Back to Community
For whom are that Rules if referee do not follow them?

Second month in a row Quantopian Open winners chosen not by contest
Rules but discretionary, using metrics and argumentation not mentioned
in the Rules.

September prize the "Sharpest" algo by Tibor Szabo was discretionary
disqualified because of its out performance in July 2015.
Is there something about that in the Rules? No.

October prize the "Sharpest" algo by Adrian-Constantin Boca was discretionary
disqualified because of its under performance on 10 year backtest and
concentration risk.
Is there something about that in the Rules? No.

The next time referee may be suspicious that most profitable day was
Wednesday or something else not in the Rules to discretionary
disqualify algo.

For whom are that Rules if referee do not follow them?

I am not against 10 year backtest.
More of that I am strongly for long term backtest at least
of one full market cycle (as of today minimum 8 years).
But add it to the Rules and just follow them.

I am not against concentration risk control eater but first You should
at least explain how you going to calculate it, add it to the Rules and
just follow them.

1 response

Writing the post 1 year ago I did not know that the same verdict will hit my algos.
This time reason- fig leaf counts.

What is written in the rules:

Your algorithm must be hedged to the market.

True. It is hedged by long position in US bonds.
corr_pt 0.07, beta_spy_pt 3.16%

It should hold both long and short positions simultaneously, or be entirely in cash.

True. It had both long and short positions.

From verdict:

The algorithm is short "enough" to get a badge on the leaderboard,
True. It had all 3 badges

but it is not hedged by any reasonable measure

False. As I mention above it is hedged by long position in US bonds.
corr_pt 0.07, beta_spy_pt 3.16%.

If I am not correct please point me to the contest rules and yours meaning hedged by any reasonable measure and how you measured it.

Dan,

There were no need to apologize.
That's me should thank You for giving a prize in contest 14.
And tell you more, as before you only disqualified the winners, by disqualifying my simple old fashioned algos you bring them to the level of potential winners in contests 15-21 and that is more important to me then $5000 which I can make in couple days.

In Sept 2016 there will be 15 winners and only 3 disqualified.

Out of sample metrics of disqualified algo:

annRet_pt 20.29%
annVol_pt 7.58%
sharpe_pt 2.48
maxDD_pt -3.37%
stability_pt 0.93
sortino_pt 3.66
beta_spy_pt 3.16%

corr_pt 0.07

Is it not hedged enough by any reasonable measure?