JL Forums

TBS3: suggested changes to the "Uniqueness" logic

0 Members and 1 Guest are viewing this topic.

TBS3: suggested changes to the "Uniqueness" logic
« on: February 25, 2017, 07:24:12 PM »
After going through the official tutorial videos, and the "Advanced Spinning techniques" by Matthew Woodward, i have some suggestions.

The advice given in the tutorials is to do sentence spinning first, and follow that with word/phrase spinning.

I have noticed that if i do sentence spinning, the "uniqueness" seems low, which is understandable, however word/phrase spins appear to be given higher weighting when added to the original sentence.

for example:
Code: [Select]
{to be or not to be|not to be or to be}
can become:
{{to be|to become} {or not|or otherwise} {to be|to get}|not to be or to be}
Unique: 6%
vs
{to be or not to be|{not to|never to} be {or to|in order to} be}
Unique: 4%

so it seems that spinning the sentence then including synonyms in the second sentence have a lower uniqueness than spinning words in the first sentence and including synonyms in the second sentence.

In appreciation of the "do sentence spinning first" logic, it would seem to be more sensible for the software to pre-select the rewritten/spun sentence as a priority therefore giving a higher percent that the article is unique in its new sentence structure, and then word/phrase spinning would not need to carry as much importance, for example.
"European travellers and geologists"
"Geologists and travellers from Europe"

as we can see here is that its not entirely practical to change the location or the job titles, but the sentence is indeed rewritten/spun, so this should yield a high uniqueness, but actually it is only 5% unique.

the reason for my suggestion is that i have spun a rather lengthy article, and even though i have completely replaced many words with their synonym in the second sentence, the uniqueness is low, but if i add synonym words and phrases to the first sentence then i am given a much higher uniqueness.

i proved this after trying to raise uniqueness in the article i have done from about 31%.. but after removing the variable choices from the first sentence and only left the completely reworded second sentence as a choice i noticed the uniqueness drop to 18%.

now while the current spinning uniqueness algorithm may be suitable for writing multiple articles, whereby the original sentence structure may be retained in its entirety or even word/phrase synonyms are used in it, it is not so appropriate when a singe article is required to be completely rewritten as to appear new.

i am aware that there is a feature for "replace" and this is very good for exchanging words for a single article, it does not allow the user to accurately calculate that their article, even after being spun, is completely unique to other articles on the internet.

i guess the reason for my suggestion is that currently im having issue with the TBS Check where it errors out when trying to check, so to have the algorithm tailored that when a single article is spun by preselecting a synonym sentence as a priority and be able to correctly identify within the software that it is indeed unique would give a greater sense of satisfaction and security despite not being able to complete the online check.

Offline Meg

  • *****
  • 7050
  • Freedup to do what you really want
    • View Profile
Re: TBS3: suggested changes to the "Uniqueness" logic
« Reply #1 on: February 25, 2017, 11:39:59 PM »
A member of this forum, SNM, gave a formula, several years ago, for calculating how many variations you could get on an article. The term "uniqueness" does NOT refer to whether the article has similar variations on the internet but to how many variations you could get, when your article was spun out.
For instance, "Mary had a little lamb" and "Mary had a little pig" would be considered "unique" because they are different. The human brain sees them as similar because we remember our nursery rhymes. SNM pointed out that the number of "unique" variations was related to the number of synonyms you included. So if you had a sentence with 4 words and provided 3 synonyms for each word, you could get 3 X 3 X 3 X3 variations = 81 different variations. TBS would provide you with each of those variations because each would be "unique".
In terms of rewriting an article, I personally start at paragraph level and work down, to sentence, then word and phrase level simply because it is much easier than the other way. Remember, plagiarism can also exist in how an article is structured, even if the words are different. That is more important if you are writing an academic article, than if you just want to rewrite an online article.

If you want to check whether your rewritten article is different from others already on the internet, I would suggest using one of the free plagiarism checkers available on the net. They are very useful.
« Last Edit: February 25, 2017, 11:43:13 PM by Meg »

Re: TBS3: suggested changes to the "Uniqueness" logic
« Reply #2 on: February 26, 2017, 01:08:39 AM »
aaah i see, so basically the more synonyms and sentence (word combinations) restructing i can do then the higher the percentage because each is relative to another including the original article...

okies that makes sense, so i guess for 1 article the best option is to simply do as your method which is paragraph>sentence>word/phrase spinning and use online checker tools to see if can ensure the article is in fact unique and not deemed to be plagiarized.

many thanks Meg :)

Offline Meg

  • *****
  • 7050
  • Freedup to do what you really want
    • View Profile
Re: TBS3: suggested changes to the "Uniqueness" logic
« Reply #3 on: February 26, 2017, 02:27:23 AM »
Yes. You have summarised in a lot fewer words than I used.  ;D

Offline andrewwilson

  • *****
  • 1496
  • catching the zeitgeist since 2005
    • View Profile
    • We Build Money Making Websites
  • Skype: andrew.wilson41
Re: TBS3: suggested changes to the "Uniqueness" logic
« Reply #4 on: February 27, 2017, 04:47:41 AM »
Two additional points to make:
1) In many topics it is almost certain that there will be some apparent duplication becasue of the jargon and stock phrases used in the subject. Copyscape, the leading plagiarism indication tool uses word groups (shingles) of three words and so it will pick up on 3 (or more) string combinations for duplication. Two word combos are so common that they are ignored.

2) The leading tool, Copyscape, is no longer a very good arbiter of uniqueness. Their index is no longer as comprehensive as it was when the Web was younger. It is still fast and useful but far from being an exhaustive check. The only way that I know of is to paste strings into Google or Bing's search tools and see what they turn up. Both Google and BIng have at least one copy of the entire Internet on their servers but even this is not foolproof as it is not possible to check an entire document and both major search engines provide a personalised, edited view of the Internet to individual searchers.

That leads to a third point, by accident. It has long been suggested that an article that is more than 30% unique compared to other pages on the Internet will be treated as unique by SEs. That seems to be roughly true, however, I have seen that content where the information load is significantly different to other versions tends to do better than articles where the words are different, paragraph order is different but the ideas and facts expressed are substantially the same.

So, a little bit of duplication does no harm at all, but the est case is where the information provided is spun as well. I do that by having several paragraphs that are spun in and out at random giving  not only a different structure but also different information.

Offline Meg

  • *****
  • 7050
  • Freedup to do what you really want
    • View Profile
Re: TBS3: suggested changes to the "Uniqueness" logic
« Reply #5 on: February 27, 2017, 05:22:45 AM »
And that kind of spinning can also be done in TBS.
I bought some PLR the other day and copyscape couldn't find any copies on the internet!

Offline andrewwilson

  • *****
  • 1496
  • catching the zeitgeist since 2005
    • View Profile
    • We Build Money Making Websites
  • Skype: andrew.wilson41
Re: TBS3: suggested changes to the "Uniqueness" logic
« Reply #6 on: February 27, 2017, 06:00:33 AM »
sometimes one hits lucky and can be the first to index a PLR product, but it won't last. :)

Of course, if one is not depending upon search traffic then duplication is not really an issue - pointing your Pinterest traffic at PLR, for example, shouldn't be a problem.

Offline Meg

  • *****
  • 7050
  • Freedup to do what you really want
    • View Profile
Re: TBS3: suggested changes to the "Uniqueness" logic
« Reply #7 on: February 27, 2017, 08:55:29 AM »
 ;D  ;D

Re: TBS3: suggested changes to the "Uniqueness" logic
« Reply #8 on: March 02, 2017, 08:12:54 AM »
Two additional points to make:
1) In many topics it is almost certain that there will be some apparent duplication becasue of the jargon and stock phrases used in the subject. Copyscape, the leading plagiarism indication tool uses word groups (shingles) of three words and so it will pick up on 3 (or more) string combinations for duplication. Two word combos are so common that they are ignored.

2) The leading tool, Copyscape, is no longer a very good arbiter of uniqueness. Their index is no longer as comprehensive as it was when the Web was younger. It is still fast and useful but far from being an exhaustive check. The only way that I know of is to paste strings into Google or Bing's search tools and see what they turn up. Both Google and BIng have at least one copy of the entire Internet on their servers but even this is not foolproof as it is not possible to check an entire document and both major search engines provide a personalised, edited view of the Internet to individual searchers.

That leads to a third point, by accident. It has long been suggested that an article that is more than 30% unique compared to other pages on the Internet will be treated as unique by SEs. That seems to be roughly true, however, I have seen that content where the information load is significantly different to other versions tends to do better than articles where the words are different, paragraph order is different but the ideas and facts expressed are substantially the same.

So, a little bit of duplication does no harm at all, but the est case is where the information provided is spun as well. I do that by having several paragraphs that are spun in and out at random giving  not only a different structure but also different information.

regarding point 3: how to check if my completely rewritten article is at least 30% unique compared to other pages on the internet? is there such a tool?


and for discussion, a question if i may.....
TBS is considered the best tool for manual control of article spinning, although i have written my first article by manually rewriting a complete article and therefore i have just 1 article.. how do i go about using the full power of TBS to generate many versions.. what i mean is.. where would i use all these generated versions of an article?
having 1 article is simple, i just post it to a site.. but it doesnt make sense to post 100 variations of an article to the same site/blog... so im curious how to use such articles... my plan is to build a website that will make some income from advertising links.. im just getting started so im curious how to maximise the site potential by funneling traffic from social media etc.... if my question is not suited to here then even pointing me in the right direction to a site or anything would be appreciated.

Offline Meg

  • *****
  • 7050
  • Freedup to do what you really want
    • View Profile
Re: TBS3: suggested changes to the "Uniqueness" logic
« Reply #9 on: March 02, 2017, 10:15:34 AM »
Some of the free plagiarism tools will tell you how much of your article is duplicated on the net. If the number is less than 70%, then your article must be more than 30% unique.

Some people have several sites in one niche. Spinning your article gives you something different to add to each one. I find spinning useful for adding comments to articles on product reviews. I can review several products and have a different review article on each.

Re: TBS3: suggested changes to the "Uniqueness" logic
« Reply #10 on: March 02, 2017, 10:14:13 PM »
wow... thats an interesting strategy.. quite different to others ive been discovering... im only new to the current state of things so loads more to learn im sure.. for now i will focus on 1 site... so i can see how everything goes together from top to bottom... then i will expand out.. and maybe use that technique :)

Offline andrewwilson

  • *****
  • 1496
  • catching the zeitgeist since 2005
    • View Profile
    • We Build Money Making Websites
  • Skype: andrew.wilson41
Re: TBS3: suggested changes to the "Uniqueness" logic
« Reply #11 on: April 11, 2017, 04:53:44 AM »
Multiple versions of content are typically used in link building. Use the various versions to point back to your site. Make sure that you don't use a version of the spun content as the target of your link building, doing so is usually obvious and, of course, provides little benefit to human visitors following such a link.