection score0/100
Base 0.0% × recency 1.00 × format 0.70

Bullet quality dimensions

Action0/100
No action verb detected → 0%.
Impact0/100
No impact verb or outcome metric → 0%.
Leadership0/100
No leadership or mentoring phrase → 0%.
Ownership0/100
No ownership phrase → 0%.
Architecture0/100
No architecture phrase → 0%.
Scale0/100
No users/requests/services/engineers scale metric → 0%.
Show calculation

1. Bullet quality

Strength of the bullet before JD relevance and recency.

  • Action

    No action verb detected → 0%.

    0.0% × 10% = 0.0%
  • Measured impact

    No impact verb or outcome metric → 0%.

    0.0% × 30% = 0.0%
  • Leadership / ownership (best)

    No leadership or mentoring phrase → 0%.

    0.0% × 20% = 0.0%
  • Technical depth (best)

    No architecture, migration, reliability, or performance phrase → 0%. Uses the best of those four.

    0.0% × 25% = 0.0%
  • Scale

    No users/requests/services/engineers scale metric → 0%.

    0.0% × 15% = 0.0%
  • Bullet quality

    Does not change when the JD changes.

    0.0%

2. JD relevance

JD importance × how strongly this bullet proves the skill.

  • JD importance

    No target skill → JD importance 0%.

    0.0%
  • Target-skill evidence

    No target skill → evidence strength 0%.

    0.0%
  • JD relevance

    Importance × evidence strength.

    0.0% × 0.0% = 0.0%

3. Selection score

How useful this bullet is for the tailored resume.

  • Bullet quality

    From section 1.

    0.0% × 50.0% = 0.0%
  • JD relevance

    From section 2.

    0.0% × 50.0% = 0.0%
  • Base selection score

    Weighted average of quality and JD relevance.

    0.0%
  • Recency multiplier

    Most recent job → ×1.00.

    × 1.00
  • Format multiplier

    Very short / weak bullet → ×0.70.

    × 0.70
  • Selection score

    Near-zero quality and relevance → selection gated to 0.

    0.0% × 1.00 × 0.70 = 0.0%