Sharpe.capital has a roadmap, white paper, and links to social accounts.
For those of you that haven’t heard of Sharpe Capital yet, here is a brief introduction. Sharpe Capital is crowdsourcing the world’s best asset sentiment via the Ethereum blockchain. Users of the Sharpe Platform earn Ether in exchange for their opinions about global equity markets. Backed by the Sharpe proprietary investment fund, rewards are paid to holders of SHP tokens on a quarterly basis.
I was lucky to get an interview with Lewis Barber, CEO at Sharpe Capital, to obtain a deeper insight into the business model and what is happening behind the scenes.
Okay Lewis, first question – When did you first get into the cryptocurrency world and what were your initial thoughts?
I’ve been involved with cryptocurrencies since discovering Bitcoin in 2011, more as a passive follower than an active participant. With a technical background in Electronic Engineering and Software Development, I was particularly fascinated by the low-level detail in the original Satoshi paper. It must have been late 2011 when I first laid my eyes upon it, and I instantly knew I’d discovered something that would significantly change the course of history.
The recent boom in ICOs gave me the courage to get more involved with the sector, and it happened to coincide with when James and I started working on the early ideas behind Sharpe Capital. I have been most surprised by the pace at which the sector moves. Many say the sector is still young and needs to mature, and I wholeheartedly agree. I just think it’ll be weeks, or months, as opposed to years!
What do you think about the current investment fund landscape?
There are huge opportunities to bridge the gap between traditional investment funds and cryptocurrency markets – bringing the wealth and breadth of opportunity in the blockchain world into the mainstream. I also think the public nature of distributed blockchain ledgers will inspire some really interesting business models based on increasing transparency in capital markets and reducing the propensity for fraudulent activity.
At Sharpe Capital, we have developed a trustless public trading ledger, which runs on the Ethereum blockchain and can be used to make investment activity public in real-time, whilst protecting sensitive position information via public-key cryptography. This is a huge growth area for us, as we see great potential to roll out our public ledger across the entire capital markets industry.
SHP holders can stake their tokens to make predictions on the future performance of assets. What happens if someone makes an incorrect prediction?
This is a great question because many people have compared our sentiment platform to prediction markets. The really novel difference between our sentiment platform and prediction markets is the fact that there can be no loss of SHP tokens. Interestingly, we’re not actually rewarding good sentiment, rather we are rewarding users for consistency. If someone is consistently bad at making predictions, that’s also useful to us.
Therefore, when users are very inconsistent in their predictions, this serves to diminish their immutable reputation score, which is stored on the Ethereum blockchain. Reputation scores are a key component to our reward formula and a low reputation score will reduce the potential Ether rewards someone can earn in the future. The only way to recover from this situation would be to start providing consistent information about future asset performance, which will slowly improve the user’s reputation sore over time.
In what ways is the investment approach at Sharpe Capital different to the one applied at projects like Numerai?
As I understand it, Numerai is crowdsourcing investment models in an attempt to reduce overfitting. We don’t believe this approach will work – what if every model is severely overfitted? There is no evidence to suggest combining many models, especially when the degree to which they’ve undergone appropriate validation and testing is unknown, would ultimately result in a less overfit model. The best – possibly the only – approach to eliminating overfitting, is to ensure that model validation strategies are robust and fit-for-purpose.
Instead, we are building our own highly accurate quantitative investment models, implementing advanced machine learning strategies to reduce overfitting, derived from James’ research during his PhD and postdoctoral career. Quantitative models that fail to take human psychology, cognitive science and decision theory into account will always lag behind strategies that seek to incorporate them. This is because the vast amount of capital traded today is still done so by human decision-making processes, which science tells us is fundamentally affect (or ‘emotion’) driven. This is why we complement this modeling with our crowdsourced asset sentiment, via our mobile application. By combining both approaches, we’re able to model both the intrinsic value of an asset and identify assets that are likely to be driven away from their intrinsic value by irrational investor sentiment.
Neural networks need to be fed large amounts of data to generate accurate models. How long do you estimate the ‘learning process’ of your algorithms to be until they can make precise predictions?
Artificial Neural Networks form only one aspect of our modeling strategy, we are complementing these with a whole suite of other machine learning techniques, these prediction sets are then consolidated using our proprietary MiDAS algorithm (Manifold-Driven Asset Scoring). We began training our ML models back in March, focusing specifically on the S&P500. We are already seeing highly accurate results in back-tests – our AI Portfolio Manager is capable of leveraging our ML models to produce an Enhanced Index-type fund with significantly better risk-adjusted returns than the S&P500. Once the token generation event is complete we plan to quickly scale these systems to begin live testing, while integrating our linguistic analysis models.
Training any given forecasting algorithm does not take an enormous amount of time (we’re talking under an hour on your average laptop to train an algorithm with 500 assets and several years historical data). The real challenge comes in determining what economic features, over what time-scales, to train our models on, and how then do we integrate the many predictions into a useful output from which our trading algorithms can make informed decisions.
This is especially true in the capital markets domain – we find the ‘rules’ that relate microeconomic data to future performance change over time, and we need to be able to adapt to that. It would not be wise, say, to simply train one neural network with decades of data for all assets. This gets even more complicated when you consider that these predictive relationships change from market-to-market: an algorithm trained on US equities would certainly not work effectively when applied to Asian equities.
Subject to regulatory approval, Sharpe Capital plans to issue the SCD, a token that pays dividends based on the performance of the fund. Do you have an approximate date in mind for the creation of this token?
Issuing SCD is a long-term objective of ours. Given the current regulatory landscape, we will have to work closely with regulators in key jurisdictions to make this a reality. We have reviewed the process for registering a security in the US, UK and Singapore, and currently favor Singapore for simplicity of registration. We believe, with our ample legal allocation, we will be able to achieve registration of SCD by early 2019.
Where do you see investment funds and Sharpe Capital in 10 years?
With the barrier to entry of financial markets reducing significantly, through innovations in blockchain technology and increased access to finance, my prediction is that we’ll see a vast array of niche funds & capital markets startups co-operating with each other to significantly disrupt the sector. Innovation is already moving away from traditional financial centers in the world, towards Asia, Eastern Europe and smaller jurisdictions like Gibraltar. Investment products available to everyday consumers are still quite restrictive and vary significantly around the world. Blockchain technology has created an ecosystem that doesn’t respect borders, where capital can move freely around the world, and services can be consumed by people anywhere on the planet. This will create fierce competition for the incumbents and drive unseen levels of innovation in capital markets over the next decade.
Sharpe Capital aims to be a significant player in this paradigm shift, through increasing transparency with our trustless ledger service and lowering the barriers to entry, with our crowd-sourced asset sentiment platform. We’re excited about working in cooperation with forward-thinking startups, and we’ve recently announced an exciting partnership with TaaS, the first closed-end fund dedicated to blockchain assets. We see our role as bridging the gap between traditional finance and the blockchain world and plan to work closely with regulators in a variety of jurisdictions to bring the benefits of cryptocurrency assets into the mainstream.
Where can people learn more about Sharpe Capital and the SHP token?
Sharpe.capital has a roadmap, white paper, and links to social accounts.
The post Sharpe Capital – The Financial Markets Protocol appeared first on Coin Noob.