By itself, the RBTI rating provides a fairly accurate assessment of trend indicators. But in theory, we may encounter such an indicator – which will perfectly adjust to the optimization period and give a huge IMP_opti and at the same time completely fail the forward tests. At the same time, the RBTI rating will be quite high.
The described situation will show a very low Prediction Coefficient (PC) and will give us the opportunity to understand that something is wrong with this indicator. At the beginning of 2025, we are only at the beginning of the journey, with the accumulation of experience we will be able to introduce additional filters to get into the rating.
For now, let's limit ourselves to the rule: if two indicators from the top 5 RBTI have similar ratings, but very different prediction coefficients, we will conduct additional testing on other symbols in order to understand the problem. Perhaps in the future this will lead to the need to modify the rating calculation formula.