m6

Some resources related to the M6-Forecasting competition.

Methodologies of top performers

• See medium reading list

Best plot ever

• RPS time-series

Official rules & contest site

• Contest site
• Rules as of Oct 7, 2021
• Rule clarifications

Example entries, and scripts

See precise/examples_m6

from precise.skatertools.m6.competition import m6_competition_entry
df = m6_competition_entry()

This will choose from dozens of different ways of estimating cov, and from a dozen different portfolio construction methods.

See also precise/examples_m6/full for example .csv's to modify as you see fit.

Related literatures

A couple of bibliographies:

• precise/LITERATURE.md
• precise/Refs

My articles on M6 contest

• How to train for the M6 Financial Foreasting Competition
• The future of forecasting competitions, according to the experts

Python time-series packages

Need to forecast something?

• Listing of popular Python time-series packages, including features, causality and so forth.
• Elo ratings of autonomous time-series prediction algorithms judged against live data.

Note that the Elo ratings contain a "special" category that might be particularly relevant to M6.

Live contests to practice on

Some of the live streams at Microprediction.Org are good fodder for predicting volatilities, covariances, et cetera. See in particular:

• Crypto streams described here
• Anything starting with 'fathom' in this list here
• See the daily $125 contest for short-term distributional prediction of crypto-currencies and stocks, including variances and co-variances, et cetera.

Random data to practice on

• Public API for time-series
• Example csv
• Tutorial on retrieving time-series data using microprediction package.
• IEX API and example scripts from Numerai

Analytical tools for 5-way rank probabilities ... for the M6 contest that wasn't :)

Ah well, in the first draft of this contest we were to rank 5 stocks 1..5. That's no longer the case. However:

• Paper soon. See notebook examples

This site maintained by yours truly, an employee of one of Intech Investments. Code is provided subject to the MIT License (and all disclaimers therein). Nothing here should be construed as investment advice, or even contest entry advice. Always refer to the official rules.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.