Bitcoin over $20k? Machine Learning Says Yes!
Bitcoin has come back into the news and the minds of many people. Over the past few months the bitcoin ecosystem has hit a few major milestones:
With both the Basics of Economics and Machine Learning as two lenses, how does the price track over the next few months?
The Basics
Let's first look at the economic indicators moving forward the next few Quarters:
1. The yield curve has inverted. This has signaled the past 7 recessions 7/7 times (100%)
2. The Manufacturing Index has contracted in June. This again has signaled a recession.
3. Interest rates have evened out and signal the fed anticipating an equities slow down.
The economy is saying recession
In a recession, the idea is to move out of equities and into scarce assets. Gold, US T-bills, and scarce cash-like resources are usually good holding assets for rough economic weather. However, both Gold and US T-bills have a few disadvantages. Gold has holding and security costs. US T-bills have the Fed printing an endless amount of money to "correct" the economy. Thus, another scare "cash-like" resource is needed.
Overall BTC beats gold and cash-equivalents in a few categories:
1. Holding Costs: BTC (minimal to none), Gold (yes, significant), US T-Bills (minimal)
2. Liquidity: BTC (Moderate to High), Gold (low), US T-Bills (High)
3. Scarcity: BTC (Exact, limited), Gold (unknown), US T-Bills (variable, decreasing)
You may wonder why gold has an unknown scarcity. The amount of gold is probably much higher than publicly know and is hoarded by various individuals of large wealth. The US has nearly 8,000tn of gold. Yet, a single mine in Alaska has enough proven gold reserves to produce nearly 56,000kg of gold through 2020. Extrapolate that over the various, more productive mines across the world, and we can estimate there is enough gold to double the US reserves in less than a decade.
BTC will be a safe asset for rough economic weather. Therefore, BTC should become more valuable than Cash Equivalents. BTC probably won't make a huge gain on gold, but may rally against gold if new traders flood the market come late August and September.
Machine Learning
With the basics supporting my initial thinking and the help of machine learning, the above curve was approximately produced. I modeled a 3 step rally using an LSTM neural network with 365 stages -- trained with 1100 training epochs. Fed into the model are three pieces of data;
(1) time series of the largest 10% of coinbase purchases
(2) the 10 most popular news articles for the relevant 15 day window. From these, sentiment and negative/positive words are extracted and fed as features to the model
The result after training and prediction is a curve that estimates a price target of $20763 on Dec 1. A theoretical modeled maximum is just shy of 33k (z>3.9) and modeled minimum just above 16k (z<3.9).
Takeaways
BTC is headed for a rally and macroeconomic factors, not whales, will drive it.
Target Price: 22k
- Largest Combined Hash Rate. This is an indication of health of the ecosystem.
- Greater than 5k gain in spot price verses the USD
- Multiple US based exchanges operating
With both the Basics of Economics and Machine Learning as two lenses, how does the price track over the next few months?
The Basics
Let's first look at the economic indicators moving forward the next few Quarters:
1. The yield curve has inverted. This has signaled the past 7 recessions 7/7 times (100%)
2. The Manufacturing Index has contracted in June. This again has signaled a recession.
3. Interest rates have evened out and signal the fed anticipating an equities slow down.
The economy is saying recession
In a recession, the idea is to move out of equities and into scarce assets. Gold, US T-bills, and scarce cash-like resources are usually good holding assets for rough economic weather. However, both Gold and US T-bills have a few disadvantages. Gold has holding and security costs. US T-bills have the Fed printing an endless amount of money to "correct" the economy. Thus, another scare "cash-like" resource is needed.
Overall BTC beats gold and cash-equivalents in a few categories:
1. Holding Costs: BTC (minimal to none), Gold (yes, significant), US T-Bills (minimal)
2. Liquidity: BTC (Moderate to High), Gold (low), US T-Bills (High)
3. Scarcity: BTC (Exact, limited), Gold (unknown), US T-Bills (variable, decreasing)
You may wonder why gold has an unknown scarcity. The amount of gold is probably much higher than publicly know and is hoarded by various individuals of large wealth. The US has nearly 8,000tn of gold. Yet, a single mine in Alaska has enough proven gold reserves to produce nearly 56,000kg of gold through 2020. Extrapolate that over the various, more productive mines across the world, and we can estimate there is enough gold to double the US reserves in less than a decade.
BTC will be a safe asset for rough economic weather. Therefore, BTC should become more valuable than Cash Equivalents. BTC probably won't make a huge gain on gold, but may rally against gold if new traders flood the market come late August and September.
Machine Learning
With the basics supporting my initial thinking and the help of machine learning, the above curve was approximately produced. I modeled a 3 step rally using an LSTM neural network with 365 stages -- trained with 1100 training epochs. Fed into the model are three pieces of data;
(1) time series of the largest 10% of coinbase purchases
(2) the 10 most popular news articles for the relevant 15 day window. From these, sentiment and negative/positive words are extracted and fed as features to the model
The result after training and prediction is a curve that estimates a price target of $20763 on Dec 1. A theoretical modeled maximum is just shy of 33k (z>3.9) and modeled minimum just above 16k (z<3.9).
Takeaways
BTC is headed for a rally and macroeconomic factors, not whales, will drive it.
Target Price: 22k
Disclaimer
The information and publications are not meant to be, and do not constitute, financial, investment, trading, or other types of advice or recommendations supplied or endorsed by TradingView. Read more in the Terms of Use.