FF-av-test 26-May-2005!

If you want to know how much samples you have to collect when the precision/accuracy level is 1 % and the reliability/confidendence level is 95 %, the result is 9609 randomly picked samples. To increase the accuracy level you have to collect drastically more samples and that is very difficult to an independent tester.

Best regards,
Firefighter!

 

The sample size for on-demand tests that AV-comparatives uses is over 386.000

FF used 2809.
8259 were used by av-comapratives for the retrospective test, which means *only new* samples appeared in the last 3 months.

 

Please, verify those detecting rates of mine to those that AV-Comparatives had in the latest 02-2005 test and you will see that all detecting rates are withing the error marqinal of 2 % against those "Total without DOS & OtherOS", except BitDefender and AntiVir, which increased the defs drastically after the AV-Comparatives test 02-2005.

Best regards,
Firefighter!

 

I saw now what you mean.

 

The most special thanks I want to send to HappyBytes. He said before that my tests were a crap ones, that's why I made this new test one. Besides, my past away father was a good chess player and he was able to beat the best chess player in the state of New Mexixo US some tens of years ago. But he was a poor chess teacher to me when he was in a closed room without a chessboard and he won me when I was at the age of 14. After that I decided that chess is a game to lonely people.

Best regards,
Firefighter!

 

Now my almost 17 year old daughter is beating me in the chess. Damn me with my history.

Best regards,
Firefighter!

 

Well Done FF! I like to see your test results, crap or not. LOL

 

Thanks for sharing the results FF.

Rich

 

Thanks from me too, its always interesting to see the results on these scanners. Any chance you could test out EZAntivirus 7?. I have DrWeb on home computer and NOD on my work box. ETrust just extended my license for me into next summer. Its kinda tempting since DrWeb lapses in a few months but I just don't know about the big EZ...

 

Good job

I-KIRJAIN käsittää mikä te omaisuus luona että , ja i-kirjain hyväksyä. Ainoastaan , haluta aito I-KIRJAIN käytetty kielenkääntäjä jotta kirjake nyt kuluva asia.

 

Firefighter,

Interesting results and thanks for taking the time to run and publish these tests. One small request though - would it be possible to add a note of "unique" detections? (e.g. the number of items that each program detected that none of the others did). This would give some indication of which ones may be useful as a backup scanner to the big hitters.

If these tests are on-demand (therefore testing the AV file-scanners), it may be worth noting that background (memory) scanners should normally give better performance, if any items in your database were compressed/crypted/rebased, etc.

 

Please test latest Norman beta

 

FF,Can you test Panda, please?

 

And AVK 2005.

 

And...... VBA32

 

Norman VC 5.81 beta2 scored only 68.1 % against those 2809 samples, so I didn't want to categorize those detectings anymore.

Best regards,
Firefighter!

 

Tnaks for testing. Just score is ok,but if it's only 68% i'm somehow dissapointed with Norman. On VB100% it has very good score,but judging by this they apperantly focus only on ITW stuff.
I wonder if IBK will test Norman (especially for proactive tests) or is Norman out of the league for good (judging by FF score)...

 

http://www.av-comparatives.org/forum/viewtopic.php?t=173

 

Thank you for the link. Looks like they've done pretty well in retrospective part.
Too bad their overall detection isn't that good (i have noticed that on Jotti too),but i rarely see Norman to detect anything there. And even detected stuff is usually covered by Sandbox and not signatures.
So my conslussion is that their heuristic part is very good,but they certanly lack signatures.

 

Norman IME has always had a low detection rate, but the heuristics were spot on. The thing that gets me with them is the stated Signature DB,

Doesn't seem to make sense, I thought it was just my collection giving it trouble. But I now see that is not the case.

 

LOL, My on demand backup to KAV 5 Real Time.

 

Just made my last correction to the test table in post 1. when 110 more samples were removed by advice of IBK. The reliability/confidence level dropped a bit to 95.889 % with the same precision/accuracy level of 2 % (see my post 5. in this thread now).

Best regards,
Firefighter!