I've never seen anyone here discuss reverse search engine algorithms. In fact it is important to do so. In the field of SEO, we need such pioneers who are willing to solve those very difficult search engine principles problems. I hope we can discuss SEO from a new level and I will help everyone solve this problem.
Disclaimer: This article is not 100% accurate. Please point out the flaws or come up with a better theory. This is science time.
Are you ready? Let's get started!
I've always heard that the pyramid link model is better than sprockets. But there is no convincing explanation. More often than not, people will use a vague concept to explain that doing so will make Google feel that the links are more chaotic and irregular. This sounds good, but it still doesn't explain the problem. Our task is to explain this problem to everyone from the perspective of search engine algorithms. Remember: what I'm talking about is just a theory, and it may be slightly different from reality. All I'm saying is the way I feel is correct and the simplest way to explain it.
First, what is the difference between the sprocket and pyramid link models? This is the general structure of a sprocket. Looking at the picture below, the nodes on the inner wheel are not only connected to the main station, but also linked to the next node. The PR value will be circulated through these pages, and finally all the weight will be aggregated to your main site. Sounds good, right?
But in fact this is not the case for Google. The weights are not transmitted as we suspected. This is perfect, completely explosive, different websites, different IP addresses. How capable are they? Sometimes, it works.
When Google successfully renders a popular technique ineffective, it makes some major changes. This is what black hat technicians need to study. We overuse this sprocket trick, just like blog comments and forum profiles, and Google easily adjusts the weight of such links. But in fact, Google does not deal with blog comments and forum profiles in this way. Google has proposed many years ago that if the links come from pages containing spam information (such as guestbook), the weight of these links will be reduced. But Guestbook Spam still works. Search engines are great, but it’s hard for Google to stop things like this. But how could they successfully identify the sprocket?
When you carefully construct a sprocket, you think Google will never discover its existence. But the result is that it turns out that as long as simple graph theory is used (Graph Theory: it is a branch of mathematics, which uses graphs as the research object, and is usually used to describe a specific relationship between certain things), the sprocket can be The structure is depicted. If you have some simple knowledge of graph theory, you can think of each page in a sprocket as a node. The shape of the sprocket is called a cycle in graph theory. A cycle is a closed loop formed by multiple nodes linked to each other. You start at a certain point and end at the same point. Here is a simple diagram.
This is the same structural formula as our sprocket. Unfortunately, Google has a way of dealing with this.
Wait, Google can detect this powerful structure but not those blog comments and personal data? Yes! Because this is a very simple data structure, every computer science student understands it.
Any sophomore student can write a program to access a graph with path directions (directed graph). It's very easy and doesn't require a lot of processing power. What the computer needs to do is walk through all possible paths, which is simple, especially in a directed graph. When the computer starts walking from any node and finally ends at this node, all nodes that have been walked will be marked as a cycle. When Google discovers that every node of low-quality content is part of a ring, it won't be polite. In fact, such severe penalties would not involve any risk of miscarriage of justice. Will there be any unexpected situations? The answer is almost impossible. No matter how many nodes there are on a ring.
This is a very simple theory, and engineers can easily solve such problems. Now let's ask engineers how to reliably determine message board spam. They may be able to solve it in the future, but it will definitely require a lot of semantic analysis and detailed standards to avoid misjudgments. In this way, the Google server needs to process a large number of instructions and requires a lot of storage space. It is almost impossible to constantly judge the possibility of each link being spam and reduce the value of such spam links. Google tried this fantastic idea as early as 2002, because at that time black hat SEOs began to use message board profiles to cheat, but Google failed. Next time you are crazy about making Profile links, you don’t need to think about "Will this thing work?" "Will those stupid engineers find this?" You only need to think "Haha, Google, what can you do to me?" (Night note: I personally have reservations here. Google has applied for a patent for profile identification: US20080243780 OPEN PROFILE CONTENT IDENTIFICATION
http://www.patsnap.com/patents/view/US20080243780.html Students who are interested in English can read it. )
Okay, this is a bit of a stretch here, but this is indeed what every black hat SEO needs to know. Back to our link structure building. Why is the pyramid link model better? Clearly, someone has tried it and it works, and the results are good long-term. Why can Google easily identify sprockets, but this pyramid model is still valid?
Let me tell you!
A pyramid has a perfect triangular shape, and everyone knows that triangles are stronger than circles, plain and simple. Instead of making a complex shape, you can layer your links. But when you consider that the pyramid is also a picture, this is actually not a big problem because it is a triangular structure. In fact it can be treated as a ring, which is the same as a ring, but an unclosed ring.
A nice shape, but where are the vector graphics?
Regardless, in pyramid linking, the graph that each of your nodes forms is known in mathematical circles as a cyclic vector graph (DAG). Such a pyramid-shaped DAG must have a special name, but I don't care what it is called. Each link in this pyramid structure is important, but the biggest feature of DAG is that it is a hidden structure. Here are a few points explaining why it’s so effective.
1. By definition, there is no ring structure in DAG
2. Starting from any node in any DAG graph, you can crawl the entire graph through hyperlinks. Without exception.
So forget about the sprocket detection algorithm and the fatal penalties. Will the deliberately crafted DAG link structure be detected starting from the main site? The answer is no.
Kind of weird, isn't it? Mathematically speaking, it is impossible for Google to detect such a pyramid link model. I bet you've never thought about this.
Now for the final thought, I'm sure you sitting in front of the screen have discovered that as long as the link between any two nodes in your sprocket structure is removed, the entire sprocket will become difficult to detect. This is absolutely correct! There is only one way from any node in the DAG to your main site, and there will be at least 2 in the sprocket.
DAG doesn't create any obviously unnatural link paths, so don't worry about perfecting that magic sprocket that will help you double your PR. The old sprocket structure is no longer effective, and now we need a new optimized structure.
or
Don't worry about linking one node to another and creating a new pyramid structure.
think about it
Can you come up with a more efficient DAG link structure? There are many, but you have to make sure it's a DAG!
If you could create a better structure, would it be possible to use tools to build or build a new pyramid structure faster and more efficiently?
Can you design an algorithm that automatically creates a better link structure for you?
Is your algorithm feasible?
I look forward to your reply!
Author: Night Breath
Article source: http://yexi.co.cc/?p=235 Please keep it for reprinting