What is a good Copyscape Result?

[wharbr]

I have used Copyscape sparingly & am not sure just what my Copyscape results indicate. I suscribe to vv spinnable article program and am starting to unspin "unique" articles.
Today I unspun an article, copied & pasted into Copyscape text search.

The Copyscape review says 90 duplicate words were found (out of 536) & I had an aggscore:10.4

Can anyone with Copyscape experience fill me in on how to tell what is an excellant, good or poor result?
How do I determine if the article is good enough to use?
Is an aggscore 10.4 good, bad?

[Jack]

Hey wharbr,

I haven't used copyscape very much. And when I have used it, it was only to see if some articles I bought were unique or copied.

I can tell you that if these articles are from the ultraspinnable articles, you are safe to use them for distributing to blog networks or your own website or anywhere else that you need unique content for Google's eyes. These articles are primarily for uniqueness to Google and they work exceptionally well for that. Google does not require complete 100% uniqueness.

Ezinearticles may not accept them because they will find parts of them on the internet already. The same thing goes for copyscape. It will find parts of them on the internet. As far as a copyscape score, I don't know what is good and what isn't.


Jack

[wharbr]

ok, thanks for the reply. Copyscape does show sentences & phrases already on the internet. Some phrases are things like "January 10th" or "maybe not" & other phrases that might be used in any article.

I have submitted two unspun articles to Ezinearticles. Each have a few sentences & phrases already on the internet. Anxious to see whether they get through or not.

I see comments in threads that their article was 50%-60%-70% changed from the original. How does anyone find the percentage change anyway?

Using what I think is good spinner software & spending a lot of time spinning I get 20%- 30% change at most. I don't think that is nearly good enough & gave up on spinners. I find it faster & easier to rewrite an article & get 100%. The spinnable articles are the only thing I have seen that are close to 100% unique.

[jamesrayenz]

I personally use the copygator.com website to find duplicated content. To me it has a number of benefits over copyscape:

1. it’s automated and brings me results instead of me searching for duplicated content. All i had to do was submit my feed and it started monitoring my feed showing me who’s republished my articles on the web.

2. i get notified by email so it contacts me when it finds copies of my articles online.

3. i use their image badge feature to alert me directly on my website when my content is being lifted.

[Vita_Vee]

Thanks a lot James.

I didn't know copygator, it looks very good.

[Andrew Wilson]

Don't forget that Copyscape is designed to do just one thing: tell you if your content has been plagiarised! You'd expect that Copyscape would flag up spins of the same article, you can't avoid it.

Basically if you are looking to check others palgiarism then you need to make your own determination of what is acceptable to you. As a duplicate content checker it is not much use because that is not its purpose.

BTW, that copygator thing looks useful.

Here's something I have been looking for for a while: http://utext.rikuz.com/en/

Measuring using shingles IS designed to check for duplication across the internet and it is very likley that this and similar tools are used by the large search engines to a greater degree than simply looking at the changes in a document and giving a percentage score which is what we do most of the time.

This uses shingles, a kind of a sliding window that is used to detect content that is similar in little blocks. The idea is that it can pick up content that has been rewritten because there will be words unwritten, similar vocabualry and the information load and structure is going to be similar.

This tool looks like fun and it'd be useful for showing subtle plagiarism but I have no way to calibrate the output for my needs right now.

But for context, I took an article from the Leading Articles Ultra Spinnable Minis about garden sheds and got the following results for two different spins:
Number of words in the shingle - 1. Similarity - 41.29%
Number of words in the shingle - 2. Similarity - 20.05%
Number of words in the shingle - 3. Similarity - 12.31%
Number of words in the shingle - 4. Similarity - 8.76%

Using the same article for the first sample and another article on the same topic, sheds, for the second sample and I got this result:
Number of words in the shingle - 1. Similarity - 17.6%
Number of words in the shingle - 2. Similarity - 1.73%
Number of words in the shingle - 3. Similarity - 0.48%
Number of words in the shingle - 4. Similarity - 0.24%

The more words in the shingle the less degree of similarity we'd expect to see. What I do not know is what is the highest figure acceptable as far as SEs are concerned. Intuitively, I'd suspect the figure is higher than 9% because I can get my spins of VV's articles ranked easily enough even alongside other customers.

I am interested in this becasue I think it is likley to be closer to measuring what SEs do and because I use content from places other than The Leading Articles and, of course, I write and spin my own stuff.

[Vita_Vee]

Hey Andrew

Excellent tool, thanks a lot for sharing the link.

I have played a little bit with it to compare variations of one USA.

The best I could get was these values that are fantastic

Number of words in the shingle - 1. Similarity - 15.64%
Number of words in the shingle - 2. Similarity - 1.74%
Number of words in the shingle - 3. Similarity - 0.84%
Number of words in the shingle - 4. Similarity - 0.42%

The worst I got was:

Number of words in the shingle - 1. Similarity - 21.54%
Number of words in the shingle - 2. Similarity - 4.87%
Number of words in the shingle - 3. Similarity - 3.14%
Number of words in the shingle - 4. Similarity - 2.69%

That would be nice if submission softwares could check the values they get from the generated variations of a spun article before they actually submit them.

[Jack]

The Best Spinner has similar feature. It can spin the number of variations you specify in the background (ten by default) and compare each one to all the others and give the results as percentages in a table. The premiums are usually 90 - 100% unique.

[serenin]

This is the first time I heard of CopyGator and it looks interesting so I’ve give it a try to check my feed with it but it cannot find my website directly so I have to use my feed link as an option. It takes 10 minutes to find my website and another hour to create my feed page result.