The only truly tight constraint in the whole defense system: reviewer accuracy
We swept the stake, probation length, ecosystem size, attacker share, and the correlation of reviewer errors — all of them are slack. The only cell where attacking starts to pay is when reviewer accuracy drops to 0.
In the post on stakes, we deleted the "exogenous entry fee" and "endorser liability." The reason was not that they had a weaker effect, but that one was actively harmful and the other was a pure loss.
But once those two paths were deleted, what does this system rely on to stop attackers?
#Five parameters swept, four of them slack
#Stake
Slack. stake=0 blocks completely; stake=2 gives a return of 2.354.
#Probation length
Very tight, but in the opposite direction. m=0 → 6.85, 0% fully blocked; m≥1 → 0.00, 100% fully blocked. One step does it all; after that there is no discrimination at all.
#Ecosystem size (E8)
Completely slack. With 2 / 3 / 5 / 10 / 20 honest participants, all 0.00, 100% fully blocked.
#Attacker share (E14)
| Attackers / total | Attacker hits | Return per attacker | Full-block rate |
|---|---|---|---|
| 1 / 6 | 2.30 | 0.144 | 5% |
| 2 / 7 | 4.00 | 0.125 | — |
| 3 / 8 | 3.80 | 0.079 | — |
| 5 / 10 | 5.40 | 0.068 | — |
| 8 / 13 | 7.45 ± 1.07 | 0.058 | 0% |
#Correlation of reviewer errors (E16)
| Correlation ρ | Attacker hits | Standard deviation |
|---|---|---|
| 0.0 (independent) | 2.40 | 0.58 |
| 0.3 | 2.30 | 0.95 |
| 0.6 | 2.50 | 0.67 |
| 0.9 | 2.45 | 1.07 |
| 1.0 (fully identical) | 2.85 | 1.98 |
#The only tight one: reviewer accuracy (E12)
| Reviewer accuracy | Routed to attacker | Attacker hits | Return |
|---|---|---|---|
| 1.0 (perfect reviewers) | 18.0 | 1.90 ± 0.30 | 0.119 |
| 0.7 (our setting) | — | 2.35 | — |
| 0.5 | — | 2.70 | — |
| 0.3 | — | 4.30 | 0.269 |
| 0.0 (no reviewers) | 33.6 | 16.65 ± 12.89 | 1.041 |
acc = 0 is the only cell in the whole table where "attacking starts to pay." A return of 1.041 means the attacker can make money.
And its standard deviation is 12.89 — more than half of the mean. This cell is not just dangerous; it is also unpredictable.
#Why all the other parameters are slack
Looked at within one framework, defensive power actually comes from three things, and only three:
| Mechanism | What it guards against |
|---|---|
| Being seen (reputation routing) | Doing bad work and still getting the next job |
| Reviewer accuracy | Bad work being mistaken for good |
| Deterministic sampled audits | Reviewers lying collectively |
Of these three, two are functions of reviewer quality. Stake, probation, ecosystem size, attacker share, reviewer correlation — all of them are downstream parameters of these three things, and they cannot change those three things.
#The practical consequences of this finding
That +0.29 pp is what "reviewer accuracy ≈ zero-intelligence baseline" means. In the E12 table, the attacker hits in the accuracy 0.7 row are 2.35 — our system actually lives in that row.
#The boundary of this conclusion}}
#Next
- Why reviewers are close to a constant: reviewer-tracks-difficulty
- How noteworthy that "zero-intelligence baseline" is: committee-loses-to-majority
- The whole defense axis: Defense axis
