A cryptocurrency buy and sell model that minimizes the volatility problem. This is the project chosen for first place in the second edition of the Datathon Challenge by Cryptocurrencies, revealed this Monday (6th). The winner received R$8,000, the second R$5,000 and the third R$2,000.
The competition was open to undergraduate students from all over Brazil and 100% online, carried out by the São Paulo School of Economics (FGV EESP) in partnership with Ripple, through the University Blockchain Research Initiative (UBRI).
The winning project is from FGV Crypto with the project “Non-directional investment strategy using cryptoactive options”. The team members are Calebe Soares (FGV EAESP), Francisco Hansen (FGV EESP), Guilherme Regueira (FGV EAESP), Isaque Sathler (FGV EESP) and Lucas Rocha (FGV EAESP).
Students developed an options-based crypto investment strategy.
Called “Gamma-Vanna-Charm Scalping”, the operation consists of assembling a Straddle (buying a call and a put with strikes as close as possible to the price of the underlying asset), and generates losses only if the asset is “parked ” (which is rare in the context of cryptocurrencies).
Applications to Bitcoin and Ethereum showed that returns were lower than traditional buy & hold, which can be explained by the recent bull period in the market.
However, the downside risk (variability of returns below the target return) was much lower. The sortino index, which measures the level of return per unit of downside risk, was considerably higher using the proposed strategy.
Students conclude that the tactic is feasible to be carried out in practice and that it has the potential to solve the biggest problem that the cryptoactive market has, so that the exposure of investors in this asset class is popularized and democratized.
In vice was the WASD team, with the project “Dapps and stablecoins as passive income”. The team assessed the differences between the lending of stablecoins in dapps (decentralized applications) and the obtaining of passive income through traditional fixed income securities.
To assess the effectiveness of the strategy, the students used the CAPM (Capital Asset Pricing Model) pricing model and the Markowitz investment portfolio optimization model with highly liquid stablecoin data.
The analysis was based on the two largest lending dapps (AAVE and Compound), and found that combinations between traditional fixed income and stable coins can maximize the expected return for the same level of risk (within-sample assessment).
The team is formed by Alexandre Wensko (FGV EESP), Cleverson Soares (FIAP), Guilherme Lima (FGV EAESP), Gustavo Jesús (UNIFESP) and Marcus Neves (USP).
In third place was team 42 with the project “Sentiment analysis based on Twitter posts and their interrelationship with the price of cryptoactives”.
To assess whether sentiment metrics are intrinsically related to BTC price fluctuations, the authors used the Scweet library to retrieve tweets that contained “bitcoin” and “btc” between June 2017 and June 2021. The approximately 62.6 1,000 tweets analyzed were evaluated in terms of daily average of tweet sentiments, daily average weighted by likes, by number of comments and by retweets.
The correlations between the sentiment variables and the BTC price proved to be high, and the Granger causality test indicated that the BTC price and volume indicators have a bi-causality relationship. The authors conclude with this that the market behaves in a mostly emotional way.
The team is formed by Felipe Gabriel (UFSC), Guilherme Terriaga (UMC), Matheus Konstantinidis (UFSC), Pedro H. Anjos (UFSC) and Vinícius Custódio (UDESC).