QuantCheck’s goal is to provide an edge to cryptocurrency traders. We do this by designing propriterary trading models.
As of today, our high-timeframe (HFT) cryptocurrency model is available to the public. To see how this model has performed on ETHUSDT, click here. We plan to publish the model on other coins, as an additional offering.
​
Subscribers can choose one of two types of subscription (1) Free (2) Premium. Free subscribers will receive delayed signals by 5 days. Paid subscribers will receive immediate notification on signal change.
About the Model
The Model is a standalone long/short automated trading system providing both long and short signals using a medium-high timeframe. To simplify, think of the model as a robot that says when to buy or sell cryptocurrencies.
There are periods in the market where the model works well and there are periods were it does not. However overall, following the model and accepting its temporary losses, leads to outsized gains in the longer term. For beginning traders and non-active market participants, following the model’s signals exactly will likely work best.
For institutions and sophisticated traders, use the model as confirmation — feel free to adjust take profit and stop loss according to the strategy you are running.
Signal Generation
The long and short signals are triggered based on changes in momentum, trend, volatility, volume, and price. It is purely a quantitative model. Every asset class has a different market structure; one of the beautiful properties of cryptocurrencies is the magnitude of price change realized. The model is able to ride the majority of trend length and compound returns.
​
There are dozens of variables that go into producing these long and short signals in an effort to be in the right direction of the trade and to not getting shaken out of the trade, despite tremendous volatility. It is important to review the prior trades executed by the model to fully understand in which situations the model takes trades. Click here to see a list of all trades.
​
Below is an annotated chart showcasing the model’s trades during Ethereum’s peak in 2022. Note, that the model doesn’t catch pico tops and bottoms; it waits for a trend to begin to form or weaken. The model works best during trending markets and suffers during choppy markets. This is by design. Capturing the meat of the move and bleeding through the choppy markets has been more profitable than any other strategy. “The person who chases two rabbits catches neither.”
Trade Sizing
We use a 100% equity assumption for each trade conducted. That means if we start of with $10,000 and make a trade that profits us with $2,000, the next trade would use $12,000 of capital. If we started with $10,000 and make a trade that ends up being a $2,000 loss, then the next trade would use $8,000. There are two main advantages with sizing this way: (1) Compounding returns (2) Can’t lose more than you have.
Stop-loss
The model has proven to work best without a stop loss; the model only exits the trade upon the occurance of a new signal. However, the primary objective of a trader is to protect capital; risk management is of utmost importance. Thus feel free to add an invalidation point to the model’s signals. We like a trailing stop loss based on the average true range. Per its definition, you can never predict a black swan; having the right tools and risk management practices is key to outperforming.
Leverage
We do not use any leverage in the backtests. Using 2-3x leverage on this Model can increase returns dramatically; anything above that and it would be detrimental to returns. We do not recommend using leverage.
Timeframe
Trades are executed on a daily basis. That means there can be a maximum of one trade per day and there could be no trades for weeks. One might think that because of the volatility and the magnitude of price change in crypto, it might be better to use a weekly timeframe to not get chopped up. There are definitely very profitable strategies based on a weekly timeframe and there are definitely less false signals than on a daily timeframe; however, weekly models tend to miss a substantial part of the move which hinders expected returns.
​
FAQ
Does the model ever go to cash?
The model is always either long or short. If the model produces a long signal then the next signal would be a short signal, and vice-versa.
When does the model perform poorly?
In choppy markets! In a ranging/choppy market as shown in the purple box below, the model doesn’t perform well. However, the areas around it make up for the loss and more.
When does the market perform well?
In trending markets!
How do I read the charts? When are the trades executed?
On the charts the blue arrow below represent longs and the red arrow from above represent shorts. The arrows are placed on the open of the next bar. For example, if today is March 2nd the model will analyze data as per 11:59 UTC on March 2nd and if a long signal is triggered, it will be placed on the open of March 3rd.
Have commission and slippage been accounted for in the backtests?
In the backtests we use a comission of 0.4%, which should be a sufficient cushion to account for slippage for liquid assets when trading with sufficient volume.
During a bullish year, the model will make about 8-12 trades and greater than 25 trades in a bearish year. 25 trades is too small of a number for comission to make a meaningful impact. So even if you’re not trading with sufficient volume and do not have preferred comission rates, it doesn’t need to be of concern.
Does the model scale entries and exits?
The model doesn’t scale into and out of positions. It assumes 100% position size immediately at each entry and exit signal. The signals are produced on daily close UTC.
Can I invest my funds with QuantCheck Labs?
QuantCheck Labs is purely a research firm that provides cryptocurrency trading signals. We do not invest funds for subscribers.