Hi Sally,

You are correct, we updated the naming convention on the winners page. Previously it said "February Contest" next to Grant's name. It now says "March Prize". We have changed the naming convention to refer to the contest winners by the month that their IB account is funded and they start trading real money rather than the contest month. The prize dates and contest deadlines remain the same.

Sorry for the confusion,
Brian

Oh, that makes a lot of sense Brian. Thank you for your answer. If I understand you properly then the August winner has already been selected and funded but her performance is not being displayed yet?

That's right. The winner is announced, but the money isn't trading yet. The announcement is here.

I repeat the apology for the confusion. Our naming convention wasn't handling the fact that there were two contests starting in August, so we started describing them in terms of the prize month, rather than the start month. July start == July contest == August prize.

Dan

Dan, thank you for your informed response. I appreciate the link to source material confirming what you said. Now it makes sense. Looking at your rules, I also read that

If the Trading Account's value drops below $90,000 USD, the algorithm
will be stopped and the positions liquidated.

I just saw that Grant Kiehne's balance dropped to $89,886.45. Does Quantopian use intra-day trading account balances for stop-loss control or end of day balances?

Thank you in advance for helping me understand the contest and the management of the crowd sourced hedge fund!

Yep, unfortunately Grant hit the $90k stop-out this morning. His account was shutdown and the positions were liquidated. I'm bummed, because I really wanted to write a check to Grant.

I'm optimistic that we get to write a check to Simon next month, instead. That will be a great day!

Thank you for the reply Dan. I agree, it must feel horrible for Grant to be the first one who won and then lose it all on a stop loss.

The bigger picture, based really on scattered observations, is that three of your contestants are long the market and one is short the market since it appears that Grant, Szilvia Hegyi and to a lesser extent Jeff Koch go down when the S&P 500 suffers a decline and have positive performance when the S&P 500 increases. On the other hand, Simon Thornington appears to have a negative beta. It may be too early to tell about Michael Van Kleeck; his performance has been great with low correlation to the market, which may be due to a new fund that was well hedged.

Is daily performance of the contest winners available for analysis? That might be a very useful exercise and either refute or confirm my admittedly anecdotal observations.

I agree that Grant looked like he was long the market. I think we fixed that, largely, with the July Prize (formerly known as the June contest) when we made low-beta a forcing factor in the scoring.

We don't have a mechanism for sharing the contest winners' results. I think you might enjoy studying the leaderboard, though. It has a CSV every day with everyone's risk metrics.

Thank you Dan for your reply. It is unfortunate that the pattern of returns for each contest winner is unavailable, perhaps you may want to consider making them available. Pretty much any trading account has daily balance information. My understanding from your contest rules is that Quantopian manages the trading account, so the daily balances must be available. While derived quantities are interesting the return information is much more illuminating.

You said "we made low-beta a forcing factor in the scoring" which brings up an entire topic for another post about the criteria used in selecting investments whether for contest winners or for algorithmic trading. Some might say that is the most important thing. I am certain to have a lot of questions once I reach that stage of my research but I am still reading prior commentary on the community board about this issue.

Thank you again for your prompt and informative replies to my silly questions and please let us all know when you start collecting and publishing daily returns for contest winners.

I know some folks have been scraping the daily winners' returns, but I don't recall who. My algorithm is based on volatility, so it probably has a negative beta by virtue of volatility having a negative relationship to the stock market indices, but it's not a direct effect. It's also worth pointing out that my algo would not place in the current contests, so it's probably futile to try and reduce it.

Wow Simon, volatility trading can be very powerful. I am impressed. I thought Quantopian only had US equities which shows what I know so far. Are options trading algorithms also possible? I know Interactive Brokers also allows trading in options. In fact, IB's parent is Timber Hill, an options market making hedge fund.

I believe you are correct when you make the assertion that "volatility having a negative relationship to the stock market indices". Which makes you wonder about algorithms chosen by maximizing Sharpe ratios...