Precision
Fact, not hype...
Death2Spam users enjoy better than 99% accuracy.
Real-time capture accuracy data is aggregated from user re-classification and training of all email filtered by Death2Spam.
Here is typical data from one of the Death2Spam servers in our hive:
| Total messages received | 15,931,337 | |
| Messages classified as good | 3,341,843 | 20.98% |
| Messages classified as spam | 12,401,324 | 77.84% |
| Messages not classified (unsure) | 37,156 | 0.23% |
| Unsure email classified as spam (miss) | 8,221 | 0.05% |
| Spam email mis-classified as good (miss) | 4,348 | 0.03% |
| Good email mis-classified as spam (false) | 722 | 0.00% |
| False negative rate = FN / (FN + spam) | 0.100% | |
| False positive rate = FP / (FP + good) | 0.021% | |
| Overall classification accuracy | 99.92% |
Many spam filters advertise accuracy rates of 95%. Consider that the difference between accuracy of 95% and 99% is 4 messages per hundred and 40 messages per thousand...
Every missed spam is not just a nuisance, it's a potential threat.
“For the first time in a very long time, I logged into email today and only got email that I actually wanted. It is amazing that overnight you have eliminated ALL my spam. Top marks for top service.”
JKL Consulting (Nov 2006)