Comment on page
Frequently Asked Questions about Predictoor
Before we answer, please understand that to make decent money, it will take hundreds or thousands of submissions, guided by AI/ML models. So doing this manually would be tedious and error-prone. Therefore a bot is the most practical way. This is why the Predictoor core team has made it easy to run a bot. Here are the getting-started READMEs Predictoor and Trader, respectively.
To the question "Is it possible to predict without a bot", the specific answer is "Yes". Here are the details
- Somehow you need to submit an up/down prediction, with OCEAN stake, as a transaction to Sapphire chain. Here are the Possible ways to do such a transaction:
The most important metric to assess confidence in a prediction is "how much OCEAN have people staked on this prediction", both up and down. It's how much they're willing to lose if they're wrong. This metric is not gameable because OCEAN has scarcity. It's "put your money where your mouth is".
What factors influence the accuracy of Predictoor? How accurate do you think the predictions will be over time?
That's up to Predictoors to figure out! Certainly, historical price information has a huge influence. After that, there are many possibilities: other CEX and DEX data, on-chain data, social media data, GitHub data, news, weather, and more. The Predictoor README gets you started, using historical price information.
How does Predictoor compare to Numerai Signals? People supplying to Numerai Signals don’t get paid out from trading (like the main Numerai product).
In Numerai Signals, payout is a function of the submitter’s “originality” compared to other signals. These signals are used as inputs for prediction models in Numerai main product.
Whereas in Predictoor, payout is a function of the the accuracy of the submitter’s prediction, compared to the true value. These signals are predictions, directly.
How does Predictoor compare to IntoTheBlock Price Predictions tool? Is has prediction feeds of crypto prices.
The IntoTheBlock tool is run internally by IntoTheBlock, to build build centralized AI models that are served up via a REST API and a webapp.
You can view Predictoor as a "next-gen" version of that tool: a decentralized version that crowd-sources predictions.
- Crowd-sourcing with accuracy incentives has potential to do much better than any centralized approach. It doesn't rely on the ideas of a single team; instead it makes the opportunity to predict open to anyone in the world. A perpetual data-science competition. The incentives mean that only the best will stick around. This will get reflected in the prediction feeds' accuracies.
- Being decentralized means one doesn't need to rely on a centralized actor running centralized services. Useful for anyone coming to rely on prediction feeds.
Finally, Predictoor's ambition is broader: to extend beyond crypto into energy, weather, agriculture and more. It's easier to expand scope of decentralized feeds, because you don't need to spin up a new team to focus on each new vertical.
Last modified 17d ago