went from hey this is great to at the moment the tests are kind of weak on Remdesivir. Right? So we thought it worked but now we think maybe it doesn't based on some clinical trials.

What about hydroxychloroquine? If you look at the news on the left and the tweets on the left, they will say it has been proven not to be effective. But all you have to do is get on Twitter and there will be people tweeting all kinds of studies, usually retrospective but some of them even clinical trials, which purport that it works just great.

Now both of these bits of data, which are opposites, they can't both be true. But they're reported with equal vigor. There's as much energy around saying hydroxychloroquine clearly works and all the information is supporting it as there is people saying it's been studied to death and there's no impact whatsoever. Both those truths are existing in full force and they can't both be true. They just can't be.

Now I'm of the opinion that if it were a big deal we would have noticed it. Meaning it would just be so obvious that you wouldn't even need a clinical trial if it were as effective as its initial proponents said. I don't know if it has no impact. I just don't think it has some kind of a magic pill impact. So we don't know about hydroxychloroquine.

We still don't know about vitamin D. We know it's good for you but we don't know how much of a difference that's making. And we're still arguing about whether masks work.

So yesterday I was tweeting about somebody. You've seen a bunch of these. Somebody will show a bunch of graphs of coronavirus infections by state or country or whatever and they'll say here's the point that masks were mandatory. And you can see that after the masks were mandatory the infections went way up and then they sort of started trailing off on their own for reasons that nobody could understand.

Now let me draw you the picture in your mind. So you're looking at a graph that's mostly slightly rising. It's closer to flat but it's slightly rising. And then suddenly it zooms up to a mountain and then the mountain crests and it goes back down. So that's what a surge of coronavirus would look like on a graph.

Now imagine on the graph that at the base of the mountain part is labeled "masks become mandatory." And it takes about a month to get to the top of that peak and before it trails off. What would be your conclusion if somebody showed you a bunch of graphs where every time the masks are mandatory the infections still get way worse after the masks are mandatory? How would you interpret those graphs given that there's a bunch of them, a whole bunch of them? What would you say? Would you say masks don't work? Because every time you have them as mandatory you can see on the graph that the infections go out of control higher. Therefore they don't work. Right? Because that's what the people tweeting those graphs are telling you. They're telling you look at the graph. It's just as plain as the nose on your face that's covered by a mask. There's the mask. They're required. And then the infections go through the roof. Clearly masks don't work. Right?

And I look at these same graphs and I say that's not what I see. I'm looking at the same graph you're looking at and I'm seeing proof that masks work. Same graph. Exactly what I just described to you.

Why? Timing. How long does it take for mandatory mask wearing to work itself through the system? Remember you've got some reporting delays and these delays could be a few weeks. So it could be that the day the masks are required you still get a bunch of infections that have not yet been reported and they're going to come in after the masks. So you're going to be reporting a whole bunch of infections that may have been last week's infections.

Secondly, do people immediately get the right kind of masks and do it universally the same day that you put a mask requirement in place? No, no. Right? People are not instantly complying. There's probably a little bit of a time lag before people get the better kind of masks. You know, maybe they start out with a bad kind of facial covering but then they buy one and now they've got a better mask. So you'd expect that the effectiveness would be a week or two before you're really all masked up. Right? Wouldn't you?

Then what about the fact that people who were infected and they don't know it yet. So they got the virus yesterday but the mask requirement goes in today. A week from now we realize that they were really sick but we just didn't know it. Where does the infection get recorded? It gets recorded at the point where you discover it. Right? So it's going to look like that infection came after the mask but in fact it started before the mask.

So here's my larger point. Can I look at those graphs and then conclude that masks work? Because there's a little bit of a time lag. But in every case, in every case on those graphs, the infections would reach a peak and then trail off very quickly. To me that's a picture of masks working perfectly just the way I would have expected them to work. I would not expect them to work on the day of implementation. I would expect to see the effect maybe three weeks later, which is about what the graphs show.

Now is my interpretation accurate? That these graphs that people are using to show that masks definitely don't work, is my interpretation correct that those same graphs are proving that they do work? Which one of us is correct? Do you know? Before I started talking about this, did you say to yourself, well obviously if infections go up like crazy after masks, obviously they don't work? Did you think that before I started talking?

Now I'm not going to make a claim that my interpretation is correct. Because if you know how to analyze data you should be asking yourself this: Where is my comparison to that same city or state that had the same problem and then they did not have masks? Right? Because you would have to compare it place to place. It'd have to be the same place to the same place. Otherwise you're not really comparing.

So the real answer is we can't tell from those graphs. Those graphs don't tell you masks work and they don't tell you the masks don't work. They don't tell you anything. Because we're really, really bad at gathering data. And there are other folks on the internet. I'm still having conversations with them. But there's some thought that even the current number of infections and deaths may be lagged by as much as months. So we don't have data we can rely on in any way about any of this stuff. It's just useless. And the thinking is useful is kind of dangerous.

All right. I asked on Twitter, I asked how many people have experienced fewer colds and regular flus this season. And I think this is again one of those things where this is purely anecdotal. But it was the summer so you shouldn't expect that there would be too many colds and flus in the summer. But let me ask all of you. Do you feel like you've had fewer regular illnesses since the coronavirus issue in say February? I feel like there are fewer of them now. I don't know if that's true. It's probably a bias. But it just feels like there are fewer of them.

There's somebody on Twitter who claims to be in the business of selling cold and flu medicines. So it's somebody who's in the industry and they say that the sales of regular cold and flu medicines is down. Meaning that regular colds and flus may be substantially down. Which would make sense, right? Because we're socially distancing now.

If it's true that regular colds and regular flus are way down, and the lack of sales of those products that treat them would suggest that's the case, then wouldn't that be pretty good evidence that masks work for coronavirus? Now that's not proof because regular colds and regular flus may have some differences that are not obvious to me as a non-medical person. But if it were true, just take this as a hypothetical. If it could be proven that our regular colds and regular seasonal flu are way down this year, if we could prove that was true, would you be willing to say that masks work? Or would you still find it? Just asking.

Now here's the latest good thinking that I've heard on masks. And it goes like this. If you and I are in a small room and let's say one of us has the coronavirus and we both have masks and you and I stay in that small room breathing our shared air for hours at a time and one of us has the infection, what are the odds that the other one will get the infection? Pretty good. Pretty good. And it's because even though the mask might be blocking some of my direct airflow going directly out, the air is going somewhere. You exhale so it's coming out the sides of the masks or whatever. So it's going somewhere. So eventually if you and I stay in the same closed room with the windows closed and bad ventilation, it doesn't matter if we have masks or not. So in that one scenario, do masks work? I'd say closer to no than yes. Meaning that if you stay in that room long enough and the ventilation is bad enough and one of you has coronavirus, the other one's going to get it. It's just a matter of time. Right?

But suppose you and I are at a bar and it's a big bar and I come up to you drunkenly and I talk a little too close to you and I've got my mask on. It's just a brief encounter. Do you think that would make a difference? I think yes. Because the mask would be dispersing my airflow out the sides and it would be out there but it wouldn't be like a hose of my bad virus directly into your mouth and your eyes. So if I'm talking to you from two feet away, I'm not jamming virus into your face. It's sort of coming out the sides. That's got to make a difference. Right? Doesn't common sense tell you that in that scenario probably it makes a difference? Whereas if you're locked in the tiny room with no ventilation, probably it doesn't. And maybe if you're outdoors the difference is so small it's not worth it. But we don't know.

I tweeted a link so you can see how the CDC estimates the number of regular influenza deaths per year. And the reason I tweeted it is for you to see how ridiculous it is. So I've been making a claim that sounds so ridiculously stupid that nobody believes it. I don't think I've convinced one person that the following is true. But I'll say it again. I like being a contrarian. We don't know how many regular influenza deaths there are every year. And almost all of our conversation about how bad the coronavirus is is compared to this number we believed was a pretty solid number. The number of regular influenza flu deaths per year, which people say is in the low tens of thousands but could be in the high tens of thousands if it's a bad year.

So look at the CDC and look at the tortured, convoluted way that they estimate it. And if you have any experience in data analysis, and I think you would need it to come to this opinion as I do, there isn't the slightest chance these numbers are good. Not any. And when I say there's not the slightest chance they're good, I don't mean they're off by 10%. Do you feel me? I mean they could be off by 200%. They could be off by 90%. There's just, when you look at how they calculate it, you can't even understand it.

There's a general rule of life that if somebody can't explain something to you, let's say you have average intelligence, if somebody can't explain it to you, it's right. And if somebody has to explain it to you with a whole bunch of word salad, it's not because they're bad at explaining necessarily. It's because there's nothing there to explain. It's just so. The CDC estimates and the way that they go about doing regular influenza is laughingly ridiculous.

And I would like to put this challenge out there. So if there's anybody who, just look in my Twitter feed. I tweeted that within the hour. Look at that link. Look how the CDC estimates the influenza deaths. And if you're experienced in data analysis, and that's the important part for this, if you're experienced at it, just look at their explanation and then tweet at me later that you think that those are useful estimates or not. I think you're just going to laugh when you see it.

And here's a little factoid to put on top of this. And I need a fact check on this. So every year we know that there's a vaccination for the regular seasonal influenza. But we also know that in various years that vaccination can either be pretty good, meaning it'll protect a lot of people, or they didn't quite get the right formulation for the virus that emerged and it's just sort of not that good.

All right. So would you say that on the years that we have the really good and strong version of the vaccine that the total number of flu deaths should be lower? Right? Because that's the year that the vaccine is working really well compared to a year where we know the vaccine wasn't a good fit for the virus and it didn't really protect many people. You would expect that those would be the years you'd have a lot of flu deaths. Right? Nope. Nope. Apparently there's no correlation between how good the vaccine is and how many people die. Now I need a fact check on that. Don't take that as true because you heard it on this Periscope. I'm explicitly asking for a fact check. This is just something I heard on Twitter and it could be untrue easily.

All right, moving on. Ian Bremmer, who is always interesting in part because I can't tell his politics, which is a compliment. Let me give Ian Bremmer one of the best compliments that a Twitter user can ever have. I've been following him for quite some time. I can't tell if he's a Republican or a Democrat or an independent. Isn't that pretty good? Because he has opinions which seem well reasoned that some are anti-Trump, some are pro-Trump. But in all cases they don't seem crazy. He doesn't seem to have any crazy opinions. You know, I don't agree with them all but when I don't agree it's usuall