You can't just use an arbitrary game to test your hypothesis like that. You need to test with something that caps out the 4090 at 100% usage and then see how a 1070 performs in comparison. It's not quite that simple of course, so just look up some benchmarks ;)
