Caduceus: The Journey to Stability

This journey, and the subsequent research that we will examine below, was brought about by witnessing the boom and bust cycles that are ever present with Bitcoin and all other subsequent altcoins. The massive volatility only supports speculation rather than the desired adoption in the form of a closed looped system. By a closed loop system I mean that there is no need for government backed fiat because the cryptocurrency has the right properties to reliably facilitate economic activity for businesses and individual users alike. Currently, the number one inhibitor to mass adoption of any cryptocurrency is not scaling, as is commonly thought by most people in the community, but really that it is not a stable storage of value for any given period of time. Once there is stability in the ecosystem, and thus people can rely on it to be predictable in purchasing power, then mass adoption can truly begin to occur. This means that there is neither the perpetual but gradual inflation experienced in fiat nor the extreme price deflation seen in Bitcoin. Caduceus aims to bring the necessary stability while still retaining the full levels of decentralization exhibited by Bitcoin.

Examining Current Currency Solutions

To first understand what is beneficial, I found it useful to quickly examine current solutions and their effectiveness in maintaining stability. The first type of currency we will look at is government backed fiat currencies. Barring unstable countries, fiat currencies are generally stable enough for usage and the inflation is relatively predictable from year to year but perpetual in nature. This perpetual inflation causes serious distortions in purchasing power over time but is benign enough that the general populace accepts it as a medium of exchange. To go along with a fairly predictable performance, the mass acceptance of fiat is also augmented via government decree. The major drawback of government backed fiat is that its creation or destruction is tied to the debt cycle of that particular nation. Unfortunately, politicians always vying for re-election, have an incentive to promise bills that require massive amounts of taxation and debt related spending. This requires more money to creation to cover national debts owed to its creditors. Consequently, this means that as debt spirals out of control the purchasing power of the currency drastically declines into a state of hyperinflation. Thus, the major downside of government backed fiat currencies are that they are subject to the debt cycle and inevitably when the debt burden becomes too great the value of the currency eventually will fall to zero.

The next type of currency we will examine are cryptocurrencies. Cryptocurrencies generally exhibit the greatest difficulties in regards to being a stable storage of value over the long term. Relative stability is absolutely necessary for cryptocurrencies to ever function on any level as a medium of exchange for large scale commerce. A popular sentiment among the community is that if you hold Bitcoin through enough of these boom and bust cycles then it will become stable enough to facilitate commerce. Unfortunately, the reality is that most individuals lose money investing in Bitcoin and this in turn provides users with a bad experience that will likely keep them from coming back in the future. Furthermore, businesses and individual users alike will not adopt cryptocurrencies until the risks of massive volatility are eventually mitigated. Currently, the volatility is unpalatable for businesses so out of necessity they use payment processors like Bitpay that converts Bitcoin back into a government backed fiat of choice. They do this because holding a large amount of Bitcoin presents a significant business risk. To compound the effect, this is on top of all of their other business risks such as competitors, a changing market, and so on. To garner true adoption from businesses would require them to accept, hold, and pay their employees with cryptocurrency. Likewise, individual users also have risks such as debt, bill payment, and other essential living expenses. If they were to try to rely on cryptocurrency, in its current state, they could find themselves in a bad position and unable to cover monthly expenses. The major strengths are that cryptocurrencies are decentralized to a degree and thus they theoretically should be less subject to censorship, seizure and they have the ability to move money across borders in an extremely efficient manner. That being said, these major benefits cannot fully be realized because of the extreme levels of volatility exhibited in the space. This is largely why, instead of a functioning currency as was originally intended, cryptocurrencies have largely been relegated to being a speculative asset.

The last, and also oldest type of currency, is a commodity currency in the form of gold or silver. These two precious metals tend to be the most stable in the long-term for a multitude of reasons. For one, they both have significant value outside of being a medium of exchange. Gold and Silver, due to their elemental properties, have demand in industries as far reaching as electronics, jewelry, solar panels, and photography. Additionally, they also operate within the means of supply and demand. By that I mean that if there is a distortion then via supply it is corrected by the invisible hand. For example, if the price of gold or silver goes up then miners start to mine in areas that were once unprofitable. This means that the supply increases to offset the price increase and a market equilibrium is once again maintained. The exact opposite happens when the price of gold and silver decreases below market equilibrium levels. The miners quit mining because it is not profitable anymore and thus new supply is significantly reduced which then brings price back to the market equilibrium. This means that short term distortions are possible but there is still relative stability over the long term. It is impossible to perfectly emulate this success via an algorithm but a better and more candid attempt can be made. My research and quest to create a more stable cryptocurrency began in trying to emulate how the purchasing power of gold and silver operate over long periods of time.

Causes of Volatility in Cryptocurrencies

The first major step in the process was to determine what are the mains causes creating the volatility within the market as a whole. Many believe that it is a lack of liquidity but in reality it is a combination of the block halving that occurs within the block reward algorithm and the total supply cap. The block halving for Bitcoin, regardless of market conditions, halves every 210,000 blocks. The block halving of Bitcoin occurs without regard for demand and thus it causes distortions with these distortions becoming increasingly larger with every cycle. This leads to a steep rise in price after a block halving and then culminates with a requisite bear market that follows. Contrary to popular belief, price deflation can be just as dangerous as price inflation to an economy. One only needs to look at the Great Depression for an example on how price deflation can create a death spiral in the cost of goods and thus employment.

The total supply cap is much more of a benign factor than the block reward halving. The issues caused by having a finite maximum supply will likely only manifest itself should Bitcoin ever gain mass adoption. For the sake of this experiment, we will pretend that by some magical and improbable turn of events Bitcoin indeed did garner complete and total adoption from the world. Even under these conditions, we would simply be trading two percent inflation for two percent price deflation caused by rising population levels. Increases in population do indeed increase the demand for a currency. Over a long period of time this would create distortions that could cause irreparable damage to the world economy. Ideally, the best currency is neither extremely deflationary or inflation and also not perpetually one or the other. If these conditions are met, this means that there can be a short term price distortions but over the long term things will remain relatively stable.

Endogenous vs Exogenous Solutions

Exogenous mechanisms measure price by an outside data source of some kind. An example of this would be measuring data from exchanges to determine the U.S dollar value of a particular cryptocurrency. This method does indeed offer the most accurate price data for a dynamic block reward algorithm. However, there are two major issues with using this method to measure price. Firstly, measuring cryptocurrencies in the U.S dollar is a flawed idea since it would subject Caduceus to fiat fluctuations. Additionally, an outside data source presents a large security flaw and should the price feed be manipulated or taken down in some way then the entire algorithm would cease to function properly. This potential security flaw and the idea of measuring price in fiat made me eliminate the idea of using an exogenous solution as a measurement of price despite the perceived accuracy it provides.

Endogenous mechanisms measure price from some internal blockchain metric. An example of this would be to use changes in transactions or changes in hashrate to represent a requisite change in price. As we all know, none of these metrics are perfectly correlated so they offer an imperfect look at changes in price and consequently in demand. That being said a metric may be used if it has a high enough correlation coefficient and it proves to be extremely difficult or particularly expensive to manipulate over long periods of time. After setting on the idea of using an endogenous mechanism, I began to search for a suitable metric to use as a basis of measuring changes in price and thus demand.

Highly Correlated Endogenous Mechanisms

In my quest to find a suitable endogenous metric, I began to do a statistical study of sorts to see how each metric correlates to daily price data. This was an especially exhaustive process as it was done for several different cryptocurrencies and for every metric imaginable. For those that do not understand statistics, a perfectly uncorrelated data set would have a correlation coefficient of 0 and a perfectly correlated data set would have a correlation coefficient of 1. Additionally, each statistic needs to also be proven to be significant. I first began the study by examining all of Bitcoin’s metrics but what I found was fairly disappointing as the highest correlation was hashrate and that came in at around a .4 correlation ratio to price. This is a medium correlation but it is not strong enough to use for the algorithm. I then began to study other currencies with similar block reward algorithms as Bitcoin but they all produced similar results.

After having failed to produce any results by examining cryptocurrencies that have a set schedule of block reward halving, I decided to first try studying those that had the most similar block reward algorithm to what I am proposing. Cryptocurrencies with no block halving and no total supply cap obviously proved to be the most similar. I decided to look at Ethereum and it did indeed prove to be beneficial to finding a suitable metric. Hashrate proved to be the most correlated metric to price with a ratio of .9. This means that it is nearly perfectly correlated which makes it suitable to be used for the algorithm. Additionally, it also validated my beliefs that Bitcoin’s block reward algorithm causes significant distortions that operate outside the norms of supply and demand. Lastly, hashrate was chosen also because it is the most expensive metric to manipulate over long periods of time.

Preventing Bad Feedback Loops

The next step in the process was to prevent bad feedback loops caused by the dynamic algorithm. Bad feedback loops are when a particular piece of data feeds another in an infinite loop and causes a perpetual and uncontrollable distortion. An example of a bad feedback loop would be if a nation is highly indebted but decides to raise interest rates to combat inflation. Unfortunately, the interest rates on variable debt would inevitably cause more debt and more inflation with the cycle repeating over and over. In the case of cryptocurrencies a bad feedback loop would be something that continually causes it to endlessly fall or rise in value via a data feed cycle. Many cryptocurrencies have tried to have a fully dynamic block reward to stabilize price and almost every single one has failed because of bad feedback loops.

To prevent bad feedback loops, I decided that the block reward would indeed move in relation to changes in hashrate but it would also have to move from a baseline target. For example, if the block reward is 5 and hashrate went up 10 percent in the last period then the next block reward would also go up by 10 percent. This would mean that the block reward for the next period is now 5.5. Additionally, the bad feedback loop would have been prevented and the algorithm is also still somewhat adjusting supply in relation to demand which should still make it considerably more stable than most cryptocurrencies today.

Accounting for Moore’s Law

One may have thought that the algorithm was finished after having prevented the bad feedback loops but this was definitively not the case. The algorithm, up to this point, assumes that all changes in hashrate are because of increased demand. However, to be a more accurate algorithm it must also account for changes in computing power via Moore’s Law. Moore’s Law states that every 2 years the number of transistors and thus computing power doubles. To compensate for this you would simply need to break down Moore’s Law into your period of correction. Basically, if your period of correction was 1 day then your Moore’s Law statistic would be equal to one days worth of Moore’s Law. Additionally, you would take one plus the delta of the hashrate change minus the Moore’s Law statistic and then multiply it by the base block reward to get your new block reward for the next period. If all of that was hard to follow in text form then the equation below should make more sense.

S2 = S1(1+ΔH-M)

Another major issue is that transitions from GPU mining to ASIC mining often exceed Moore’s Law by several magnitudes of order. For this reason, Caduceus will launch on the Equihash algorithm because it is already being mined by ASIC’s and it is fairly easy to run at home due to being compatible with 120V or 220V. This means that changes in the hashrate of miners will be much more in line with Moore’s Law. The Moore’s Law statistic, although used to compensate for changes in computing power, can be used for other means that we will explore in the next section.

Front Loading Volatility

It once again seemed like the algorithm was complete but there was again one another important issue left unaccounted for. I realized that I also needed to account for the actual changes in hashrate and the essential security that it provides to the network. By this, I mean that if we are guiding the price within a range then we are also limiting changes in hashrate on the network. To compensate for this, I decided that the best thing to do would be to use the Moore’s Law statistic to allow the price to move in a fairly incremental upwards manner via overestimating Moore’s Law. Additionally, it made sense to add in an exponential decay formula so that the price movements were mostly front loaded but became more accurate as time goes on. This potentially allows the network to be bolstered by additional hashrate that will make the chain more secure against attacks. The real trick of this is in predicting how long it will take to attain the desired security and then afterwards maintaining or simply slightly increasing that security over time. The exact details will of course be made public in the whitepaper that will be released much closer to launch. Below is the formula to calculate the Moore’s Law Statistic.

M = a(1-r)^x+b

The variable M represents the Moore’s Law statistic, a represents the original value before decay, r represents the decay rate, x represents the number of intervals that have passed, and b is the minimum value that you wish to keep the Moore’s Law statistic for the long term. By structuring it this way you can gain the necessary hashrate for security purposes without sacrificing the need for a more stable currency.

Gaining Freedom From Bitcoin

Another factor in obtaining a more reliable and stable currency would be to release the tether between Bitcoin and all other altcoins. Bitcoin, due to its bad economic foundation and first movers advantage, wildly drags the rest of the market along with it. To decouple from Bitcoin, the most essential thing to do will be to establish unique fiat trading pairs and fostering Caduceus as a good trading pair for other cryptocurrencies. Caduceus, if working as designed, will be more stable than Bitcoin but still retain the majority of the decentralization aspects that stable coins like USDT lack. You would essentially be getting the best of both worlds and this is essential for a worldwide and decentralized currency to operate. Additionally, fostering a closed loop system in which fiat and Bitcoin are not needed at all is essential. This means that businesses transact in it, purchase goods with it, individuals get paid in it, and ultimately rely on it to hold a relatively steady value so that bills can be paid for in a reliable manner.

Extra Added Benefits

In addition to the benefits listed above, there are also other benefits that the algorithm brings to the users of Caduceus. For one, it increases the quality of the distribution of wealth within the network. With block halving present in Bitcoin there is a disproportionate amount of coins rewarded to early adopters. Of course, from a logical standpoint those that take risks deserve the greatest reward but in reality it is still not good for the future of any currency for one percent of addresses to have over 90 percent of the coins as seen in BTC. This can lead to rampant manipulation at times and this is not preferable to the long term health of the community. I would expect that early adopters would still be rewarded more than later adopters because of less hashrate shortly after launch but still not as disproportionately as seen with Bitcoin.

Another positive side effect of a new block reward algorithm is that it decreases the incentive to speculate on a coin. The algorithm instead provides an incentive to use it for real world economic activity. Speculation adds no value to the ecosystem as there must be losers for there also to be winners. True adoption adds significant amount of value and also legitimacy to the market as a whole. If a person were to try to speculate on Caduceus, I expect that they would be disappointed to some degree because of the self correcting nature of the algorithm. If they try to manipulate the price in the upwards direction then it would cause a distortion which would increase hashrate and then the block reward would increase to compensate for this. Since miners generally have to sell to cover expenses this would bring the price down to the market equilibrium. The same would occur if someone tried to manipulate the price in the downward direction. The block reward would be decreased and then it should eventually send price back to a market equilibrium.


The algorithm is not perfect in the nature of its correction but it should prove to be functional and effective in stabilizing price to palatable levels to facilitate economic activity and also for developers to build real solutions on top of Caduceus. The algorithm was largely developed out of a necessity after seeing and experiencing first hand the massive boom and bust cycles that coincided with block reward halving events and the requisite demand shifts. The market most obviously needs some level of stability to gain legitimacy and a foster a real use case in the world economy. In a market that only cares about technology it seems that economics has been forgotten as an essential foundation for a functional currency. However, that time is over as Caduceus launches in the early part of 2019. There will be more articles and of course a whitepaper to come in the next few months so please be sure to follow me and the social media below if the project sounds interesting to you.