Can you explain if the following scenario about vehicular homicide is in any way truthfully applicable?

1) the guy is an idiot for drinking more. Police are entitled to draw adverse inferences.

2) if they do a blood test (as opposed to breath test) then they can see what percentage of the alcohol has metabolized and calculate back to when he drank it.

If he killed someone, he’d probably be breathalyzed right away and taken down to the stationhouse. So he’d not have an opportunity to drink more prior to police intervention.

No, the test would not be able to determine the original alcohol/blood content.

I agree with @kritiper . The answer is “No”. The test cannot tell the difference – unless it was taken immediately after the person drank the second time.
The alcohol from the second slug would be in the drinker’s stomach but would not have entered the bloodstream yet. In that case a followup test would theoretically show the BAC actually increasing. With third and fourth tests taken a fixed time later you could theoretically estimate the before and after accident consumption.
Note I say theoretically because the test precision and accuracy and the driver’s own metabolism are probably not accurate and repeatable enough to infer both consumption times and quantities.

No. And that defense has been used and was successful in the past. At least it was successful for avoiding a DUI charge. Reckless operation and vehicular manslaughter are another story altogether.