Remove failed requests from benchmark tables

README.md

@@ -46,38 +46,38 @@ Failed requests count as incorrect decisions: a failed issue is a miss; a failed
 
 ### Public — 2,220 pairs
 
-| Model | Recall | False-positive rate | Youden’s J | Failed requests | Inference cost |
-| --- | ---: | ---: | ---: | ---: | ---: |
-| [GPT-6 Luna Decisions](https://developers.openai.com/api/docs/guides/decisions) | 48.7% | 38.7% | **10.0%** | **0** | $1.102 |
-| [DiffusionGemma Jev](https://github.com/razorback16/openjev) | 33.4% | 27.1% | 6.3% | **0** | $0.239 |
-| [Microsoft-Decision-1](https://openrouter.ai/microsoft/microsoft-decision-1) | 24.0% | 18.8% | 5.1% | **0** | $0.481 |
-| [Jev](https://typesafe.ai/blog/introducing-system-one-models-and-jev) | 17.7% | 12.7% | 5.0% | **0** | $0.551 |
-| [SemiF](https://github.com/TheoLeeCJ/SemIf) | 43.0% | 39.7% | 3.3% | **0** | $0.239 |
-| [Mercury Decide](https://openrouter.ai/inception/mercury-decide) | 8.8% | 5.6% | 3.2% | 1 | $0.224 |
-| [TypeLLM](https://typellm.ai/docs) | 8.3% | 5.5% | 2.7% | **0** | $0.576 |
-| [Span-01](https://www.respan.ai/docs/documentation/span-01/concept) | 25.3% | 22.6% | 2.7% | **0** | $0.182 |
-| [Liquid d1](https://openrouter.ai/liquid/d1) | 5.4% | 3.6% | 1.8% | 1 | $0.875 |
-| [CodeQL](https://codeql.github.com/)* | 5.3% | 4.2% | 1.1% | — | — |
-| [CLM-v0.1-8B](https://huggingface.co/Contrastive-LM/CLM-v0.1-8B) | **81.9%** | 80.9% | 1.0% | **0** | ~$0.210 |
-| [GLiNER-2.5-Decide](https://huggingface.co/fastino/GLiNER2.5-Decide) | 2.3% | **2.1%** | 0.2% | **0** | **$0.175** |
+| Model | Recall | False-positive rate | Youden’s J | Inference cost |
+| --- | ---: | ---: | ---: | ---: |
+| [GPT-6 Luna Decisions](https://developers.openai.com/api/docs/guides/decisions) | 48.7% | 38.7% | **10.0%** | $1.102 |
+| [DiffusionGemma Jev](https://github.com/razorback16/openjev) | 33.4% | 27.1% | 6.3% | $0.239 |
+| [Microsoft-Decision-1](https://openrouter.ai/microsoft/microsoft-decision-1) | 24.0% | 18.8% | 5.1% | $0.481 |
+| [Jev](https://typesafe.ai/blog/introducing-system-one-models-and-jev) | 17.7% | 12.7% | 5.0% | $0.551 |
+| [SemiF](https://github.com/TheoLeeCJ/SemIf) | 43.0% | 39.7% | 3.3% | $0.239 |
+| [Mercury Decide](https://openrouter.ai/inception/mercury-decide) | 8.8% | 5.6% | 3.2% | $0.224 |
+| [TypeLLM](https://typellm.ai/docs) | 8.3% | 5.5% | 2.7% | $0.576 |
+| [Span-01](https://www.respan.ai/docs/documentation/span-01/concept) | 25.3% | 22.6% | 2.7% | $0.182 |
+| [Liquid d1](https://openrouter.ai/liquid/d1) | 5.4% | 3.6% | 1.8% | $0.875 |
+| [CodeQL](https://codeql.github.com/)* | 5.3% | 4.2% | 1.1% | — |
+| [CLM-v0.1-8B](https://huggingface.co/Contrastive-LM/CLM-v0.1-8B) | **81.9%** | 80.9% | 1.0% | ~$0.210 |
+| [GLiNER-2.5-Decide](https://huggingface.co/fastino/GLiNER2.5-Decide) | 2.3% | **2.1%** | 0.2% | **$0.175** |
 
 \* CodeQL scored 682 pairs in Python, JavaScript, TypeScript, C and C++; 255 additional pairs could not be scored because source recovery, source verification or analysis failed. [Details](CODEQL.md) · [Method decisions](codeql-comparison/)
 
 ### Private — 498 pairs
 
-| Model | Recall | False-positive rate | Youden’s J | Failed requests | Inference cost |
-| --- | ---: | ---: | ---: | ---: | ---: |
-| [GPT-6 Luna Decisions](https://developers.openai.com/api/docs/guides/decisions) | 52.4% | 41.3% | **11.1%** | **0** | $0.247 |
-| [DiffusionGemma Jev](https://github.com/razorback16/openjev) | 35.8% | 27.3% | 8.6% | **0** | $0.054 |
-| [TypeLLM](https://typellm.ai/docs) | 11.3% | 5.8% | 5.6% | **0** | $0.129 |
-| [Microsoft-Decision-1](https://openrouter.ai/microsoft/microsoft-decision-1) | 25.5% | 20.6% | 4.8% | **0** | $0.108 |
-| [Jev](https://typesafe.ai/blog/introducing-system-one-models-and-jev) | 20.0% | 15.6% | 4.5% | **0** | $0.123 |
-| [Mercury Decide](https://openrouter.ai/inception/mercury-decide) | 9.4% | 5.1% | 4.4% | **0** | $0.050 |
-| [Span-01](https://www.respan.ai/docs/documentation/span-01/concept) | 28.5% | 24.3% | 4.2% | **0** | $0.040 |
-| [SemiF](https://github.com/TheoLeeCJ/SemIf) | 44.3% | 41.8% | 2.6% | **0** | $0.054 |
-| [Liquid d1](https://openrouter.ai/liquid/d1) | 6.6% | 4.2% | 2.4% | **0** | $0.197 |
-| [GLiNER-2.5-Decide](https://huggingface.co/fastino/GLiNER2.5-Decide) | 1.7% | **2.1%** | -0.4% | **0** | **$0.039** |
-| [CLM-v0.1-8B](https://huggingface.co/Contrastive-LM/CLM-v0.1-8B) | **79.5%** | 80.1% | -0.6% | **0** | ~$0.047 |
+| Model | Recall | False-positive rate | Youden’s J | Inference cost |
+| --- | ---: | ---: | ---: | ---: |
+| [GPT-6 Luna Decisions](https://developers.openai.com/api/docs/guides/decisions) | 52.4% | 41.3% | **11.1%** | $0.247 |
+| [DiffusionGemma Jev](https://github.com/razorback16/openjev) | 35.8% | 27.3% | 8.6% | $0.054 |
+| [TypeLLM](https://typellm.ai/docs) | 11.3% | 5.8% | 5.6% | $0.129 |
+| [Microsoft-Decision-1](https://openrouter.ai/microsoft/microsoft-decision-1) | 25.5% | 20.6% | 4.8% | $0.108 |
+| [Jev](https://typesafe.ai/blog/introducing-system-one-models-and-jev) | 20.0% | 15.6% | 4.5% | $0.123 |
+| [Mercury Decide](https://openrouter.ai/inception/mercury-decide) | 9.4% | 5.1% | 4.4% | $0.050 |
+| [Span-01](https://www.respan.ai/docs/documentation/span-01/concept) | 28.5% | 24.3% | 4.2% | $0.040 |
+| [SemiF](https://github.com/TheoLeeCJ/SemIf) | 44.3% | 41.8% | 2.6% | $0.054 |
+| [Liquid d1](https://openrouter.ai/liquid/d1) | 6.6% | 4.2% | 2.4% | $0.197 |
+| [GLiNER-2.5-Decide](https://huggingface.co/fastino/GLiNER2.5-Decide) | 1.7% | **2.1%** | -0.4% | **$0.039** |
+| [CLM-v0.1-8B](https://huggingface.co/Contrastive-LM/CLM-v0.1-8B) | **79.5%** | 80.1% | -0.6% | ~$0.047 |
 
 ## Security
 
@@ -87,38 +87,38 @@ Failed requests count as incorrect decisions: a failed issue is a miss; a failed
 
 ### Public — 1,311 pairs
 
-| Model | Recall | False-positive rate | Youden’s J | Failed requests | Inference cost |
-| --- | ---: | ---: | ---: | ---: | ---: |
-| [Liquid d1](https://openrouter.ai/liquid/d1) | 52.3% | 40.7% | **11.5%** | **0** | $6.132 |
-| [Jev](https://typesafe.ai/blog/introducing-system-one-models-and-jev) | 34.6% | 23.6% | 11.0% | **0** | $0.507 |
-| [TypeLLM](https://typellm.ai/docs) | 27.8% | 17.4% | 10.5% | **0** | $0.566 |
-| [GPT-6 Luna Decisions](https://developers.openai.com/api/docs/guides/decisions) | 54.1% | 44.5% | 9.7% | **0** | $1.469 |
-| [Mercury Decide](https://openrouter.ai/inception/mercury-decide) | 27.1% | 17.8% | 9.3% | **0** | **$0.131** |
-| [DiffusionGemma Jev](https://github.com/razorback16/openjev) | 33.1% | 25.8% | 7.3% | **0** | $0.208 |
-| [Microsoft-Decision-1](https://openrouter.ai/microsoft/microsoft-decision-1) | 43.9% | 38.2% | 5.6% | **0** | $0.455 |
-| [Span-01](https://www.respan.ai/docs/documentation/span-01/concept) | 21.0% | 17.2% | 3.9% | **0** | $0.251 |
-| [CodeQL](https://codeql.github.com/)* | 4.8% | 1.5% | 3.3% | — | — |
-| [GLiNER-2.5-Decide](https://huggingface.co/fastino/GLiNER2.5-Decide) | 2.1% | **1.6%** | 0.5% | 6 | $0.434 |
-| [SemiF](https://github.com/TheoLeeCJ/SemIf) | **100.0%** | 99.9% | 0.1% | **0** | $0.208 |
-| [CLM-v0.1-8B](https://huggingface.co/Contrastive-LM/CLM-v0.1-8B) | **100.0%** | 100.0% | 0.0% | **0** | ~$1.432 |
+| Model | Recall | False-positive rate | Youden’s J | Inference cost |
+| --- | ---: | ---: | ---: | ---: |
+| [Liquid d1](https://openrouter.ai/liquid/d1) | 52.3% | 40.7% | **11.5%** | $6.132 |
+| [Jev](https://typesafe.ai/blog/introducing-system-one-models-and-jev) | 34.6% | 23.6% | 11.0% | $0.507 |
+| [TypeLLM](https://typellm.ai/docs) | 27.8% | 17.4% | 10.5% | $0.566 |
+| [GPT-6 Luna Decisions](https://developers.openai.com/api/docs/guides/decisions) | 54.1% | 44.5% | 9.7% | $1.469 |
+| [Mercury Decide](https://openrouter.ai/inception/mercury-decide) | 27.1% | 17.8% | 9.3% | **$0.131** |
+| [DiffusionGemma Jev](https://github.com/razorback16/openjev) | 33.1% | 25.8% | 7.3% | $0.208 |
+| [Microsoft-Decision-1](https://openrouter.ai/microsoft/microsoft-decision-1) | 43.9% | 38.2% | 5.6% | $0.455 |
+| [Span-01](https://www.respan.ai/docs/documentation/span-01/concept) | 21.0% | 17.2% | 3.9% | $0.251 |
+| [CodeQL](https://codeql.github.com/)* | 4.8% | 1.5% | 3.3% | — |
+| [GLiNER-2.5-Decide](https://huggingface.co/fastino/GLiNER2.5-Decide) | 2.1% | **1.6%** | 0.5% | $0.434 |
+| [SemiF](https://github.com/TheoLeeCJ/SemIf) | **100.0%** | 99.9% | 0.1% | $0.208 |
+| [CLM-v0.1-8B](https://huggingface.co/Contrastive-LM/CLM-v0.1-8B) | **100.0%** | 100.0% | 0.0% | ~$1.432 |
 
 \* CodeQL scored 305 pairs in Python, JavaScript, TypeScript, C and C++; 18 additional pairs could not be scored because source recovery, source verification or analysis failed. [Details](CODEQL.md) · [Method decisions](codeql-comparison/)
 
 ### Private — 439 pairs
 
-| Model | Recall | False-positive rate | Youden’s J | Failed requests | Inference cost |
-| --- | ---: | ---: | ---: | ---: | ---: |
-| [GPT-6 Luna Decisions](https://developers.openai.com/api/docs/guides/decisions) | 40.5% | 31.5% | **9.0%** | **0** | $0.530 |
-| [Liquid d1](https://openrouter.ai/liquid/d1) | 34.7% | 29.4% | 5.3% | **0** | $2.362 |
-| [Jev](https://typesafe.ai/blog/introducing-system-one-models-and-jev) | 20.6% | 16.7% | 3.9% | **0** | $0.182 |
-| [DiffusionGemma Jev](https://github.com/razorback16/openjev) | 21.9% | 18.5% | 3.3% | **0** | $0.074 |
-| [TypeLLM](https://typellm.ai/docs) | 8.8% | 6.2% | 2.5% | **0** | $0.203 |
-| [Mercury Decide](https://openrouter.ai/inception/mercury-decide) | 11.1% | 9.0% | 2.1% | **0** | **$0.047** |
-| [Microsoft-Decision-1](https://openrouter.ai/microsoft/microsoft-decision-1) | 18.1% | 16.7% | 1.4% | **0** | $0.163 |
-| [GLiNER-2.5-Decide](https://huggingface.co/fastino/GLiNER2.5-Decide) | 2.0% | **1.5%** | 0.6% | 1 | $0.142 |
-| [CLM-v0.1-8B](https://huggingface.co/Contrastive-LM/CLM-v0.1-8B) | **100.0%** | 100.0% | 0.0% | **0** | ~$0.528 |
-| [SemiF](https://github.com/TheoLeeCJ/SemIf) | **100.0%** | 100.0% | 0.0% | **0** | $0.074 |
-| [Span-01](https://www.respan.ai/docs/documentation/span-01/concept) | 8.0% | 8.1% | -0.1% | **0** | $0.091 |
+| Model | Recall | False-positive rate | Youden’s J | Inference cost |
+| --- | ---: | ---: | ---: | ---: |
+| [GPT-6 Luna Decisions](https://developers.openai.com/api/docs/guides/decisions) | 40.5% | 31.5% | **9.0%** | $0.530 |
+| [Liquid d1](https://openrouter.ai/liquid/d1) | 34.7% | 29.4% | 5.3% | $2.362 |
+| [Jev](https://typesafe.ai/blog/introducing-system-one-models-and-jev) | 20.6% | 16.7% | 3.9% | $0.182 |
+| [DiffusionGemma Jev](https://github.com/razorback16/openjev) | 21.9% | 18.5% | 3.3% | $0.074 |
+| [TypeLLM](https://typellm.ai/docs) | 8.8% | 6.2% | 2.5% | $0.203 |
+| [Mercury Decide](https://openrouter.ai/inception/mercury-decide) | 11.1% | 9.0% | 2.1% | **$0.047** |
+| [Microsoft-Decision-1](https://openrouter.ai/microsoft/microsoft-decision-1) | 18.1% | 16.7% | 1.4% | $0.163 |
+| [GLiNER-2.5-Decide](https://huggingface.co/fastino/GLiNER2.5-Decide) | 2.0% | **1.5%** | 0.6% | $0.142 |
+| [CLM-v0.1-8B](https://huggingface.co/Contrastive-LM/CLM-v0.1-8B) | **100.0%** | 100.0% | 0.0% | ~$0.528 |
+| [SemiF](https://github.com/TheoLeeCJ/SemIf) | **100.0%** | 100.0% | 0.0% | $0.074 |
+| [Span-01](https://www.respan.ai/docs/documentation/span-01/concept) | 8.0% | 8.1% | -0.1% | $0.091 |
 
 ## Perch Lint Bench
 
@@ -128,35 +128,35 @@ Failed requests count as incorrect decisions: a failed issue is a miss; a failed
 
 ### Public — 214 pairs
 
-| Model | Recall | False-positive rate | Youden’s J | Failed requests | Inference cost |
-| --- | ---: | ---: | ---: | ---: | ---: |
-| [Mercury Decide](https://openrouter.ai/inception/mercury-decide) | **81.8%** | 6.5% | **75.2%** | **0** | $0.020 |
-| [GPT-6 Luna Decisions](https://developers.openai.com/api/docs/guides/decisions) | 76.6% | 4.7% | 72.0% | **0** | $0.093 |
-| [TypeLLM](https://typellm.ai/docs) | 61.7% | 2.8% | 58.9% | **0** | $0.050 |
-| [Jev](https://typesafe.ai/blog/introducing-system-one-models-and-jev) | 50.0% | 0.5% | 49.5% | **0** | $0.048 |
-| [DiffusionGemma Jev](https://github.com/razorback16/openjev) | 52.8% | 5.1% | 47.7% | **0** | $0.018 |
-| [Liquid d1](https://openrouter.ai/liquid/d1) | 40.7% | 0.5% | 40.2% | **0** | $0.039 |
-| [Span-01](https://www.respan.ai/docs/documentation/span-01/concept) | 50.0% | 11.2% | 38.8% | **0** | $0.014 |
-| [Microsoft-Decision-1](https://openrouter.ai/microsoft/microsoft-decision-1) | 38.3% | 1.9% | 36.4% | **0** | $0.041 |
-| [CLM-v0.1-8B](https://huggingface.co/Contrastive-LM/CLM-v0.1-8B) | 38.8% | 37.4% | 1.4% | **0** | **~$0.010** |
-| [SemiF](https://github.com/TheoLeeCJ/SemIf) | 32.2% | 31.3% | 0.9% | **0** | $0.018 |
-| [GLiNER-2.5-Decide](https://huggingface.co/fastino/GLiNER2.5-Decide) | 0.0% | **0.0%** | 0.0% | **0** | **$0.010** |
+| Model | Recall | False-positive rate | Youden’s J | Inference cost |
+| --- | ---: | ---: | ---: | ---: |
+| [Mercury Decide](https://openrouter.ai/inception/mercury-decide) | **81.8%** | 6.5% | **75.2%** | $0.020 |
+| [GPT-6 Luna Decisions](https://developers.openai.com/api/docs/guides/decisions) | 76.6% | 4.7% | 72.0% | $0.093 |
+| [TypeLLM](https://typellm.ai/docs) | 61.7% | 2.8% | 58.9% | $0.050 |
+| [Jev](https://typesafe.ai/blog/introducing-system-one-models-and-jev) | 50.0% | 0.5% | 49.5% | $0.048 |
+| [DiffusionGemma Jev](https://github.com/razorback16/openjev) | 52.8% | 5.1% | 47.7% | $0.018 |
+| [Liquid d1](https://openrouter.ai/liquid/d1) | 40.7% | 0.5% | 40.2% | $0.039 |
+| [Span-01](https://www.respan.ai/docs/documentation/span-01/concept) | 50.0% | 11.2% | 38.8% | $0.014 |
+| [Microsoft-Decision-1](https://openrouter.ai/microsoft/microsoft-decision-1) | 38.3% | 1.9% | 36.4% | $0.041 |
+| [CLM-v0.1-8B](https://huggingface.co/Contrastive-LM/CLM-v0.1-8B) | 38.8% | 37.4% | 1.4% | **~$0.010** |
+| [SemiF](https://github.com/TheoLeeCJ/SemIf) | 32.2% | 31.3% | 0.9% | $0.018 |
+| [GLiNER-2.5-Decide](https://huggingface.co/fastino/GLiNER2.5-Decide) | 0.0% | **0.0%** | 0.0% | **$0.010** |
 
 ### Private — 71 pairs
 
-| Model | Recall | False-positive rate | Youden’s J | Failed requests | Inference cost |
-| --- | ---: | ---: | ---: | ---: | ---: |
-| [Mercury Decide](https://openrouter.ai/inception/mercury-decide) | **87.3%** | 8.5% | **78.9%** | **0** | $0.006 |
-| [GPT-6 Luna Decisions](https://developers.openai.com/api/docs/guides/decisions) | 84.5% | 7.0% | 77.5% | **0** | $0.030 |
-| [TypeLLM](https://typellm.ai/docs) | 59.2% | 5.6% | 53.5% | **0** | $0.016 |
-| [Jev](https://typesafe.ai/blog/introducing-system-one-models-and-jev) | 53.5% | 1.4% | 52.1% | **0** | $0.016 |
-| [DiffusionGemma Jev](https://github.com/razorback16/openjev) | 53.5% | 8.5% | 45.1% | **0** | $0.006 |
-| [Microsoft-Decision-1](https://openrouter.ai/microsoft/microsoft-decision-1) | 43.7% | 1.4% | 42.3% | **0** | $0.013 |
-| [Liquid d1](https://openrouter.ai/liquid/d1) | 40.8% | 1.4% | 39.4% | **0** | $0.012 |
-| [Span-01](https://www.respan.ai/docs/documentation/span-01/concept) | 45.1% | 8.5% | 36.6% | **0** | $0.005 |
-| [CLM-v0.1-8B](https://huggingface.co/Contrastive-LM/CLM-v0.1-8B) | 50.7% | 49.3% | 1.4% | **0** | **~$0.003** |
-| [GLiNER-2.5-Decide](https://huggingface.co/fastino/GLiNER2.5-Decide) | 0.0% | **0.0%** | 0.0% | **0** | $0.004 |
-| [SemiF](https://github.com/TheoLeeCJ/SemIf) | 42.3% | 45.1% | -2.8% | **0** | $0.006 |
+| Model | Recall | False-positive rate | Youden’s J | Inference cost |
+| --- | ---: | ---: | ---: | ---: |
+| [Mercury Decide](https://openrouter.ai/inception/mercury-decide) | **87.3%** | 8.5% | **78.9%** | $0.006 |
+| [GPT-6 Luna Decisions](https://developers.openai.com/api/docs/guides/decisions) | 84.5% | 7.0% | 77.5% | $0.030 |
+| [TypeLLM](https://typellm.ai/docs) | 59.2% | 5.6% | 53.5% | $0.016 |
+| [Jev](https://typesafe.ai/blog/introducing-system-one-models-and-jev) | 53.5% | 1.4% | 52.1% | $0.016 |
+| [DiffusionGemma Jev](https://github.com/razorback16/openjev) | 53.5% | 8.5% | 45.1% | $0.006 |
+| [Microsoft-Decision-1](https://openrouter.ai/microsoft/microsoft-decision-1) | 43.7% | 1.4% | 42.3% | $0.013 |
+| [Liquid d1](https://openrouter.ai/liquid/d1) | 40.8% | 1.4% | 39.4% | $0.012 |
+| [Span-01](https://www.respan.ai/docs/documentation/span-01/concept) | 45.1% | 8.5% | 36.6% | $0.005 |
+| [CLM-v0.1-8B](https://huggingface.co/Contrastive-LM/CLM-v0.1-8B) | 50.7% | 49.3% | 1.4% | **~$0.003** |
+| [GLiNER-2.5-Decide](https://huggingface.co/fastino/GLiNER2.5-Decide) | 0.0% | **0.0%** | 0.0% | $0.004 |
+| [SemiF](https://github.com/TheoLeeCJ/SemIf) | 42.3% | 45.1% | -2.8% | $0.006 |
 
 <!-- END TRADEOFF_RESULTS -->
 

test_tradeoff.py

@@ -19,7 +19,7 @@ class GraphTests(unittest.TestCase):
             return ET.fromstring(output.read_text())
 
     def dots(self, root):
-        return {e.attrib['data-model']: e for e in root.findall('{http://www.w3.org/2000/svg}circle')
+        return {e.attrib['data-model']: e for e in root.iter()
                 if 'data-model' in e.attrib}
 
     def test_points_use_recorded_cost_and_youden_j(self):
@@ -30,9 +30,9 @@ class GraphTests(unittest.TestCase):
         self.assertAlmostEqual(float(dot.attrib['data-j']), .6)
         low,high=float(root.attrib['data-cost-low']),float(root.attrib['data-cost-high'])
         y_min,y_max=float(root.attrib['data-j-min']),float(root.attrib['data-j-max'])
-        self.assertAlmostEqual((float(dot.attrib['cx'])-100)/1130,
+        self.assertAlmostEqual((float(dot.attrib['data-x'])-100)/1130,
             (math.log10(.05)-math.log10(low))/(math.log10(high)-math.log10(low)),places=4)
-        self.assertAlmostEqual((float(dot.attrib['cy'])-175)/400,(y_max-60)/(y_max-y_min),places=4)
+        self.assertAlmostEqual((float(dot.attrib['data-y'])-175)/400,(y_max-60)/(y_max-y_min),places=4)
         labels=[e.text for e in root.findall('{http://www.w3.org/2000/svg}text')]
         self.assertIn('Run cost (USD, log scale)', labels)
         self.assertIn('Youden’s J (recall − false-positive rate)', labels)
@@ -40,8 +40,8 @@ class GraphTests(unittest.TestCase):
     def test_more_false_positives_reduce_j_at_same_cost(self):
         current=self.rows();current['typellm']['confusion']={'tp':7,'fn':3,'fp':10,'tn':10}
         dots=self.dots(self.render(current))
-        self.assertEqual(dots['jev'].attrib['cx'],dots['typellm'].attrib['cx'])
-        self.assertLess(float(dots['jev'].attrib['cy']),float(dots['typellm'].attrib['cy']))
+        self.assertEqual(dots['jev'].attrib['data-x'],dots['typellm'].attrib['data-x'])
+        self.assertLess(float(dots['jev'].attrib['data-y']),float(dots['typellm'].attrib['data-y']))
 
     def test_frontier_excludes_strictly_dominated_models_and_keeps_ties(self):
         def row(cost,tp,fp):

tradeoff.py

@@ -11,7 +11,7 @@ import argparse
 from collections import defaultdict
 import html
 import json
-from math import ceil, floor, isfinite, log10
+from math import ceil, cos, floor, isfinite, log10, pi, sin
 from pathlib import Path
 from tempfile import TemporaryDirectory
 from urllib.request import urlopen
@@ -36,14 +36,20 @@ NAMES = {
     "mercury-decide": ("Mercury Decide", "https://openrouter.ai/inception/mercury-decide"),
     "openai-decisions": ("GPT-6 Luna Decisions", "https://developers.openai.com/api/docs/guides/decisions"),
 }
+# Shapes distinguish models independently of color, including in grayscale.
 COLORS = {
-    "microsoft-decision-1": "#4aa4ef",
-    "mercury-decide": "#b2bf3c",
-    "openai-decisions": "#36b894",
-    "jev": "#1976d2", "typellm": "#c64f81", "liquid-d1": "#7638d6",
-    "diffusiongemma": "#e08a22", "span01": "#168d80", "clm": "#677586",
-    "semif": "#b14b39",
-    "gliner25-decide": "#9b7a39",
+    "jev": "#56b4e9", "typellm": "#ff63b5", "liquid-d1": "#b98aff",
+    "diffusiongemma": "#ff9234", "span01": "#00c49a", "clm": "#a0a8b5",
+    "semif": "#ef5b5b", "gliner25-decide": "#ffe05b",
+    "microsoft-decision-1": "#f5f5f5", "mercury-decide": "#6e8fff",
+    "openai-decisions": "#b8e186",
+}
+MARKERS = {
+    "jev": "circle", "typellm": "diamond", "liquid-d1": "triangle",
+    "diffusiongemma": "square", "span01": "triangle-down", "clm": "hexagon",
+    "semif": "cross", "gliner25-decide": "pentagon",
+    "microsoft-decision-1": "plus", "mercury-decide": "star",
+    "openai-decisions": "ring",
 }
 ORDER = tuple(NAMES)
 OLD_PACK_SHA = "bb53c8a6ca30920140e2bf33b2905d159e7b1709b29d1ea896637905dbf9047d"
@@ -186,15 +192,14 @@ def results_table(rows: list[dict]) -> str:
     best_recall = max(round(metrics[row["model_id"]][0] * 100, 1) for row in full_rows)
     best_false_positive_rate = min(round(metrics[row["model_id"]][1] * 100, 1) for row in full_rows)
     best_net_recall = max(round(metrics[row["model_id"]][2] * 100, 1) for row in full_rows)
-    best_failures = min(row.get("failed_methods", 0) for row in full_rows)
     priced_costs = [round(row["inference_cost_usd"], 3) for row in full_rows if row.get("inference_cost_usd") is not None]
     best_priced_cost = min(priced_costs) if priced_costs else None
 
     def mark(value: str, winner: bool) -> str:
         return f"**{value}**" if winner else value
 
-    table = ["| Model | Recall | False-positive rate | Youden’s J | Failed requests | Inference cost |",
-             "| --- | ---: | ---: | ---: | ---: | ---: |"]
+    table = ["| Model | Recall | False-positive rate | Youden’s J | Inference cost |",
+             "| --- | ---: | ---: | ---: | ---: |"]
     for row in ordered:
         recall, false_positive_rate, j = metrics[row["model_id"]]
         cost = row.get("inference_cost_usd")
@@ -207,7 +212,6 @@ def results_table(rows: list[dict]) -> str:
             mark(percent(recall), not is_codeql and round(recall * 100, 1) == best_recall),
             mark(percent(false_positive_rate), not is_codeql and round(false_positive_rate * 100, 1) == best_false_positive_rate),
             mark(percent(j), not is_codeql and round(j * 100, 1) == best_net_recall),
-            "—" if is_codeql else mark(str(row.get("failed_methods", 0)), row.get("failed_methods", 0) == best_failures),
             "—" if is_codeql else mark(money(row), cost is not None and round(cost, 3) == best_priced_cost),
         )) + " |")
     return "\n".join(table)
@@ -257,6 +261,34 @@ def plot_axes(current: dict[str, dict]) -> tuple[float, float, list[float], floa
     return low, high, ticks, bottom, top, step
 
 
+def marker(model: str, x: float, y: float, size: float = 7,
+           attributes: str = "", title: str = "") -> str:
+    shape, color = MARKERS[model], COLORS[model]
+    style = f'fill="{color}" stroke="#101318" stroke-width="1.2" stroke-linejoin="round"'
+    if shape == "circle":
+        element = f'<circle r="{size}" {style}/>'
+    elif shape == "ring":
+        element = f'<circle r="{size-1}" fill="#101318" stroke="{color}" stroke-width="3"/>'
+    elif shape == "square":
+        element = f'<rect x="{-size}" y="{-size}" width="{size*2}" height="{size*2}" {style}/>'
+    elif shape in ("plus", "cross"):
+        thickness = size * .35
+        path = f'M {-thickness} {-size} H {thickness} V {-thickness} H {size} V {thickness} H {thickness} V {size} H {-thickness} V {thickness} H {-size} V {-thickness} H {-thickness} Z'
+        element = f'<path d="{path}" {style}' + (' transform="rotate(45)"' if shape == "cross" else '') + '/>'
+    else:
+        sides = {"diamond": 4, "triangle": 3, "triangle-down": 3,
+                 "hexagon": 6, "pentagon": 5, "star": 10}[shape]
+        angle = pi/2 if shape == "triangle-down" else -pi/2
+        points = []
+        for i in range(sides):
+            radius = size * (.45 if shape == "star" and i % 2 else 1.15)
+            theta = angle + i * 2*pi/sides
+            points.append(f'{radius*cos(theta):.2f},{radius*sin(theta):.2f}')
+        element = f'<polygon points="{" ".join(points)}" {style}/>'
+    tooltip = f'<title>{html.escape(title)}</title>' if title else ''
+    return f'<g transform="translate({x:.2f} {y:.2f})" data-marker="{shape}" {attributes}>{tooltip}{element}</g>'
+
+
 def svg(domain: str, current: dict[str, dict], output: Path) -> None:
     width, height = 1280, 690
     left, right, top, bottom = 100, 50, 175, 115
@@ -282,7 +314,7 @@ def svg(domain: str, current: dict[str, dict], output: Path) -> None:
     for i, model in enumerate(models):
         col, row = i % 4, i // 4
         lx, ly = left + col * 280, 102 + row * 23
-        out += [f'<circle cx="{lx + 8}" cy="{ly}" r="5" fill="{COLORS[model]}"/>',
+        out += [marker(model, lx + 8, ly, size=6),
                 f'<text x="{lx + 24}" y="{ly + 5}" fill="#e1e7ef" font-family="Arial,sans-serif" font-size="15">{html.escape(NAMES[model][0])}</text>']
     for value in cost_ticks:
         px = x(value)
@@ -303,10 +335,11 @@ def svg(domain: str, current: dict[str, dict], output: Path) -> None:
         row = current[model]; cost = plot_cost(row)
         recall, fpr, score = operating_metrics(row)
         px, py = x(cost), y(score)
-        tooltip = html.escape(f'{NAMES[model][0]}: Youden’s J {score:.1%}; recall {recall:.1%}; false-positive rate {fpr:.1%}; run cost ${cost:.4f}' + (' (GPU estimate)' if model == 'clm' else ''))
+        tooltip = f'{NAMES[model][0]}: Youden’s J {score:.1%}; recall {recall:.1%}; false-positive rate {fpr:.1%}; run cost ${cost:.4f}' + (' (GPU estimate)' if model == 'clm' else '')
         if model in frontier:
-            out.append(f'<circle cx="{px:.2f}" cy="{py:.2f}" r="11" fill="#101318" stroke="#ffb454" stroke-width="2"/>')
-        out.append(f'<circle data-model="{model}" data-cost="{cost}" data-j="{score}" data-frontier="{str(model in frontier).lower()}" cx="{px:.2f}" cy="{py:.2f}" r="6" fill="{COLORS[model]}" stroke="#fff" stroke-width="1.5"><title>{tooltip}</title></circle>')
+            out.append(f'<circle cx="{px:.2f}" cy="{py:.2f}" r="13" fill="#101318" stroke="#ffb454" stroke-width="2"/>')
+        attributes = f'data-model="{model}" data-cost="{cost}" data-j="{score}" data-frontier="{str(model in frontier).lower()}" data-x="{px:.2f}" data-y="{py:.2f}"'
+        out.append(marker(model, px, py, attributes=attributes, title=tooltip))
     footer = 'CLM cost: GPU runtime estimate'
     if 'codeql' in current:
         footer += ' · CodeQL cost: unrecorded'