Sensitivity study
10brands · prompt1 × 10
Stability by metric
Spread across trials, in score points. A pair counts as unstable when its standard deviation across trials exceeds 5 points.
| Metric | Pairs | Mean SD | Median SD | Max SD | Within variant | Between variants | Unstable |
|---|---|---|---|---|---|---|---|
| Cluny Affinity | 45 | 1.217 | 1.153 | 2.373 | 1.217 | — | 0 |
| Brand affinity | 45 | 1.375 | 1.332 | 2.928 | 1.375 | — | 0 |
| Market proximity | 45 | 1.344 | 1.315 | 2.669 | 1.344 | — | 0 |
Against the production matrix
Compared with run 4d388d02 (published, universe brands, 100 brands). Only pairs present in both are counted.
| Metric | Pairs | Mean delta | MAE | RMSE | Max delta | Pearson r | Spearman ρ |
|---|---|---|---|---|---|---|---|
| Cluny Affinity | 45 | 2.899 | 3.329 | 3.899 | 7.177 | 0.983 | 0.985 |
| Brand affinity | 45 | 0.432 | 2.082 | 2.534 | 6.542 | 0.989 | 0.988 |
| Market proximity | 45 | 4.276 | 4.444 | 5.21 | 10.353 | 0.973 | 0.958 |
Trials
Each trial is preserved as its own affinity run and can be exported on its own with brand-map:export.
| Run | Variant | Replicate | Status | Complete pairs |
|---|---|---|---|---|
| dc8a4ac7 | prompt1 8b4598c4b705 | 1 | completed | 45 |
| 884756ba | prompt1 8b4598c4b705 | 2 | completed | 45 |
| 1bbad344 | prompt1 8b4598c4b705 | 3 | completed | 45 |
| 27e73320 | prompt1 8b4598c4b705 | 4 | completed | 45 |
| 644df0ad | prompt1 8b4598c4b705 | 5 | completed | 45 |
| 0daf53a3 | prompt1 8b4598c4b705 | 6 | completed | 45 |
| e84afff1 | prompt1 8b4598c4b705 | 7 | completed | 45 |
| 61e2bcca | prompt1 8b4598c4b705 | 8 | completed | 45 |
| 42826e6c | prompt1 8b4598c4b705 | 9 | running | 0 |
| a87d1ee0 | prompt1 8b4598c4b705 | 10 | completed | 45 |
Least stable pairs
Ranked by spread across trials, worst first, for each metric.
| Pair | Mean | SD | Range | Within variant | Between variants | Production | Delta |
|---|---|---|---|---|---|---|---|
| Le Labo · Rare Beauty | 42.472 | 2.373 | 7.213 | 2.373 | — | 38.625 | 3.847 |
| Flamingo Estate · Supergoop! | 54.043 | 2.129 | 6.838 | 2.129 | — | 51.95 | 2.093 |
| EltaMD · La Mer | 47.866 | 1.898 | 5.787 | 1.898 | — | 49.538 | -1.672 |
| Flamingo Estate · Rare Beauty | 39.2 | 1.845 | 5.625 | 1.845 | — | 38.613 | 0.587 |
| MALIN+GOETZ · Supergoop! | 67.682 | 1.818 | 4.875 | 1.818 | — | 63.163 | 4.519 |
| EltaMD · Rare Beauty | 48.36 | 1.692 | 4.838 | 1.692 | — | 49.213 | -0.853 |
| Aesop · EltaMD | 48.038 | 1.616 | 4.55 | 1.616 | — | 47.275 | 0.763 |
| EltaMD · Merit | 56.175 | 1.516 | 5.6 | 1.516 | — | 53.575 | 2.6 |
| Aesop · Supergoop! | 59.311 | 1.51 | 5.225 | 1.51 | — | 54.488 | 4.823 |
| La Mer · Rare Beauty | 40.936 | 1.501 | 4.125 | 1.501 | — | 35.588 | 5.348 |
| Flamingo Estate · La Mer | 56.421 | 1.475 | 3.937 | 1.475 | — | 55.188 | 1.233 |
| MALIN+GOETZ · Rare Beauty | 49.92 | 1.456 | 4.425 | 1.456 | — | 49.275 | 0.645 |
| Aesop · Merit | 66.002 | 1.43 | 4.588 | 1.43 | — | 62 | 4.002 |
| Rare Beauty · Supergoop! | 73.853 | 1.409 | 3.925 | 1.409 | — | 71.725 | 2.128 |
| Le Labo · Merit | 63.433 | 1.398 | 4.025 | 1.398 | — | 56.925 | 6.508 |
| Aesop · Rare Beauty | 44.253 | 1.304 | 4.7 | 1.304 | — | 42.125 | 2.128 |
| MALIN+GOETZ · Merit | 68.012 | 1.282 | 3.262 | 1.282 | — | 64.438 | 3.574 |
| La Mer · Supergoop! | 54.234 | 1.264 | 3.612 | 1.264 | — | 48.875 | 5.359 |
| Flamingo Estate · Merit | 61.841 | 1.225 | 3.525 | 1.225 | — | 57.825 | 4.016 |
| Merit · Oribe | 64.602 | 1.223 | 3.475 | 1.223 | — | 57.425 | 7.177 |
| Aesop · Oribe | 70.82 | 1.188 | 3.562 | 1.188 | — | 65.825 | 4.995 |
| La Mer · Merit | 56.31 | 1.168 | 3.775 | 1.168 | — | 50.5 | 5.81 |
| La Mer · MALIN+GOETZ | 56.888 | 1.153 | 4.112 | 1.153 | — | 50.488 | 6.4 |
| EltaMD · Oribe | 46.614 | 1.133 | 3.201 | 1.133 | — | 42.125 | 4.489 |
| Oribe · Supergoop! | 59.821 | 1.094 | 3.225 | 1.094 | — | 52.738 | 7.083 |