Skip to main content

Spam Blog False Positive Classifications Do Exist

No matter what Blogger does with improving the spam classification process, they will always generate some false positive classifications.

The more that they improve the classification process overall, the more obscure the false positive classifications will be. Some false positives will require intense investigation of the blogs, to identify the relationships between the blogs involved - and to prove, to Blogger Policy Review staff, that a false positive classification is present.

Some time ago, in Blogger Help Forum: Get Help with an Issue, I had occasion to deal with a blog owner who was a currency trader, by profession. This person had published a cluster of blogs, where he discussed legitimate issues involved in currency trading.

Currency trading is an essential part of exchange of merchandise, between different countries (the export / import business) - and the Blogger spam classification process had deleted his blogs, as suspected spam hosts.

For you to understand the situation that I'm recounting, you need to understand one principle of currency trading.

When someone sells US Dollars ("usd") and buys Great Britain Pounds ("gbp"), to help a client in the USA import merchandise from the UK, there has to be a second trader, buying US Dollars and selling Great Britain Pounds. The first trader is doing a "usdgbp" trade, and the second, a "gbpusd" trade.

The blog cluster, that was setup by the trader, discussed issues about trades between Great Britain Pounds, Japanese Yen ("jpy"), and US Dollars. The trade combinations possible, with just 3 currencies, would be
Most traders handle dozens of currencies, in various trades. A staff of a dozen traders would handle all of the world currencies - with some trades involving 2 traders working together, on either end of the trade.

The trader in question had published a cluster of 6 blogs - all using the same base blog name (we will call it "Currency Trading"), which gave him 6 blogs - with legitimate names.

Each blog discussed specific - and very trade sensitive - issues about trades between the specific currency pairs. Each blog provided relevant and unique discussions about the issues.

Thanks to splog detection, a false positive detection - based on legitimate splog classification technique - had, nonetheless, left this blog owner with 6 deleted blogs.

Thanks to one of my previous IT responsibilities, as tech support for a bank currency trading desk network, I recognised the naming convention used by the 6 blogs, and the reason for the 6 blogs - and was able to convince Blogger Policy Review staff about the problem.

Some blogs are part of legitimate blog clusters, which use well defined naming conventions. Others have names that are coincidentally similar to other blogs, owned by different people.

In either case, the reasons for the naming similarities are where we start, in identifying any spuriously deleted blogs - and in describing the reasons for the spurious deletions. Blogger Policy Review staff have to be made aware of the legitimate reason for a naming similarity, and any other details, so they can make an informed decision.

Knowledge about obscure relationships, such as the naming convention for currency trading blogs, are not part of every Blogger helper or staff persons background. Sometimes, the blog owner has to provide details, objectively and patiently, to help convince Blogger Policy Review to restore the blogs involved.

Unfortunately, not all blog owners are patient enough, to be interviewed, to some degree of success. Also unfortunately, some self stated blog owners are actually hackers or spammers, misusing Blogger Help Forums resources.

Some owners of legitimate blogs claim that their being subjected to an interview shows suspicion placed on them, rather than an attempt to identify the problem - and successfully prove, to Blogger Policy Review staff, that a false positive spam classification is present.

The complicated and devious relationships, involved in some spam review request investigations, make a Tom Clancy novel seem like a 5 minute segment of the 6:00 news.

>> Top


Popular posts from this blog

What's The URL Of My Blog?

We see the plea for help, periodically I need the URL of my blog, so I can give it to my friends. Help! Who's buried in Grant's Tomb, after all? No Chuck, be polite. OK, OK. The title of this blog is "The Real Blogger Status", and the title of this post is "What's The URL Of My Blog?".

Where's The Dashboard?

We see this confusion, a couple times a week, in Blogger Help Forum: How Do I? . Where is the dashboard? In the Classic Blogger GUI, the display which contained the "Blog List" (at the top), and the "Reading List" (at the bottom) was labeled "Dashboard". Many people also called the "Settings" / "Template" screens for the various blogs, linked from the Blog List, the dashboard. The New Blogger GUI has no page with the label - and no links "To The Dashboard". The Navbar (another unlabeled feature) has two links - "Design" and "New Post" - which lead to different dashboard sections, when you are appropriately logged in to Blogger . And, the "B" logo at the far left of the navbar will, similarly, take you to the Blog List / Reading List.