Please note - whilst we discuss trading, financial analysis, it is simply an example to highlight challenges with "The Wall" and is not financial advice.

What is The Wall?

A wall, is a page which continually updates with new streams of information. It is intended to keep users engaged, reacting to new wall posts which may or may not be important to them. Users who interact with The Wall can go through a range of emotions.

Serious thinking must be undertaken when considering the true benefits of wall applications.

Feeds of content, typically other users or information, continually streams nudging users to interact with it. They can become addictive and can cause many individuals emotional stress. Most platforms adopt hierarchical top-down approaches, permitting users with higher follower counts to push their ideas and policies. This tends to marginalise ideas with lower follower count, which in itself can lead to polarisation of large numbers of users to take potentially harmful action.

Ignoring any nefarious elements of walls, we want to understand why they are not really the best way to create meaningful and long-term engagement with target audiences.

The Wall has a present and barely a past

Feeds of content, typically other users or information, continually streams nudging users to interact with it. Only the latest news is relevant, even what happened a day ago is no longer relevant.

Each wall participant is the platform's product

If nobody posted on a wall, it would serve no purpose. Without continued participation and engagement, the platform would cease to be. The very nature of this means we need continual stimuli and responses to maintain engagement. This leads to the addictive nature and narcistic projection of our perfect self - whether we are a business or an individual, image and reputation are vital. At other times, we are agnostic to our own portrayal, instead it is the ideas which we promote to influence others that matters most.

A simple example of a wall as a treadmill - cryptocurrency YouTubers

The cryptocurrency space is constantly changing, prices of currencies collapse and rise far more regularly than other sectors. Development of new technologies is very fast paced. YouTube personalities job is to keep themselves relevant, get paid for viewer numbers, by reacting to what is happening now and giving their opinions on it.

We have to ask how useful is most of this? What benefit is it to the individual? It is for certain, the YouTuber won't tell their audience to sell their Bitcoin at the precise moment before a major collapse. Even if they did, they would be unlikely to be able to continually make these calls right. Does this mean there is not a place for content producers who are doomed to being wrong quite a bit? We don't think so.

Getting useful information from streamed information

Sticking with our cryptocurrency YouTuber example, what we really want is mechanisms which can attach similar significance to the past, the present, and the future. We want opinion which accurately focuses on past events, just as much as current events, and can not only try to predict future events but also consider their predictions on future events once those moments have passed.

We may prefer to simply have mechanisms which automates what we would do optimally in terms of buying and selling cryptocurrency.

Bitcoin recently hit �50k in the UK (November 2021). There were no shortage of YouTubers jumping out to tell us how Bitcoin is going to �70k, �100k, �150k, �500k. Yet, quite predictably to me, it is now at �42k. There will have been many YouTubers warning about a pullback whilst stating they are bullish.

Weighted responses for streamed information

What happens right now, often seems important, often though - it is incorrect, opinion based, salacious and unhelpful to those who consume that information. At other times, small snippets of current insight are ignored because the signals aren't sufficient for us to take action.

A weighted response model

Event of type A happens > (Perceived Positive | Perceived Negative) > Action Taken (Positive Action |No Action | Negative Action) > Decision Review (Decision Good | Decision Neutral | Decision Negative).

We can of course think of this in terms of reinforcement learning. However, does it need to be that complicated?

If we think about a typical Bitcoin Youtuber, he may say;

• Bitcoin is forming a head and shoulders.
• If it drops below this neckline, it could fall.
• If it breaks out of it, it could see a sharp price increase.
• I am cautiously bullish.

If we consider there to be three alternative futures;

• The price rapidly rose.
• The price stayed the same.
• The price dropped 10%.

The YouTuber can say;

• I took massive profits on this one.
• It is trading sideways and seems to have broken the head and shoulders for now. Am waiting for a signal to the upside.
• I took a slight loss and I will look for the opportunity to buy back in.

If we think about the above opinions of the YouTuber, what they have actually done is to fully cover every eventuality and avoid being wrong. This is no different to many; politicians, customer relationship managers, doctors, or lawyers.

The wisdom of crowds?

Ultimately, actors in this space are either;

• Always wrong.
• Often wrong.
• Neither right nor wrong.
• Mostly right.
• Always right.

We simply want to know when new information is relevant to us and what the correct way to interact with that information is. We should feel free to interact with streamed information, but limit the negative consequences of continually interacting with it - for our own sanity as much as anything else. We would ideally be able to not be influenced or affected by those who are not providing optimal information.

Adding visual perspectives to streamed information - Dashboards

Sticking with our cryptocurrency price example, we know that there are different futures at any present. We know that if we can at least try to account for Bitcoin prices, we may not want to spend our time continually in the present. Consider these possibilities for Bitcoin;

• In the last month, Bitcoin has traded within it's 25% and 75% quartiles of it's daily price range 400 times.
• When BTC wicks above it's daily close, it will retest that intraday high within the next three days.
• When Bitcoin has less than a 2% price volatility, prices drop 10% after 5 consecutive days.

Just taking simple examples like above, suddenly, we don't need to constantly keep watching YouTube, price charts, or news articles. Instead, we can take a probabilistic based approach, and use this information to adapt our dashboard to suit what we think.

Burton Malkiel's famous book - "A random walk down Wall Street" debunked the idea that market analysts can consistently outperform market averages. Many market professionals think the opposite and use indicators, signals, machine learning to try to out perform the market. Technical Analysts uses charting to try and do this. Some use charts to try and spot fractals, and see when similar moves are happening.

Thinking about The Wall and benefitting from it algorithmically

Can we really model all human interactions on walls, incorporate data, to try and give us a critical advantage from these types of systems? Or are they inherently human, and flawed?

We see that streamed information rarely gives individuals a chance to reflect, and to assess their own response to it, let alone determine the positives and negatives of the stream.

Those thinking about designing platforms which presents a stream of information should put dashboards and mechanisms in place to act as checkpoints on previous decision points. Those platforms interrogating natural language will have a harder job than those which use quantitative information but it is worth saving your users the angst of constantly interacting with the wall.

Another interesting possibility is offering rewards to those who provide good insights and information. This means they can be more focused on providing good information, than spending time chasing users and audiences.

We hope you enjoyed our musings on the challenges of streamed information, sound decision making, and possible ways to get more benefit from platforms engaged in streaming content.

Written with StackEdit.