This table is an experimental balance watch list, not a forecast. It does not know Valve's plans. It looks at item history, item type, cost, API item stats, win-rate movement, and usage signals to produce an exploratory model score, not a calibrated probability.
The L1 logistic screen is the lasso-style step. It tries to zero out weaker predictors before the tree models run. Stage 1 predicts whether an item changes. Stage 2 only looks at changed items and predicts buff versus nerf.
156 items are included. The table is sorted by highest experimental score first, then bootstrap stability.
| Item | Experimental score | Direction | Direction confidence | Bootstrap stability | Notes |
|---|---|---|---|---|---|
| 98% | Nerf | 68% | 60% | Moderate nerf signal. Review with patch context. | |
| 98% | Buff | 67% | 78% | Strong stable buff signal. | |
| 98% | Buff | 75% | 97% | Strong stable buff signal. | |
| 97% | Buff | 54% | 85% | Change signal is present, but buff versus nerf direction is unclear. | |
| 97% | Buff | 74% | 87% | Strong stable buff signal. | |
| 97% | Buff | 73% | 100% | Strong stable buff signal. | |
| 97% | Buff | 74% | 92% | Strong stable buff signal. | |
| 97% | Nerf | 56% | 78% | Change signal is present, but buff versus nerf direction is unclear. | |
| 97% | Buff | 55% | 68% | Change signal is present, but buff versus nerf direction is unclear. | |
| 97% | Buff | 73% | 97% | Strong stable buff signal. | |
| 97% | Nerf | 75% | 92% | Strong stable nerf signal. | |
| 97% | Nerf | 58% | 78% | Change signal is present, but buff versus nerf direction is unclear. | |
| 97% | Buff | 70% | 82% | Strong stable buff signal. | |
| 96% | Nerf | 64% | 77% | Strong stable nerf signal. | |
| 96% | Buff | 51% | 62% | Change signal is present, but buff versus nerf direction is unclear. | |
| 96% | Buff | 76% | 100% | Strong stable buff signal. | |
| 96% | Buff | 58% | 82% | Change signal is present, but buff versus nerf direction is unclear. | |
| 95% | Buff | 55% | 77% | Change signal is present, but buff versus nerf direction is unclear. | |
| 95% | Buff | 88% | 98% | Strong stable buff signal. | |
| 95% | Buff | 84% | 98% | Strong stable buff signal. | |
| 95% | Nerf | 75% | 82% | Strong stable nerf signal. | |
| 95% | Buff | 58% | 85% | Change signal is present, but buff versus nerf direction is unclear. | |
| 95% | Buff | 69% | 77% | Strong stable buff signal. | |
| 95% | Buff | 79% | 93% | Strong stable buff signal. | |
| 95% | Nerf | 63% | 90% | Strong stable nerf signal. | |
| 95% | Buff | 81% | 92% | Strong stable buff signal. | |
| 95% | Buff | 66% | 77% | Strong stable buff signal. | |
| 95% | Buff | 76% | 70% | Moderate buff signal. Review with patch context. | |
| 95% | Nerf | 68% | 77% | Strong stable nerf signal. | |
| 95% | Buff | 73% | 92% | Strong stable buff signal. | |
| 95% | Buff | 53% | 57% | Unstable prediction. Treat this as a watch-list item, not a call. | |
| 95% | Buff | 66% | 97% | Strong stable buff signal. | |
| 94% | Nerf | 80% | 97% | Strong stable nerf signal. | |
| 94% | Nerf | 73% | 95% | Strong stable nerf signal. | |
| 94% | Buff | 62% | 63% | Moderate buff signal. Review with patch context. | |
| 94% | Buff | 68% | 87% | Strong stable buff signal. | |
| 94% | Buff | 53% | 80% | Change signal is present, but buff versus nerf direction is unclear. | |
| 94% | Nerf | 73% | 78% | Strong stable nerf signal. | |
| 94% | Nerf | 69% | 83% | Strong stable nerf signal. | |
| 93% | Buff | 70% | 77% | Strong stable buff signal. | |
| 93% | Buff | 68% | 60% | Moderate buff signal. Review with patch context. | |
| 93% | Nerf | 54% | 72% | Change signal is present, but buff versus nerf direction is unclear. | |
| 93% | Nerf | 63% | 93% | Strong stable nerf signal. | |
| 93% | Buff | 86% | 98% | Strong stable buff signal. | |
| 93% | Buff | 62% | 78% | Strong stable buff signal. | |
| 93% | Nerf | 72% | 72% | Moderate nerf signal. Review with patch context. | |
| 93% | Buff | 55% | 68% | Change signal is present, but buff versus nerf direction is unclear. | |
| 93% | Buff | 87% | 98% | Strong stable buff signal. | |
| 93% | Buff | 86% | 92% | Strong stable buff signal. | |
| 92% | Nerf | 74% | 90% | Strong stable nerf signal. | |
| 92% | Buff | 52% | 63% | Change signal is present, but buff versus nerf direction is unclear. | |
| 92% | Buff | 87% | 95% | Strong stable buff signal. | |
| 92% | Buff | 81% | 92% | Strong stable buff signal. | |
| 92% | Nerf | 72% | 92% | Strong stable nerf signal. | |
| 92% | Buff | 70% | 92% | Strong stable buff signal. | |
| 92% | Buff | 61% | 72% | Moderate buff signal. Review with patch context. | |
| 92% | Buff | 61% | 83% | Strong stable buff signal. | |
| 92% | Nerf | 51% | 70% | Change signal is present, but buff versus nerf direction is unclear. | |
| 92% | Buff | 80% | 65% | Moderate buff signal. Review with patch context. | |
| 92% | Nerf | 58% | 62% | Change signal is present, but buff versus nerf direction is unclear. | |
| 92% | Nerf | 73% | 100% | Strong stable nerf signal. | |
| 92% | Buff | 55% | 62% | Change signal is present, but buff versus nerf direction is unclear. | |
| 92% | Nerf | 79% | 68% | Moderate nerf signal. Review with patch context. | |
| 91% | Buff | 72% | 98% | Strong stable buff signal. | |
| 91% | Buff | 81% | 95% | Strong stable buff signal. | |
| 91% | Nerf | 87% | 90% | Strong stable nerf signal. | |
| 91% | Buff | 83% | 97% | Strong stable buff signal. | |
| 91% | Buff | 70% | 72% | Moderate buff signal. Review with patch context. | |
| 91% | Buff | 77% | 88% | Strong stable buff signal. | |
| 91% | Buff | 70% | 87% | Strong stable buff signal. | |
| 90% | Buff | 77% | 98% | Strong stable buff signal. | |
| 90% | Buff | 79% | 88% | Strong stable buff signal. | |
| 90% | Nerf | 55% | 73% | Change signal is present, but buff versus nerf direction is unclear. | |
| 90% | Nerf | 66% | 72% | Moderate nerf signal. Review with patch context. | |
| 90% | Buff | 90% | 98% | Strong stable buff signal. | |
| 90% | Buff | 88% | 98% | Strong stable buff signal. | |
| 90% | Nerf | 70% | 52% | Unstable prediction. Treat this as a watch-list item, not a call. | |
| 90% | Buff | 80% | 100% | Strong stable buff signal. | |
| 90% | Nerf | 53% | 78% | Change signal is present, but buff versus nerf direction is unclear. | |
| 90% | Nerf | 73% | 98% | Strong stable nerf signal. | |
| 89% | Nerf | 81% | 98% | Strong stable nerf signal. | |
| 89% | Buff | 83% | 88% | Strong stable buff signal. | |
| 89% | Buff | 72% | 85% | Strong stable buff signal. | |
| 89% | Nerf | 81% | 98% | Strong stable nerf signal. | |
| 88% | Buff | 80% | 82% | Strong stable buff signal. | |
| 88% | Buff | 73% | 68% | Moderate buff signal. Review with patch context. | |
| 88% | Nerf | 57% | 62% | Change signal is present, but buff versus nerf direction is unclear. | |
| 88% | Nerf | 73% | 92% | Strong stable nerf signal. | |
| 88% | Buff | 55% | 63% | Change signal is present, but buff versus nerf direction is unclear. | |
| 88% | Buff | 85% | 100% | Strong stable buff signal. | |
| 88% | Buff | 79% | 85% | Strong stable buff signal. | |
| 88% | Nerf | 81% | 95% | Strong stable nerf signal. | |
| 88% | Buff | 64% | 78% | Strong stable buff signal. | |
| 87% | Nerf | 69% | 73% | Moderate nerf signal. Review with patch context. | |
| 87% | Nerf | 69% | 82% | Strong stable nerf signal. | |
| 87% | Nerf | 86% | 90% | Strong stable nerf signal. | |
| 87% | Nerf | 70% | 90% | Strong stable nerf signal. | |
| 87% | Nerf | 50% | 58% | Unstable prediction. Treat this as a watch-list item, not a call. | |
| 87% | Buff | 81% | 95% | Strong stable buff signal. | |
| 87% | Nerf | 59% | 78% | Change signal is present, but buff versus nerf direction is unclear. | |
| 87% | Nerf | 58% | 53% | Unstable prediction. Treat this as a watch-list item, not a call. | |
| 86% | Buff | 84% | 92% | Strong stable buff signal. | |
| 86% | Buff | 86% | 98% | Strong stable buff signal. | |
| 86% | Nerf | 62% | 93% | Strong stable nerf signal. | |
| 86% | Buff | 85% | 90% | Strong stable buff signal. | |
| 86% | Buff | 81% | 75% | Strong stable buff signal. | |
| 86% | Nerf | 89% | 90% | Strong stable nerf signal. | |
| 86% | Buff | 77% | 88% | Strong stable buff signal. | |
| 85% | Nerf | 73% | 62% | Moderate nerf signal. Review with patch context. | |
| 85% | Buff | 76% | 80% | Strong stable buff signal. | |
| 85% | Buff | 65% | 62% | Moderate buff signal. Review with patch context. | |
| 84% | Nerf | 67% | 78% | Strong stable nerf signal. | |
| 84% | Buff | 82% | 90% | Strong stable buff signal. | |
| 84% | Nerf | 57% | 65% | Change signal is present, but buff versus nerf direction is unclear. | |
| 84% | Nerf | 51% | 72% | Change signal is present, but buff versus nerf direction is unclear. | |
| 83% | Buff | 83% | 88% | Strong stable buff signal. | |
| 82% | Nerf | 65% | 80% | Strong stable nerf signal. | |
| 81% | Nerf | 52% | 58% | Unstable prediction. Treat this as a watch-list item, not a call. | |
| 80% | Nerf | 68% | 95% | Strong stable nerf signal. | |
| 80% | Nerf | 70% | 93% | Strong stable nerf signal. | |
| 80% | Buff | 68% | 75% | Strong stable buff signal. | |
| 79% | Nerf | 60% | 80% | Strong stable nerf signal. | |
| 78% | Nerf | 68% | 77% | Strong stable nerf signal. | |
| 77% | Buff | 65% | 68% | Moderate buff signal. Review with patch context. | |
| 77% | Buff | 71% | 82% | Strong stable buff signal. | |
| 76% | Buff | 81% | 68% | Moderate buff signal. Review with patch context. | |
| 75% | Buff | 59% | 62% | Change signal is present, but buff versus nerf direction is unclear. | |
| 74% | Nerf | 51% | 75% | Change signal is present, but buff versus nerf direction is unclear. | |
| 74% | Buff | 66% | 88% | Moderate buff signal. Review with patch context. | |
| 74% | Buff | 86% | 98% | Moderate buff signal. Review with patch context. | |
| 70% | Nerf | 61% | 73% | Moderate nerf signal. Review with patch context. | |
| 45% | Nerf | 82% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 38% | Nerf | 55% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 37% | Buff | 50% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 36% | Buff | 72% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 34% | Buff | 77% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 34% | Buff | 60% | 48% | Unstable prediction. Treat this as a watch-list item, not a call. | |
| 34% | Nerf | 85% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 34% | Buff | 65% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 33% | Nerf | 84% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 33% | Buff | 65% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 33% | Nerf | 79% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 32% | Buff | 56% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 31% | Buff | 89% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 31% | Buff | 83% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 29% | Nerf | 53% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 28% | Buff | 87% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 28% | Buff | 58% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 27% | Buff | 83% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 26% | Nerf | 80% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 26% | Nerf | 78% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 26% | Nerf | 57% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 26% | Buff | 56% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 25% | Buff | 77% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 24% | Buff | 86% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. | |
| 24% | Nerf | 73% | Disagreement | Random Forest and Gradient Boosting's bootstrap re-run land on different final labels. This doesn't directly contradict the score above, but treat this row as unreliable. |