 So once I pick a particular state, I have all the counties of that state available. I'm going to be choosing Richland County, which is where I live. And what we are going to do is look at intervention date of June 1st. So the result is available here. So what we notice here is that again, the Prodolent Treatment Group, they behave similar before, and then there is a divergence. This is the actual for the county that we are interested in, and this is the counterfactual. So Richland County would have had 76% more cases, and which equates to 240 deaths, right? Now let's consider if we had done the intervention on July 1st, okay? So notice this was 76% more in the deaths. If they had done this intervention on July 1st, right? Then the difference would have been 61%. So notice that from 76% difference of the intervention, now this has reduced to 61%. And further, if I make it as, and the difference in deaths would have been 190. So if you remember in June 1st, the difference in deaths would have been 240, right? Now it is 190. So already there is a difference of 50, which you could have seen. So I did not tell you when actually Richland County did something. They actually imposed on July 2nd, okay? What we will see is that taking a mask intervention at any time is helpful, okay? But its impact is highest when you do it early. For June 1st, the difference would have been 76%. July 1st was 61%. Now we'll see the difference is about 60%, 60.45, okay? So what you notice is that the incremental benefit, right? You could have saved 15% more, right? If you had acted on June 1st, but you acted on July 1st, so you delayed it. August 1st would have still made a difference, but that's like not too much of a difference. So between July and August, not too much of a difference, but between June and July are very high difference. Of course, you cannot save this for every county around the country. You have to do it case by case basis. There's also a limitation in the sense that this survey for mask adherence was only done in summer, whereas when we are looking at intervention period, things keep changing, right? So you have to account for it. This is a data-driven tool. But what you see is that at any given time when you put a mask intervention, right? At first it makes a positive impact and the earlier you do it, the better it is.