How adding more indicators does not increase profits

To answer this in a deterministic manner and not to confuse you further, I would first ask you to throw another important question – whether the indicators you have posses any real edge in market? By edge I mean winning tendency over and above 50%, so if your indicator have winning tendency of 56% then it has edge of 56-50 = 6%, notice that 50 is discarded, and does not become the part of our overall edge calculation. To be clear, I should mention that win rates are calculated using strategy back-testing over a sufficient time series data.

Why, we eliminated 50 part of winrate is because all kinds of financial market provides you a default edge of 50%, without doing anything on your part, basically it just means that, in an efficient markets at any points of time the tendency to go up or go down are always the same, no matter what strategy we deploy. We know that markets are not 100% efficient but for most mathematical model based strategies this still holds a solid ground. Therefore 50% is not an edge in reality, when we combine anything of 50-50 probability it will not enhance our win rate anyway, this is what our probability theory also tells us.

Lets run through a very simple example for our understanding, lets say we have three indicators A, B and C of win rates 50% each, in a mathematical language, when an indicator A is true state we have 50% probability of winning, and same goes for B and C. Now we are asking when A, B or A, B, C have simultaneously true states that time whats the winning edge ? is it more than 50 or less !!

As you figured out, they all have 0 (50-50) edge, therefor when system is stacked by A+B+C,additional stacking do not change existing edge any further.

To put it in simple words, only if an indicator have over and above 50% edge, then combined with another indicator could results in further improvements of overall winning probabilities, and we know that most of our indicators have 50% or less (most are in range of ~35%) win rates, they become less than 50% because of brokerage and late entries, so by now it should be clear, that such combination stacked up together cannot lead you to a more profitable trading system.

Does adding more indicators increase profits ?

Anyone with some experience in using algo/signals, would have certainly wondered, can adding more indicators lead to better results than single indicator? Here by term indicator I mean builtin indicators like MACD, RSI, Bollinger Bands etc or Custom indicators too.

Lets say by some logic, indicator “A” of 50% of win rate and we add another indicator “B” of 50% winrate, a compound probability should be more > 50 ?

It doesn’t work that way in financial markets, even by simple math adding two systems with  50% win rates would yield max of 50% in result. Also note that most indicators have below 50% of winrate. So basically here we are not gaining anything instead it will take a hit on your trade system.

Biggest price you pay when include many indicators, complicate your trading system which will now behaving randomly and skip signals when expected and lose synchronicity with manual analysis. Lets see why it doesn’t improve profitability.

First, all indicators are derivatives of price history so they don’t complement each others, therefore don’t expect good from such combination. Then fact remains the markets are much random and fractal in nature, in such markets a 50% win rate is not at all a win rate, rather its a random probability of zero, funny thing is from figure 50% you may think that you have got an edge of 50 out of 100.

Just think about it, if it would have performed even in single occasion, it doesn’t take expert programmers to combine 100s of indicators and trade in huge quantity.