Wouldn’t it be better if I just ‘customise’ my own STI ETF by taking maybe say only the top 5 performing stocks within the ETF instead of investing in STI ETF?

Yes, true, past performance is not indicative. But this is really an interesting question, the same as how would a "S&P 100" or "S&P 5" compare to the "S&P 500".

Small caps could be more successful so maybe with larger diversifcation performance could be better. However, given the tech stock dominance of the current S&P500 an "S&P 5" could be even extremely successful. For the STI things - I feel - would be completely different.

For knowing whether STI or a subchoice of it would do better, You needed (at least for some insights) to have analysis of long-term data over the last 10-20 Years.

Maybe, yet, it is much easier - with less time and thinking spent - to do it the very easy way (maybe all You need could be: ... only ... some VUSA, some MSCI World ETF, some physical gold).

everything else here:

https://seedly.sg/questions/what-is-your-genera...

1. As you mentioned, being the top performers in the past period does not mean they will continue to be the top performers for the next period

2. In a way timing the market (which you could be too late)

3. You are increasing specific risk by concentrating into few industries

4. Not cost efficient if you have to liquidate and buy new stocks frequently

5. You minimise your potential returns since you go for the uptrend instead of going into stocks that are underperforming

6. Many variables in your model (If stock is top performing but overvalued, do you still buy? During a market downturn, do you sell your holdings and buy stocks that drop the least?)

7. Hassle for a long term strategy (esp. for a retail investor)