Hi, so I wanted to cover Nomics and our data and why we're different. We found that most price aggregators and most market data services are failing in a number of ways that I think we've solved for and I wanted to cover that first.

A little bit about the company: we are an API first product company, so out of everything that we do our API comes first. We built the API before we built anything else. If you do go to Nomics that entire website was built with our API so anything you see on the website we have available but we also have a lot of data available that is not on our website.

Our investors include Coinbase Ventures, Digital Currency Group, CityBlock Capital, and a bunch of people that we really respect in the space.

So I think perhaps the way to start this out is by talking about data and data quality. So, our service and most of what we do is based around raw trade data, right. So, for the majority of the exchanges that we have data from we have literally every trade on every trading pair on that exchange. So, we have essentially the entire trading history of that exchange and from those trades we construct candles and from those candles we construct tickers.

Here we have on this chart, trades. As you can see, this is fairly high fidelity. This is Binance BTC/USD market. Here we have candles for BTC/USD. As you can see, there's just a lot that's left out—you can actually hide a lot of fake volume in candles. And then you have tickers—which a lot of our competitors are gathering ticker data rather than candles or trades. And ticker data is pretty bad... You essentially get tickers whenever they're computed, you don't necessarily get them at a specific time, so if you want to find out what an asset was priced at the end of a given time period you can't do that with tickers. There's just a lot of problems...

So probably a good way to think about data and how we do data is around this idea of a data pyramid. So at the bottom kind of underlying everything that we do is gapless historical raw data. So let's say, for example, that you wanted to price Ethereum. We start out by gathering every—let's start with a trading pair on an exchange, right, because there's a lot of trading activity on Ethereum that isn't with USD or fiat pair. So, we would get, for example, every trade on the Ethereum to BTC pair (ETH/BTC) on Coinbase Pro. We'd start off with all the trades on all the Ethereum pairs—and this is an example of one. Then we would move to creating exchange candles based on this pair. So, for example, if we have all these trades, we can create the candle for the Ethereum to BTC pair (ETH/BTC) on Coinbase Pro. From there, we can create aggregated candles for Ethereum/BTC across all exchanges, and then we would aggregate those and arrive at a price. So there's a lot that goes into this and a lot of our competitors just are ingesting tickers or candles and we normalized the way that we compute candles based on the raw trade.

So, what we found in some cases is that exchanges are reporting candle data that is, in fact, inaccurate, right. They'll pump up the volume by just adding volume numbers to their candles and when you actually count—when you have gapless historical raw trade data—you can actually like count each individual trade and add it up and get to the volume and see if the math checks out, and often it doesn't. So, because we have the trades, we can compute the candles ourselves.

So trade data is better than candle data, is better than ticker data, which is the worst and this is what our data set looks like: We have raw trade data and from those raw trades we can construct candles and from those candles we can construct tickers and that's for exchanges that do have raw trade data from. If an exchange only provides candle data then we will get the candle data and will calculate tickers but we won't use their tickers—we'll calculate them ourselves. And then the worst case scenario is you're in exchange that only provides tickers.

I think the beauty of our data approach is that we have a database that allows raw trade data to coexist with candle data to coexist from ticker data as the primary source data from exchanges and we inform you about what kind of data you're getting and how the numbers that you're asking for are derived from these data points.

So if an exchange has great data we'll get it and if they have terrible data we'll get that too because people often do want data from these crappy exchanges. So we'll log it all—whereas others often only have tickers from exchanges. In other words, they're ingesting tickers and then constructing candles from those tickers and that's something that I think is pretty important to talk about. A lot of our competitors, what they're doing is they're ingesting tickers like ticker feed data in real time and they're constructing candles from that.

So let's say you want to construct a 1-minute candle and then you've got 24-hour tickers coming in so a ticker is basically like a 24-hour candle that you get whenever you get it—whenever it's computed—it isn't computed on specific time intervals that you can rely on. So let's say you're ingesting data from an exchange that only provides ticker data (that's all that they do) and you want to construct a 1-minute candle. Well, let's say that only one ticker comes in during this one minute—that means that the open/close high and low for this candle is all the same price. Similarly, let's say you want to create a 1-hour candle and you've got the steady stream of tickers coming in. You know whenever they send them to you, well, you can't use. So, let's just go all the way down. So let's say you do luck out, you hit the lottery and you do get a ticker that gives you a data point at the exact time of this candle opening and let's say you get some additional points that you are going to believe are the high and low. The low at least they're the highest and lowest prices of the ticker points that you have—which are not a lot—during this period and let's say the last ticker you get before the close of this candle is at 2:50 257. Well, you have to just taken this price that you got at 2:57, just assume that it's close (if you are constructing tickers from candles), which is generally a bad idea. This isn't how we do things.

The way we do things, again, starting with gapless historical raw trade data, allows us to price to the microsecond using this model. So, anyway, there's a lot I can talk about here.

I think it's probably worth discussing a little bit our transparency ratings. So you might recall that Bitwise put out a report where they identified 10 exchanges that they say have actual volume and in order to do this analysis they looked at Bitcoin to USD (BTC/USD) and Bitcoin to Tether (BTC/USDT) markets and they looked at 80 exchanges. They did not look at exchanges that did not have Bitcoin to USD in Bitcoin to other markets and we were looking at this data and we found something interesting... We found that of the 10 exchanges that were deemed to be trusted by Bitwise, that 8 out of these 10 exchanges provided historical gapless raw trade data. And why would that be, right? I think the reason this would be the case is that just like the IRS if you provide a lot of data and you're doing something wrong you're likely to be caught. So we have found that providing historical gapless raw trade data is correlated with being a good exchange. And then of the exchanges that Bitwise identified as being suspect, that they explicitly called out as being suspect, all but two of those did not provide historical gapless raw trade data.

We care quite a bit about how we approach data. I can tell you a little bit about our data services. Basically, we can create customized endpoints for you. Often, there's analysis that people want that requires them to download a whole lot of data and then analyze that data and often—because we have all the data in our database—we can just give you an API endpoint that just outputs the number that you're looking for that just sort of does the analysis for you. So, that's one of the things that we do.

Let's start off with the first one. I'm not going to go through all these slides but we do custom asset pricing so what we found is that a lot of hedge funds and funds that calculate nav for investors, that they want to calculate prices according to a specified methodology. So they might say, "We want to calculate prices based on only these ten exchanges and even just in and only based on Fiat pairs on these ten exchanges," and so they specify and they want to "calculate end-of-day prices based on the end of the day" in their time zone. Let's say they're in California... Then they would calculate these based on end of day prices in the Pacific time zone...

Another thing that we do is we provide low latency data. So if you need super low latency order book snapshots and trading data, that's something we can do. We can get order book snapshots down to 100 milliseconds.

Another thing that we do—and this is more for exchanges—but we can power white label market data API. So if you're an exchange and you do have a data API, we can run that for you.

And, finally, we can stand up market data websites for you. So let's say you have an investor portal and you want to give your investors like, you know, real-time access to what's happening with the price of a whole bunch of different cryptocurrencies and you want to give them real-time access to maybe an index or prices on the exchanges that you guys
or gals are trading on, then we can do that for you.

Yeah so anyway, I kind of went through all [these slides] anyway...

On behalf of all of us, thank you.