Skip to content

Commit 0f93b84

Browse files
NeaByteLabNeaByteLab
committed
docs(personal): add 6D+1M claim check doc 🚶
- Add 6D+1M claim check framework doc under PERSONAL - Add 6D+1M claim check link to PERSONAL README - Remove shortcut sections from root README
1 parent 40f592a commit 0f93b84

File tree

3 files changed

+168
-37
lines changed

3 files changed

+168
-37
lines changed

‎PERSONAL/6D-1M.md‎

Lines changed: 167 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,167 @@
1+
# 6D+1M — 30-Second Claim Check
2+
3+
_Because "sounds right" isn’t a basis._
4+
5+
## Table of Contents
6+
7+
- [Overview](#overview)
8+
- [The Framework](#the-framework)
9+
- [30-Second Flow](#30-second-flow)
10+
- [Rubric (Fast Scoring)](#rubric-fast-scoring)
11+
- [One-Line Template](#one-line-template)
12+
- [Examples](#examples)
13+
- [Common Failure Modes](#common-failure-modes)
14+
- [When to Escalate](#when-to-escalate)
15+
16+
---
17+
18+
## Overview
19+
20+
6D+1M is a tiny framework to filter statements fast, without turning your brain into a full-time fact-checking department.
21+
22+
It answers one question:
23+
24+
Can I treat this claim as **actionable**, **tentative**, or **noise**?
25+
26+
### Quick Takeaways
27+
28+
1. **Definition** — If terms are vague, the claim is vague
29+
2. **Data** — No observable anchor, no confidence
30+
3. **Design** — Correlation isn’t causation; design limits conclusions
31+
4. **Distortions** — Confounds and biases create fake “truth”
32+
5. **Degree** — Effect size matters more than vibes
33+
6. **Mind-change** — Define what evidence would update your belief
34+
35+
---
36+
37+
## The Framework
38+
39+
### 1) Definition
40+
41+
Goal: make the claim _operational_, not poetic.
42+
43+
- What does the key term mean **exactly**?
44+
- Who/where/when does it apply? (population, context, horizon)
45+
46+
### 2) Data
47+
48+
Goal: locate the observable anchor.
49+
50+
- Is there **measurement** (numbers/records), or only assertion?
51+
- Primary source or hearsay? Is the method transparent?
52+
53+
### 3) Design (Study / Evidence Design)
54+
55+
Goal: know what kind of inference is allowed.
56+
57+
- RCT / natural experiment / longitudinal / cross-sectional / case study?
58+
- Is the claim **correlational** or **causal**?
59+
60+
### 4) Distortions (Confounds / Bias)
61+
62+
Goal: name the 1–2 strongest illusion generators.
63+
64+
- What likely confounds could explain it? (SES, selection bias, reverse causality)
65+
- Were they controlled, discussed, or tested for robustness?
66+
67+
### 5) Degree (Effect Size / Practical Impact)
68+
69+
Goal: separate “real” from “technically significant.”
70+
71+
- How big is the effect? (effect size, RR/OR, points, %, practical change)
72+
- Is it **material** in real life, or only a p-value trophy?
73+
74+
### +1) Mind-change (What Would Update My Mind)
75+
76+
Goal: avoid identity-based beliefs.
77+
78+
- What evidence would make you revise? (meta-analysis, prereg replication, better design, stronger data)
79+
- What result would falsify the claim?
80+
81+
---
82+
83+
## 30-Second Flow
84+
85+
Ask these in order:
86+
87+
1. **Definition**: do I understand the claim in measurable terms?
88+
2. **Data**: what observation backs it?
89+
3. **Design**: what inference is allowed (cause vs association)?
90+
4. **Distortions**: what confound/bias could fake it?
91+
5. **Degree**: is it big enough to matter?
92+
6. **Mind-change**: what would change my mind?
93+
94+
Output: **Actionable / Tentative / Noise**
95+
96+
- **Actionable**: terms clear, evidence decent, distortion addressed, magnitude meaningful
97+
- **Tentative**: partial evidence, unclear magnitude, distortions plausible
98+
- **Noise**: vague terms + no data + unsupported causal leap
99+
100+
---
101+
102+
## Rubric (Fast Scoring)
103+
104+
Score each axis 0–2.
105+
106+
- **0**: missing / vague / not addressed
107+
- **1**: partial / weak / unclear
108+
- **2**: clear / strong / directly addressed
109+
110+
Suggested interpretation:
111+
112+
- **9–12**: actionable (with normal caution)
113+
- **5–8**: tentative (ok to test with low-risk experiments)
114+
- **0–4**: noise (ignore or ask for better grounding)
115+
116+
---
117+
118+
## One-Line Template
119+
120+
Use this as the default output (copy-paste friendly):
121+
122+
> This claim is **[well-defined / vague]**, grounded in **[data + design]**, vulnerable to **[main distortion]**, the effect is **[small / medium / large]**, and I’d update if **[evidence X]**.
123+
124+
---
125+
126+
## Examples
127+
128+
### Example A — “Cold shower makes you disciplined.”
129+
130+
- **Definition**: “disciplined” = what behavior, measured how?
131+
- **Data**: anecdotes vs tracked adherence metrics
132+
- **Design**: correlation (people who already train hard do cold showers) vs intervention
133+
- **Distortions**: selection bias, novelty effect
134+
- **Degree**: does it improve adherence meaningfully, or just mood for 30 minutes?
135+
- **Mind-change**: “If a well-run RCT shows no adherence change vs control, I downgrade belief.”
136+
137+
Verdict: usually **tentative** → test it low-risk with tracking.
138+
139+
### Example B — “X causes Y.”
140+
141+
Instant checks:
142+
143+
- If the source is cross-sectional, treat “causes” as **overclaim**.
144+
- If there’s no counterfactual design, ask for **mechanism + robustness**.
145+
146+
Verdict: **tentative/noise** until design supports causality.
147+
148+
---
149+
150+
## Common Failure Modes
151+
152+
- **Paper worship**: “No paper = false.” Wrong. It may be unstudied or context-specific.
153+
- **Vibe worship**: “Feels right = true.” Also wrong.
154+
- **Significance trophy**: p-value without magnitude/practical impact.
155+
- **Survivorship bias**: you only see winners posting outcomes.
156+
- **Reverse causality**: Y causes X, not the other way.
157+
158+
---
159+
160+
## When to Escalate
161+
162+
Use the 30-second check first. Escalate to deeper research when:
163+
164+
- Stakes are high (health, money, irreversible decisions)
165+
- You’re about to build a habit/system around the claim
166+
- You see strong incentives for the source to mislead
167+
- The claim is being used to justify a policy or moral judgement

‎PERSONAL/README.md‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -4,6 +4,7 @@ Real stories from the development trenches.
44

55
## Contents
66

7+
- [6D+1M Claim Check](./6D-1M.md) - 30-second claim filter framework
78
- [Dev Stories](./STORIES.md) - Tales from the keyboard trenches
89

910
---

‎README.md‎

Lines changed: 0 additions & 37 deletions
Original file line numberDiff line numberDiff line change
@@ -7,41 +7,4 @@ class Me {
77
}
88
```
99

10-
# About This Space
11-
12-
Architecture, principles, dev journey, and the occasional "why did I think `any` was a good idea?" moment. 🤦‍
13-
14-
## Shortcuts
15-
16-
**Architecture**
17-
18-
| Doc | One-line |
19-
| -------------------------------------------------------- | ------------------------------ |
20-
| [README](./ARCHITECTURE/README.md) | Index |
21-
| [Code Organization](./ARCHITECTURE/CODE-ORGANIZATION.md) | Where everything lives and why |
22-
| [Decision Tree](./ARCHITECTURE/DECISION-TREE.md) | "It depends" needs a structure |
23-
24-
**Principles**
25-
26-
| Doc | One-line |
27-
| -------------------------------- | --------------------- |
28-
| [README](./PRINCIPLES/README.md) | Index |
29-
| [DRY](./PRINCIPLES/DRY.md) | Don't Repeat Yourself |
30-
| [KISS](./PRINCIPLES/KISS.md) | Keep It Simple Stupid |
31-
32-
**Personal**
33-
34-
| Doc | One-line |
35-
| ------------------------------------ | -------------------------------- |
36-
| [README](./PERSONAL/README.md) | Index |
37-
| [Dev Stories](./PERSONAL/STORIES.md) | Tales from the keyboard trenches |
38-
39-
**Deep Notes**
40-
41-
| Doc | One-line |
42-
| -------------------------------- | ------------------------------ |
43-
| [README](./DEEP-NOTES/README.md) | Catalog (e.g. Time Complexity) |
44-
45-
---
46-
4710
> _"The fuller the rice stalk, the more it bends down 🌾"_ — My father, since I was 5 years old

0 commit comments

Comments
 (0)