Using On-Chain Metrics to Visualize Cryptocurrency Network Behavior

Transcript

is dino it says here in content and seo i guess that was in a past life i currently do sort of i wear several heads sentiment a little bit of business development a little bit of market analysis i've been with sentiment for more than two years now um and yeah today the topic of this discussion which i really hope is going to be more of a discussion than a presentation so feel free to you know raise your hand and ask questions if there is anything that you don't really understand while i talk because i will be sort of introducing some new concepts and some new metrics and the way that we visualize uh cryptocurrency behavior so anything that you feel like i could provide a little bit more clarifications on just let me know so the topic of uh today's presentation is how to use onshore

metrics for cryptocurrency behavior analysis and now let's see if this works it does not all right good um okay am i doing something wrong maybe you can just switch it manually i mean this can work oh

sorry yeah all right beautiful thank you yeah all right so uh a few words of introduction about sentiment for those of you that are not familiar with the project so sentiment was established established in late 2016 uh maxine velasquez our founder he essentially back in 2016 thought that uh the cryptocurrency market in the way

that people analyze cryptocurrency market is fairly deficient uh and effective uh mainly because of the fact that back in 2016 and very much so today 201 a lot of what people and most sort of cryptocurrency enthusiasts use in order to uh contextualize cryptocurrency behavior is they use price and volume data so financial data usually taken from a variety of cryptocurrency exchanges aggregated into a single data so uh the great thing about cryptocurrencies is that we now have a wealth and breadth of new data sources alternative data sources with which we can sort of figure out what's going on with any particular network we can gauge the health of a cryptocurrency project

from a variety of different perspectives that weren't really uh available in traditional finance right so i'm not only talking about data directly from the blockchain i'm also talking about data uh from thousands of public github repositories that nearly every cryptocurrency project uh provides which can allow us to more easily visualize just how active a project is from a development perspective and that sometimes points to you know specific um you know specific ways to analyze the health of a project from a variety of perspectives so very briefly the main goal of sentiment is to use alternative data sources and we have three that we have sort of built our entire platform around to identify and then clearly track or visualize the

behavior of what we believe are key stakeholders for any cryptocurrency now this information can obviously be used for trading but also it is used our customers range from um quantitative developers and engineers to retail traders uh large digital asset management platforms cryptocurrency exchanges journalists anybody that wants to get a better sense of or better sort of context about everyday market events and the health of a particular network so for us we are sort of focusing on several large key stakeholders anything from the retail sector or the crowd to cryptocurrency developers the so-called whales or large addresses which usually have a disproportionate impact on the uh

health and the price performance of a particular cryptocurrency exchanges as a major factor today because most cryptocurrencies are still sort of largely speculative in nature retail traders early adopters as well as long-term holders of a particular cryptocurrencies uh team and infrastructure wallets and currently or more recently liquidity pools nfds and all that uh sort of you know beautiful new market narratives that are emerging in 2021. so over the last three and a half years we have developed about 100 different metrics uh some of them unique for the first time sort of built exclusively in-house by our team when you uh use all of the sort of versions of our different metrics and you compile them together that comes out to about 800 of them

obviously i'm not going to be talking about all of them uh today we're hopefully going to cover about five or six of them that i think um give a pretty good sense of how we use on-chain data specifically to visualize the behavior of a particular network and then what to do with that information if we have some time in the end i will also try to show a little bit about our social data as well which is uh still sort of an untapped sector in the cryptocurrency space uh and you know we are one of the first companies or first projects that have for the last three years been doing a lot of social listening monitoring about 1500 different cryptocurrency related social media channels and that data we have used to create a bunch of new metrics and indicators that seem to have uh

quite an interesting correlation with a coin's performance over time all right so we're to start with uh something that's a little bit of a mouthful uh so basically i'm actually let me go back a little bit um the first so so i mentioned three sources of information that the entire sentiment platform was built around these three sources of information are ong data social activity or social media data and development activity data so mostly today i'm going to be talking about on-chain data we collect launching data directly from the blockchain nodes there is no intermediary between the sentiment platform and the blockchain itself we currently provide on-chain data for about 1500 different cryptocurrency projects which to my knowledge anybody can feel free to correct me if they can find one uh we

are the most comprehensive platform in terms of on-chain coverage so our uh cryptocurrency launching coverage includes bitcoin ethereum all erc20 coins or coins using drc20 standards in other words coins that are built on the ethereum platform ripple litecoin uh eos and we are now introducing polka dot as well as cardano which have been sort of very uh you know popular by our with our community uh and we're constantly trying to expand into new blockchains but obviously uh you know the infrastructure for building and scraping data from every specific blockchain is quite resource intensive so for now we are focusing on what believe will have the most sort of longevity on the market all right so um i'm going to start with something that uh we have developed

exclusively at sentiment it's called the mean dollar investigator i think it's although it's a little bit of a mouthful i'll try to explain it to the best of my abilities and then feel free to ask me any follow-up questions but i think it's a good uh sort of macro indicator to give you a sense of how we use on-chain data to try and visualize what is going on with cryptocurrencies and its network in a very sort of macro way and from a macro perspective so mean dollar investor age is a metric that we've developed back in 2019 and essentially what we do is we're gonna use bitcoin uh as an example but the same sort of patterns work for any cryptocurrency right so the way that this metric works is we collect data from the blockchain and then

we try to determine how long each unit of bitcoin has stayed in their current address and then we compute the average age of all those bitcoins so by age i mean the amount of time that a particular unit of bitcoin has stayed in their current address let's say uh i have acquired 30 days ago so that unit of bitcoin has an age of 30 days right so we do that we determine how long each unit of bitcoin has stayed in their current address in their respective address we compute the average of all those ages for all units of bitcoin in existence and then we adjust it by the price at which each unit of bitcoin was last transacted on the blockchain so i know that's a few different sort of relatively complex

things but the end result is very simple and maybe it's best that i just show it on a chart and show you what we believe it uh represents so this purple line is mean dollar invested age for bitcoin uh for the show showing the last i think seven or eight years of bitcoins uh life cycle and then in green obviously is the price section of bitcoin which has been uh pretty impressive so what generally mean dollar invested age does is when you see a rising slope in this metric overall as a rule of thumb that means that the average age of all units of bitcoin is rising meaning that units bitcoin are becoming older or in other

words they are not as actively being moved on the blockchain in other words most bitcoin investors most bitcoin holders in this part of the cycle um are content with sitting on their backs there's there's quite a lot of value accrual and value accumulation not a lot of value redistribution so this is basically a period where the uh network-wide sentiment is mostly tailored to hodling or holding cryptocurrencies rather than uh moving them between addresses right and then when you see these large drop-offs or declines a declining slope in bitcoin's mean dollar invested age this is a period in time where the average age of units of bitcoin is becoming lower in other words uh bitcoins are becoming younger by which i mean simply they're

being more actively transacted between addresses right and as a result this for portion of bitcoin's mean dollar invested h chart usually points to periods of redistribution um of bitcoin from you know stakeholders a and b whatever they might be to stakeholders x and y so overall what's interesting with bitcoin's mean dollar invested age is that as you can see several times over bitcoin's live cycle in terms of its price action there have been periods of large declines in bitcoin's mean dollar invested rate charts which have coincided with the sort of start or a mid part of bitcoin's bold cycle so the first thing happened that actually the first time it happened was back in 2013

although you can't really see it just because well has gone crazy since um the second time it happened was in 2016 the next time it happened was in 2019 and then the last sort of big drop from bitcoin's million dollar investigations happened in 2020 so essentially like i said most of these large uh drop-offs points to a period of time where there is a massive redistribution between addresses we believe that in 2013 this first drop-off or value in bitcoin's million dollar investor age has pointed to a period of redistribution from sort of early blockchain nerds to first wave of cryptocurrency enthusiasts in early adopters back in 2017 this was the first time that cryptocurrencies really saw or sort of taken center stage so this

points to a period of redistribution from those early adopters to the first wave of mainstream cryptocurrency investors and then lastly there's large drops in 2020 we think uh coincides with the general market narrative that in 23 the um institutional investor has taken a a huge interest in bitcoin in particular but cryptocurrencies in general and this sort of coincides with this massive redistribution from sort of weaker hands to large institutional addresses and later on i will show a few additional metrics that seem to uh confirm or validate that theory for us so this is how overall bitcoins mean dollar investigation works and then on the other side of that point we also see that several times as bitcoin's million

dollar investor age sort of uh hits the bottom and then moves into the opposite direction or starts uh sort of um charting an upward trade trajectory it seems to coincide with the period of price depreciation for uh the price of that point so we've seen it back in 2017 back in 2019 as well as back in 2020. so here we think what typically explains this is that the redistribution cycle is over most investors are now moving into dormancy and overall the network-wide sentiment seems to be moving into a little larger sort of more of a lethargic mode there's not a whole lot of value being created on the network anymore there's

not a lot of value being transferred on the network anymore so from a fundamental perspective um there's not a whole lot of fundamental justification for the rally that's going to continue because value is sort of sitting on the sidelines right in general you can see very similar patterns in ethereum's mean dollar invested age um what is uh obviously sort of great but also our drawback for a bit for million dollar investor agency metric is that it works best for coins that have a lot of history so you can sort of clearly visualize market cycles right so this isn't going to work for i don't know if anybody's interested in shiba inu or dogecoin uh but it's probably not going to work for those uh simply because there's just not a whole lot of history uh for a lot of meme coins as of yet there will be other

metrics but not this one in particular all right yes please decisions so one second so we've written a couple of articles uh back when uh bitcoin's mean dollar investigation sort of started moving so we sort of measured the uh divergence of the bitcoin dollar investigator in the price and we've written i think one article around 2019 where we basically said that it seems like the mean dollar investor age is moving into a bearish direction and you know several months have passed uh the sort of coincidence with this

with this conclusion of ours so uh in our opinion this isn't necessarily a lagging indicator but can be used in real time we haven't done a uh sort of create a strategy or a back test of this particular metric we have done for several others um but yeah there are people in our sort of uh institutional uh some of our institutional customers that use this metric particularly to rebalance their bitcoin portfolios um yeah i have a question of course also about the i mean bitcoin is obviously a great source of storage of value and it's not really a currency that has a lot of utility yes uh what what is what is

a chart what this will look like for ethereum or even like bnb where you can use for different kinds of defy and ft's there's a lot of movements great question um even though in general it seems like uh i would love to share it uh live the chart for it here maybe after the presentation i can show you um but in general it actually seems to follow fairly similar patterns so even for coins that have not just a speculative purpose but a lot of sort of utilitarian use cases i think that's mostly currently ethereum and a few other coins but most of them are still fairly speculative in nature it seems to follow the same patterns where eve you know their sort of the utility uh line is sort of the same or rises but there are still periods where on average

on a network level um holders are moving into a more dormant period or dormant sentiment uh and then there's a period where both holders are redistributing speculators are more active as well as the utility of the network is rising so it sort of accounts for it um in both ways it's a fairly similar chart overall um yeah yeah i'll be happy to sort of share it after that do your other metrics matrices do they account for some extent as to what's causing these fluctuations because that's fairly straightforward it's you know it's strange enough to see that but then really take into account sort of macro political circumstances regulators said how how how can we

relate what you're showing us there to what the influencing factors are of course so that's uh a great question um we try to so we try to uh sort of conflate or juxtapose these metrics to a bunch of different uh metrics that try to explain different parts of the uh cryptocurrency behavior so for instance uh what i'm going to show in a little bit is that this sort of period of or massive drop-off in bitcoin's mean dollar invested has coincided with a period of fairly large accumulation by addresses holding between 100 and 10 000 bitcoin so that's sort of large between millionaire segment of the bitcoin audience so there obviously when it comes to unchained

data there are uh inherently some assumptions to be made we can't account for every single uh transaction and know its full intent but the best we can do is that uh we can use additional metrics specifically from the blockchain uh to account for what is the most likely reason that we are seeing massive drop-offs or massive uh inflows so for instance like i said uh massive accumulation large addresses has coincided with this period as well as generally very large growth in the total addresses that are interacting with bitcoin on a daily basis which is sort of flat enough for much of 2019 and then it started peaking back in 2020 suggesting that there is sort of a new rejuvenated interest in bitcoin as a

network on the whole so i have a few of those metrics that i'll be um happy to share as we go along um any questions more on bitcoin's million dollar investor age i hope that how much time do we have i usually lose track of it yeah it's been 20 minutes all right so we might want to go into a few more of those and then you can feel free to tell us about the future yeah uh great question um i'm not sure so generally speaking what this seems to be telling us at least at this point in time is that uh there has been a a general sort of general uh paradigm shift in terms of bitcoin's holders uh once again towards dormant you know this has changed a little bit uh over the last

maybe a week or so and this uh i have a few other metrics that sort of seem to explain why but overall um from a uh fundamental perspective we are not a big fan of periods of time where uh money is standing on the sidelines and not a lot of value is being uh transacted on the network uh specifically for bitcoin but also ethereum and other coins we think that uh periods of dormancy on the network and periods of low value transfer you typically coincide with a um sort of less of a push or less of a potential for massive sustained price rallies so i'm not surprised i don't think you know bitcoin can go up and down but from a sort of macro perspective uh if this sort of

charge continues to trend in the upper trajectory i would sort of be a little bit concerned for a couple of years for bitcoin's price action obviously it's all cyclical but it does seem to point to um a period of concern for some bitcoin holders not investment advice obviously um yeah so the next metric that i want to talk about is probably uh the most popular orange metric right now it's called the mbrv ratio i'll show what it looks like here before we go into explanations uh it has been created back in 2018 by nick carter and like in very laymen terms what indian ratio does or tries to do is it

tries to compute the average profit or the average loss of a group of addresses of all or all addresses that have acquired uh tokens within a certain time frame so i can explain how that works as a simple formula so mvr ratio stands for market value to realize value ratio so it's a fairly simple simple formula where you just uh divide two uh parameters so you take the market valuation or market cap of a coin which is simply total supply of a coin times the current price of that coin that's how you get the market uh cap of a coin and then you divide it by the realized value of a coin so realize value takes the total supply current supply of the coin but it multiplies it by the price at which each unit of

that coin was last transacted on the blockchain which the assumption is was the time at which they were acquired so it's basically the difference between the total amount of money that would currently take to buy let's say all of bitcoin versus the total amount of money that was taken to purchase all bitcoin in existence right so as a result we get this sort of uh ratio and for example when i made this chart before it was still surrounded at a 65k uh range and it was at the the 30 day of the yearb ratio was at 18.12 so what that points to the 30 day mbrv ratio essentially points to the average profit or the average loss of all addresses that have acquired bitcoin in

the 30 the previous 30-day span so it's basically looking at the average profits of short-term a bitcoin holders at 18.12 that means that the average profit of these swing traders or short-term bitcoin holders was uh plus 18 they were plus 18 on their initial investment on average right so what does this tell us or how do people generally use this metric right let's move back so the way that function analysts use this usually is that they believe that because as nprv ratio rises that means that there are more unrealized profits sitting on the sidelines by a specific category of coin holders so as nvrv ratio rises the

higher the mvrv the more unrealized or unclaimed profits there are and as a result the more likely it is that holders are going to start claiming those profits at some point in time and applying massive cell side pressure so in general the higher the nvrv ratio the more un overvalued uh and potentially risky a coin is and then the lower the mvrv ratio is on the other side of the aisle the more undervalued or the more open to time it can be to accumulate once again not financial advice um so what with that sort of context what function analysts then do is then they used a bunch of data historical data and they look at historical peaks and valleys and we have done some vectors on this that i'll show in a moment where they

essentially look at periods in time where uh bitcoins or a point every ratio tends to peak both in the upper direction or a lower direction so what are the peaks and valleys of incur of a specific mvrv ratio to look for historical trends inter or historical danger and opportunity zones for a particular coin so for 30 day of verve ratio what we see is that there is a tendency or there has been a tendency for bitcoin's mbrv ratio over the last three to four years to um peak between 15 and 20 there are several short-term outliers but even those last for no longer than several days and generally points to it overvalued

territory or sort of unsustainable price rise right um so between 15 and 20 seems to be the average uh sort of dangers for bitcoin short-term holders or that seems to be a period in time or sort of profit margins at which short-term bitcoin holders seem to like to take some profits off the table uh between 15 and 20 and then on the other side side of the aisle around minus 15 all the way to another support level at minus 30 percent these seem to be two levels in which bitcoins third of the 30-day mvrb ratio seems to historically bottom out and points to a potential for upward price uh recovery now we have done some backwards oh but let me show this one as well so this is the 365-day

mvrb ratio which is showing the average profit or average loss of all addresses that have acquired bitcoin in the previous 365 days so it's sort of a longer term uh view or gauge of investor sentiment um and overall i think we're starting here 2017 there was one even larger i believe in 2014 so the mba ratio especially the long-term mbmb ratio has shown a tendency to sort of flatten out over time which seems to point to a general um mature maturation of the cryptocurrency uh speculation so with every subsequent market cycle it seems like the uh highest level of nvme ratio or the highest profit margins or the levels at which uh

both long and short-term holders like to sell seems to uh out a little bit pointing to more of a risk of on behavior with every subsequent market cycle is people seem to no longer expect um outside returns like they did in 2017 but are you know completely fine with state taking or starting to take profits at relatively lower levels we're still talking about 200 or 300 x right from their initial investment so this is a bad test that we've done um i believe about 12 months ago where we wanted to see if there is a possibility to create a real-time model where you would use an nvrv ratio to determine when to uh sort of enter into

a trade and exit out of a trade right so what we have done is we have taken the entire history of bitcoin and we have split it into 30-day chunks and in each of those 30-day chunks we have looked at historical sort of local peaks that is in local bottoms for a particular mbrv ratio here we're looking at a 30-day of derivation so for each 30-day chunk we look at the the local extrema or local maxima local minima and then we compute the median values of all of those local peaks and bottoms and we get two values we get the median peak and medium bottom for an nvrv ratio and then we use those two values and as signals for to buy and to sell actually to sell and to buy so when bitcoin's

nvrv ratio hits this predetermined threshold based on historical data if it hits a peak mvrv threshold we sell and then if he hits a bottom mvary threshold we would buy simply using on-chain or sort of crypto native metrics so this was these were our fur attempts at trying to visualize this as you can see there are several interesting uh positions where our model was able to sort of recognize that the market is becoming a little bit uh overvalued for its own good and gonna hand periods where the market seems to be um sort of overselling and looks uh a little bit too frightened for its own good right that being said obviously it's not a perfect indicator for example we would trigger a signal here even though uh

there would be another sort of more uh or better or more opportune time to accumulate so we're still tinkering with the thresholds uh with what would make the best possible um real-time trading strategy but we've actually created an index uh that is based specifically on this metric where you can rebalance your portfolio uh of as many points as you like by looking at a 30-day mvrv ratio and rebalancing when the things become too heated or things become sort of too undervalued i think to have another example this for ethereum so same idea looking at the 30-day mdrv ratio these were would be the times that our real-time signals have triggered and then these would be the times where our

real-time buy signals have triggered as well so a few interesting sort of uh ways to visualize cryptocurrency network behavior and also maybe create a models to um you know try and get a get a get advantage in the cryptocurrency trading space specifically using uh crypto native metrics or the behavior of the actual network so any questions on this before we move forward yes [Music] uh no this actually uh it's it's not uh it's the uh not the weighted average it's just the standard

average we don't there is a metric that uses a weighted average it's called the mvrv z ratio but we've so far had better uh success with just using the normal average for for these uh for the uh vectors that we've done all right now we're actually going to get into it's a weird way that i've made this presentation because we're getting now into some something a little bit more simple so maybe we should have started with this um this is um what we like to call uh price day divergence so dna stands for daily active addresses and i think most people here know what daily active addresses sort of represents it's one of the most basic metrics that you can use in figuring out the performance of the

behavior of a network daily active address that simply stands for the total amount of unique addresses that are involved in transactions or that are interacting with a coin on a daily basis by interacting with a coin i simply mean uh buying sorry um yeah buying or selling uh a particular cryptocurrency on a daily basis or just moving it on the blockchain so based on both uh bad tests that we have done and just anecdotal evidence it seems like um sustained price rallies typically go hand in hand with periods of large growth or sustained growth in daily active addresses which a lot of people watching analysts use sort of as an on-chain proxy for the amount of demand in a coin right as daily active

addresses rise it seems to point to another sort of wave of fresh blood as well as another period of uh value accrual um and value redistribution within a network so that sort of makes sense that from a fundamental perspective you you would like to see a price growth coincide with growth in terms of just the net effect of a particular coin right now what's interesting there is that we have then tried to uh we have seen sort of the opposite trend as well so we have seen periods where it coins let's say bitcoin uh bitcoin's price action continues to rise however the price or or sorry daily addresses or let's say network activity or address activity starts to flatten out or decline so in

that point of time we are seeing a sort of divergence between the uh price or market valuation of the coin while the total value transfer and the total network fit in the coin is declining and for us that seems to generally point to a bearish divergence and a potential for um you know local correction as the crowd as the market seems to be a little bit overvalued and then the other thing also applies or sort of vice versa also applies when you see periods of price depreciation for a coin but sustained uptick in its daily active addresses and on chain activity that also seems to point to a bullish divergence where the network seems a little bit undervalued relative to the amount of activity and the amount of network effect it is currently accruing for itself

so we have then uh back tested this a couple of years ago um we i have a chart that shows what's happened since um we backtested this idea a couple of years ago where okay this is the chart that i want to show where we basically done a similar thing with the mvrv ratio so we looked at a difference between a coin's price action on the one hand and the coins network activity on the other hand so if there's a massive uptick in a particular points price action but there's a considerable downside or sort of a lethargic network activity that sort of bearish divergence will trigger a sell signal in real time and then if the price of a coin is declining but its network activity seems to be on a

sustainable growth that would point or sort of trigger a buy signal so these would be some of the signals uh the first sort of attempt of a uh price the divergence model where we would sort of sell at these levels and then we would buy that into a coin at these uh orange levels so several times that we were able to sort of uh see especially back in 2018 just how sort of overvalued the network all of a sudden become even though the price was rising like crazy the network activity seems to have largely uh been unable to sort of keep afloat or sort of keep up with the price this is what this the divergence model looks like on our platform so we have all of these

charts are live uh currently on sentiments website and you can just log in and see all of them for yourself um i'll share the link at the end of the video so this is the this model applied in real time on our bitcoins price action so the these large uh dropbox would be periods where uh relative to its price action bitcoin's network activity has declined quite heavily and then these green periods or green bars would be periods where relative to its network to its price action bitcoin's network activity or address activity has started to really blossom this was in 2017.

this was in 2019 as you can see we have been able to catch several uh sort of local tops as well as local bottoms for btc however this is what's happened in 2021 so it seems like we need to adjust our thresholds again because where exactly do you sell here do you sell here do you sell here like right what do you do do you buy here or do you buy here so even though there seems to be definitely value in the idea of a divergence between a coin's price and coins in the daily active addresses it seems like over time the threshold defining when a potential signal buy signal or sell signal uh should trigger needs to be

adjusted so this is what oh okay uh yeah i thought i had it for ethereum as well yeah man we can show it on the chart later um so this is what the this sort of divergence looks like and we have a similar metric that sort of works in the same way uh called the nvt ratio and i promise this is the last metric that i'll show i know usually uh after four or five for people just sort of zoom up uh um i've been dead so i know um almost there so this last metric that i wanted to show is sort of the takes the same idea where it compares a market valuation of a coin versus the network activity of a coin except here the mvt ratio

it's an auction surrogate for what in traditional finances you know you used to know or is known as the price to earnings ratio where you take the market value per share and then you divide it by the earnings per share to sort of see whether a stock is all valid or undervalued the problem with that and applying it to cryptocurrencies is that bitcoin doesn't have a balance sheet so you can't really track its earnings right so what do you do well a few uh inventive individuals have tried solving this problem by using on-chain data and on-chain activity to account for the lack of you know transparent balance sheets for a particular coin so what we do with the nvt ratio is we divide a coin's market valuation by the total amount of unique

tokens moving on the blockchain daily so what that sort of tries to do is it tries to compare the total market cap or market value of a coin to the daily value transfer of that coin which we are using here as a surrogate for earnings right and so when the nbt ratio is high uh it suggests that a coin's market valuation is too high relative to the amount of value being created on the network daily and then when it's too low uh it means that it could be potentially undervalued because market valuation is fairly substandard or some prime relative to the amount of unique coins moving or being in circulation on a daily basis on the network so this is what our mvt ratio model looks like and so apparently at least we are having

some signs of of health in between surprise section and then on chain metrics um but yeah it sort of points to uh this is something that we've done using several weighted moving averages where we've tried to sort of show periods in time where bitcoin's price action is too high relative to its um token circulation or general amount of uh the daily value transfer of the network and when it's too low so when these red bars point to extremely high market valuation relative to network activity and then these low bars going to relatively or extremely low market valuation relative to the amount of value that is being created and transferred or redistributed on the network on a daily basis so we were at

the very least able to sort of the first time our energy ratio has uh started growing green was back in october of 2020 which also coincides with that uh let me move to a bunch it's too too slow i wanted to show the mean dollar invested h chart uh once again because you will see sort of a period the the middle resolution started to drop off exactly here so at the point in time where there was all of a sudden a massive amount of new units of bitcoin being daily transacted on the blockchain this is the point of time where we saw the twenty dollar nest page start to drop as well and then in march 2021 um from a fundamental perspective this was a period of time where bitcoin started first exhibiting

some concerning signs or concerning patterns where uh when it could all come higher relatively i think even about a month earlier before it could all come high um its price action was no longer sustainable at least from a fundamental perspective so based on the total amount of uh value accrual and value creation on the network its price actually was no longer fundamentally justifiable and so far it seems to have been on point so like i said this only takes into account this last green ticker only takes it took out about a seven day stretch to keep it with a grain of salt but generally speaking what we like to see is we have to see periods of increased growth in terms of on-chain network activity and that means both in terms of the amount of daily addresses that are interacting with the coin as well as the

amount of unique coins that are moving on the network yes the hedge is simply sold when i see the first non-green candle yes doing really really well you would be doing okay we actually uh haven't uh done a real-time model on this one like we've done with the mdrv ratio and price the a divergence but this is a fairly popular model in on-chain uh we are not the only uh company that has

it we are the only people or sort of company that has created this model which uses i think four different moving averages for nvt ratio rather than just a total difference between market evaluation and token circulation but you would do okayish uh with the with this this uh discharge or this metric so generally speaking the whole point of this exercise of this presentation is showing several um you know ways that we only look at potential investment opportunities to do diligence on a project based on crypto native metrics nothing else no sort of uh price and volume related data order book data so obviously i think uh that's a relatively limited approach we're not against using technical analysis and price and volume data but we believe

there's just a lot more discurrently still being sort of unused or underutilized in terms of potential metrics to try and uh beat the market which is what everyone wants to put in um if you're interested i will show you later on uh the reasons why i sold my btc at 65k uh three or four indicators that have sort of started flashing um on on our platform that made me i want to at least for a while get out of btc and then maybe get into it at the later date so yeah feel free to think there are some secret questions do you have any kind of ensemble decision making [Music]

so the question is do we have sort of an ensemble uh basically a bunch of a bundle of metrics where we just kind of put it into an algorithm and then it says buy or sell right we haven't no uh and in general what we like to do is we like to take uh sort of or just a pose or compare all of these metrics individually rather than we were not able to create a simple algorithm to point to use all of them into a sort of a uniform uh uniform way generally speaking the way that we and our community and our users use this metric is they look at them on a daily basis it takes a bit of time it's not a a very automated process but um yeah usually i like to look between like seven to ten

metrics on a daily basis to get a general sense of what the uh current network activity is like um but yeah if you could create a uh sort of an umbrella model we have tried a few of them uh they've performed oakish at times but nothing that would like massively beat the market uh based on all of the metrics in in um in our repertoire um okay so i have a question yes did your very clever um uh ability to to hit 65 to bring it out yes did that also um allow you to come back in for this this last 30 days and and yeah so to achieve that so not at the moment i

mean i'm still sort of out of bitcoin if that's your question right now yeah but i've also been traveling a lot so i don't know which one is uh the culprit for it um but yeah i mean bitcoin only hit 65k what like a week ago maybe 10 days ago so that's when i oh i see okay yeah yeah that's right no the previous one no this new all-time high yeah it's not always that i sell the tops but when i do i like to brag about it um so yeah all right so the last book uh that i wanted to bring up i guess we'll have no time for social metrics which are super interesting so if anybody's interested in more information about non-on-chain sources and how to use them to figure out what is going on with the markets please bring me after after the

presentation um the last sort of metric that i wanted to talk about is the behavior of large addresses and i only have one example here but um there's more in the um document that i've prepared for the for the presentation so generally speaking um we are of the opinion that uh you know large network addresses still have a rather disproportionate impact on the price performance and the general market performance of the coin uh there was an article that was on front page of reddit yesterday i believe which said that uh bitcoin is largely had by a minority of large addresses and sort of concerted uh interest groups rather than sort of a very uh decentralized and distributed um

coin so based on a recent report that was done by analysis uh there was a massive amount more than 60 percent of d5 transactions so all decentralized transactions in the first uh and second quarter of 2021 have account have come from these larger addresses and addresses that hold more than 10 million and actually transactions that are worth more than 10 million so it's not it's not very much cryptocurrencies today are still very much a whale's game so it then makes sense that we would need to look into the behavior of these whales and addresses to figure out uh what they're doing and whether there are any patterns that we can spot now this is a very difficult thing to do it's not as easy as it might seem so this is sort of a most basic way that we do it um on the sentiment platform what are we seeing

here so this current chart shows the total holdings of addresses that are holding between 100 and 10 000 bitcoin so quite a quite a strategy but these are all sort of millionaire addresses versus the price of a coin of bitcoin over the last what is it like a year uh a year of data so um generally speaking we have seen uh back in 20 uh back in the august in september of 2020 we have seen sort of the first um signs of large accumulation by these uh massive addresses so anybody that owns at least 100 bitcoin or more started to sort of accumulate more btc this was about 250 000 bitcoin being accumulated but these addresses in a span of about

two and a half months so a massive amount of new btc being scooped up not by retail holders but by these large uh you know possibly institutional addresses and then here even though we can't really see it super clearly there have been periods where uh there are sort of a short-term offloading events so periods of entire where there is a decline in the total holdings of these addresses this was about 60 000 btc 60 000 btc as well here around the old time high here about 40 000 bitcoin so sort of short-term potential broadly taken activities by these large addresses and then over time as bitcoin started to correct uh there has been a sort of left side of uh lack of confidence by these large addresses since then however we

have seen a another sort of accumulation effort by these addresses particularly over the last month and a half so we tend to use this metric as sort of a very general gauge of whale sentiment now there are obviously several caveats here first of all not all addresses that are large are wheels specifically there are a lot of massive addresses that belong to cryptocurrency exchanges right so if you are just accounting for all addresses there is a possibility that you are also taking into account um like binance wallets kraken wallets coin based wallets and then if you're seeing a accumulation in those wallets it's a

very different context than if you're seeing an accumulation in whale addresses right if they if i as or somebody as a just individual large address accumulates a bunch of bitcoin that's a potential bullet signal but if there's a lot of bitcoin being moved to coinbase that could be a signal that there's a rise in sell side pressure right so this is why usually we don't like to look at any addresses worth more than ten thousand are holding more than ten thousand btc because there's only about a hundred of them and they're mostly large cryptocurrency exchange wallets and the second thing and this is what sentiment is sort of working on uh very hardly if that's going to work in the last six months or so is we're working on improving our labels so this is where uh it's not no longer enough to just scrape

data directly from the blocks you need to call it a day um for us and for as like with every subsequent mars cycle the market cycle there are additional complexities that you have to account for when it comes to understanding engine data to really understand what um whales are doing you need to understand you need to properly label uh hundreds and thousands of wallets you need to go through each of those wallets individually and you look to for patterns or heuristics that will tell you how likely it is that this if this wallet is already not labeled as an exchange wallet how likely it is that it behaves as an exchange wallet so there is a level of assumptions that you need to make when doing this type of analysis but generally speaking this is what we

are uh the sentiment team is sort of working on um and putting a lot of resources and effort over the last uh six months and this is where we think is not necessarily the holy grail but generally um what is going to sort of uh distinguish us from most market competitors is having the best labels on the market so that you can uh contextualize the behavior of every specific stakeholder and know exactly like whether you know these specific groups that you can filter these specific groups of addresses by a label so um you know for example show me all of the addresses that hold between 100 and 10 000 bitcoin but then filter only those that belong to individual whales filter those that belong to cryptocurrency exchanges those

are those that belong to yield aggregators farms and so forth so we already have a prototype for this metric um it's on our platform um yeah i'll be happy to show it to you after the the presentation so this is these were five different metrics so mean dollar invested age nvrv ratio price va divergence nvt ratio in the whales like i said we currently have around 100 different metrics on our website i'm not going to pretend that all of them are uh equally effective uh they're not and not all of them are going to tell you what's going to happen with the market or what's going to happen with the market in the next week there are some indicators that we believe are super valuable even though they have nothing to do necessarily with short-term market uh volatility for

example we are very proud with the development activity metrics that we've created where essentially we take data from thousands of public github repositories from cryptocurrency projects and we showed we filter or sort of clean that data where we filter out any non-development related events in the projects github repositories to show what is the level of investment that cryptocurrency projects are putting into their uh project on a daily weekly monthly yearly basis how much are they investing in coding manpower in particular because at the end of the day developers are probably the most valuable assets in cryptocurrency projects is that investment growing over time is it dependent on a coins price action or is

it completely sort of independent of it and for us the best projects on a very long term scale are projects that show sustained growth in developer development activity regardless of market cycles regardless of price action uh you know teams that are not really looking at the market at all but rather just continuing to push code and continue to continue to build so that sort of metric won't help you to determine what happens with you know bitcoin or ethereum tomorrow but it can help you understand uh from a more sort of due diligence perspective whether the project that you might be interested in investing in is sound and whether the theme is sound so i'm going to skip over all of the social metrics very sad yes uh i'll be happy to share this

presentation with anyone that's uh interested um yeah this is sort of the end because i think we're already about 55 minutes in yeah if you are interested type of data uh several things that you can do first of all check out our website app.sentiment.net is our token intelligence platform so everything that you've seen right now or or in the last hour for bitcoin in particular we have this data not just for btc but for ethereum all your c20 coins and a variety of other blockchains that we're currently integrating even if we don't have on-chain data for a specific point we still have several unique sources of information like social activity data and development activity data then you can engage uh on our platform you can also tap in if you're interested if you're a quant or

somebody that has a you know knowledge of back testing and creating your own trading models you can tap into our api use all of this beautiful raw data to create your own dashboards models and try to beat the market um so yeah that's it for me i hope you found this useful although i'm sure that a lot of it was uh especially after the first three metrics maybe went over your head i'll be more than happy to stay here or you can find me at the booth later so we can talk more and yeah if you're interested in a few indicators that i use to sort of sell btc at the 65k um yeah you can come here one of your questions which metrics do you look at it do you look at personally so you would uh or you would advise your family yeah

so i never advise my family and friends just because i'm super risk averse so i just would feel at the end of the day you know i can be fairly confident that you know bitcoin might grow based on a bunch of fundamental indicators but it's never 100 so i would feel trouble if they made an investment choice based on my behavior but uh generally speaking the several metrics that i've used first of all i've used the actual the 30-day mvrv ratio they have shown here so we do these weekly market streams on every friday on our youtube channel so if you're interested 4 p.m central eastern time actually in a couple of hours without me this time but usually i'm there as well we just talk about the market and uh two two weeks ago or about a week ago actually um we talked about

how bitcoin's 30 am vrv ratio which is sort of that level with which uh short-term bitcoin holders like to take profits are starting to reach that historical proverbial danger zone between 15 and 18 and 20 so once it hit about 18 it was written fairly after sort of a couple of hours after uh it reached all-time high that was the first uh sign of concern for me generally likely to show that the chart there are several times in the past that bitcoin's mdmv ratio was able to even like get to 40 but it never lasts for a couple of days and i'm not one to sort of only sell at the top so i'm fine with just when there's a red signal i'll you know i'll be fine with the profits that i think so that was one the actually the other metric that i've used is the

sentiment or sort of social data that we have so we like i said we currently track about 1500 different cryptocurrency related social media channels this includes telegram groups crypto subreddits uh discord channels twitter accounts uh professional trader chats and we're expanding that list every day we're now adding nft different nft channels different d5 neil aggregated channels so what we're trying to do is we're really trying to anywhere that the main topic of discussions or cryptocurrencies we try to listen it and so over the last three years we've built a massive database we still collect about 100 000 new daily messages on on every every day um and we've built a bunch of indicators using that data so one indicator in particular sort of social volume shows dimensions of a cryptocurrency from all of these uh

different sources and we've done some back tests on this idea that when you see massive spikes in social volume as well as social sentiment of a project uh especially if that coincides with the okay actually let me just show it so yeah so this is social volume this is dimensions of a coin plotted against the price of that coin so generally speaking we see very often these obviously it's not always going to look as pretty as here but we see periods where there's a massive uptick in a social mentions or chatter towards a particular coin when it coincides with a massive price appreciation of that coin what that usually indicates is uh the fact that the market is becoming a little bit and

especially the retail sector is becoming a little bit too overzealous and too greedy for its own good so these new mentions they are not people talking about matic from a fundamental perspective and talking about how great the project is and how much um you know it's going to change the world it's just you know a bunch of people spending rocket emojis and fire emojis and right so usually it's uh the sort of retail crowd uh becoming hyper aware of the rally uh and as a result uh that usually tends to coincide you know our founder always thinks that you should bet against the crowd so anything that the crowd is doing you should go the opposite way and this is a metric that seems to follow that pattern to a t um it also points these massive spikes in social volume points to a period in time

where everybody is aware of the rally so there's also going to be less of a buy side pressure uh moving forward which is going to sort of potentially negatively impact the price of a coin so i've used um okay so this is a back test that we've done so we have this list on the website called social trends it's very simple it's a list of 10 words that is getting updated every hour which shows the top 10 words that are gaining the most social volume or sort of the most mentioned encrypted social media are relative to their previous two-week average so we're trying to figure out at this point in time what are people talking about the most in cryptocurrencies right um so very early on when we created this list what we notice is when a coin's name or

a ticker ends up in the top three positions on this list it's usually bad news for the coin um so what we've done is we've done a back test where we looked at about 650 different times or events where a coin's name or a ticker has ended up on this list of ours and the average price performance of these coins 20 days before and 20 days later so 20 days before especially six days before you can see the massive price appreciation of these coins on average which makes perfect sense because the reason they're appearing on our list in the first place is because everybody's talking about them however afterwards uh there's typically a period of downtrend for the coin uh so about 13 on average seems to be a price depreciation in the coins that we have um that we have uh

examined so this was done about uh seven months ago so it was before this final cycle so some things might have changed take it with a grain of salt but yeah this is uh how we we use this data and then lastly uh with our social social metrics we've also gone a step forward further and we've used machine learning to not only show you the amount of mentions of a particular coin but also try to figure out what is the mood of those dimensions in other words you know is the spike in mentions of bitcoin uh is that people being very bullish about btc or is it people being very scared about btc because and we here employ the same idea we're extremely bullish sentiment we typically think is pretty bad for a coin

um and then generally speaking when a coin is super bearish or people are super bearish about the coin this is where we like to be from a uh investment perspective like for example you can see here this massive spike in bitcoin's sentiment in right particular is the correction started this was the time that tesla announced that they will no longer accept bitcoin as payment and a lot of people were happy about it uh because it allowed them to buy the dip and move into a bitcoin at a discount um so generally speaking seeing that type of massive spike in a social sentiment um told us and there was also massive spike in by the dip mentions or a platform so we thought that this points to a period of time where people are still

failing or sort of to euphoric or too excited for their own good and it might not be the end of the correction which eventually uh it wasn't until we moved into a predominantly bearish atmosphere that bitcoin was able to uh start recovering and so this is bitcoin at around 59k when it got to 65k we basically sort of rivaled these levels so it was another indicator for me that it might not happen immediately but it's going to happen probably very fairly soon that we're going to see uh uh short-term price depreciation short term is the the operating word i don't think this works sort of on a longer term scale go ahead i'm so so sorry it's the end you could go on for hours i'm sorry yes very very interesting topic but uh thank you so much