On balance volume indicator showing different signal on different time frames

Roll 2 6-sided dice twenty times, and put the results in a spreadsheet.

Let's say the results were:

7
6
2
9
11
5
2
9
5
5
4
4   
3
3
4
7
10
10
12
12

Now do a 'trend analysis' on the last 4 rolls. "Averaging 10, this baby is going up!".

Now do a 'trend analysis' on the last 10 rolls. "Averaging 6.9, this is a little below average, better hold!"

Now do a 'trend analysis' on the full 20 roll run. "Averaging just 6.5, SELL SELL SELL!"

The point is that you can easily adjust the data being looked at, to present the 'trend' or conclusion you want. It is trivial to find patterns in random data, and impute a causal relationship to project a false estimate of the future. Finding patterns is more or less mankind's greatest strength and also a modern weakness.

You are using questionable techniques to 'invest' [gamble] on a speculative asset class. You are looking for logic where none exists.

Some indicators are more susceptible to “take away” effect of early data (early data points are removed from the calculation as new ones are added). An example of this would be moving averages or any indicator based on moving averages. A 10 period MA might be trending up whereas a 20 or 50 period MA might be trending down.

Because the RSI is a ratio, it eliminates the problem of needing large amounts of historical data and therefore is less susceptible to this “take away” effect.