That is, if the price were to fall from the present $38.44 to $0.9 (really improbable) I'd have bought 38 grams.
In this scenario, you'd have 38 grams of gold at an average cost of $19.50 per gram. That gold is now worth $0.90 per gram, for a 95% loss. Granted if you had bought it all at $38.44, your loss would have been 98%, but I would not call that a winning strategy.
Your assumption is that the prices will always go up from where you bought each lot, which is not a certainty.
What you are describing is similar to dollar cost averaging, meaning you buy periodically so that sometimes you buy high and sometimes you buy low. The problem is it has not been proven to be a good long-term strategy. It only helps reduce losses when prices go down because you don't buy at the initial higher price. In stable markets it makes no difference, and in rising markets you're worse off because you could have bought at the initial lower price.
If you think the price of gold will rise over the next 10 years, then buy it now and wait 10 years. if you think it will fall over the next 10 years, then don't buy it (or short it if you can). If you have no clue, then don't speculate in gold.