I Love Negative Feedback
No, not that kind. It hurts when you tell me I’m bad repeatedly. I don’t like adverse feedback loops either, the kind which drive things from bad to worse, as described by Xabo’s board member Larry Summers here. I am referring to the engineering control theory where a system is self correcting.
Negative feedback loops sound like a bad thing, but they are actually good. Their opposites are positive feedback loops, which make unstable systems. Stable systems need negative feedback loops. Positive and negative aren’t normative statements, but describe how a system responds to changes.
Positive feedback is when a system variable X changes, the system responds to be more sensitive to X. Since X is now more important, the system does even more X. Positive feedback loop system will either implode or grow beyond bounds and break. A good example of positive feedback is when a sound system starts screeching during an old school stage performance. It gets louder and louder, feeding on itself until it can’t get any louder.
Positive and negative feedback can be found in financial systems too. Stop loss orders and forced liquidation induce the sale of assets when the price drops, causing the price to drop further. In the US, Trading Curbs are implemented to prevent the market from imploding.
Negative feedback loops are seen in systems that are self correcting. When a system input changes to do more of Y, the system responds by being less sensitive to Y. Eventually the system will settle at a new equilibrium based on the new input. A good example of negative feedback is the overall price system acting through supply and demand. When something is more scarce, it becomes more expensive, and demand drops.
Nuclear reactors are judged on their safety based on their temperature coefficient, which describes how their feedback loops work. Reactors with negative temperature coefficients self correct to cool themselves down automatically if they start to overheat. If a reactor gets into a state with positive feedback, bad things happen. Negative feedback is important to say the least.
Bitcoin uses negative feedback too. It uses feedback to keep block creation at a 10 minute average. When more miners join the network, the difficulty level responds to make finding a block less likely. If blocks on average take more than 10 minutes, the difficulty drops. This is why even though there is thousands and thousands times more hashpower than in 2010, block times still average 10 minutes.
Factom also has a negative feedback loop. The Factoids have a supply and a demand. The Factoid supply rate is fixed, but the demand is variable. If you remember, the Factoids are created at a fixed rate and granted to the Federated and Audit servers. The Factoids are created at a rate independent of any prices.
The Factoids are demanded and removed from circulation when users buy Entry Credits (EC) to use Factom. The Factoid/EC exchange rate is set by the market and matched internally by the Federated servers. The goal is to have a stable price for ECs. This will allow Factom users to ignore the Factoid price, just like like how Gmail users can ignore the Google share price.
Factom has a source and a sink for Factoids. Both the sink quantity and the price of Factoids are variable. The price of Factoids is at the whims of the market, but here is where the negative feedback loop comes in. The conjecture is that Factoid price will reach an equilibrium directly proportional to Entry Credit usage.
Let’s conduct some thought experiments. First assume that the EC usage is constant, because the cost to submit Entries and usefulness of the system is independent of the Factoid price. Let’s postulate that the market drives the Factoid price to a low level. The Federated servers will match this internally and raise the Factoid/EC exchange rate. If the price drops to 1/10th of the earlier price, then 10x more Factoids will be needed for the same number of ECs. This means that now 10x more Factoids are being removed from the market. This will lower supply, which will tend to push the Factoid price up.
Alternatively, assume that the market pushes the Factoid price very high. If the Factoid price goes up 100x, the Factoid/EC exchange rate will be 1/100th what it was before. The increase in supply will also go up 100x relative to the demand. This increased supply would tend to drive the Factoid price down.
These negative feedback loops should stabilize the Factoid price, driving it towards a value directly proportional to how much Factom is used.
