The Curves!
For the first time, I am releasing the performance of one of the main unleveraged accounts I am managing. This performance doesn’t include any leverage and the graph is scaled to an initial equity of 100,000. The management started around 4 months back, 25th of March to be precise.
Technically, the profit shown here is the bare minimum, “open-for-audit-if-you-want-it”, type of profit. Includes friction and all the costs a real trading account will face. But in its reality the real profits are slightly more than this. The extra component comes from a few option trades I have taken to run a side speculation via one of the other systems I designed recently. Quite notable was the option trade on June 8th, where I loaded up 4400-4500 puts, quite heavily, to see their price shoot through the roofs, 10 days later. Also involves a few duds, like the one I am currently in, where I loaded up 4400 calls, at 150, but still couldnt show a very encouraging trade[ as I am writing, it tanked from around 250 to 139ish levels in 2 hours straight, Nifty falls 4%]
The Profit Curve
Till date, the return shown is 37%[since March].
Hmmm.. Warren E. Buffett in the making?
Keep it rising
But would have wanted you to register more before, er, recording your readings in public.
Thank You! M’lady.
I just wanted to have it posted regularly every single quarter
Interesting. As Jaishree mentioned, a bit of more data would have given more weight. Is this Cash Segment based – i.e., Cash Stocks or Index?
Also, if I am not wrong, is the trade of 10 and 11 just before Elections?
Do you have any OOS testing results. Since you use AB, if you have data which is not used for testing, maybe you can use the same for a longer period.
True, very true.
Actually, I was myself reluctant to put it in public domain, because of such low number of trades. But there are two things, its an unleveraged portfolio, hence does not take many trades.
More importantly, its a slower strategy, hence we have got just 16 odd trades in this period.
10th trade is closed, on 8th May; 11th is closed on 9th June. Both were initiated in 4th April.
As for OOS, it doesn’t hold much water in this case. These quarterly results are rather a platform to make my results -good,bad or ugly, public.
Its a place where “you eat what you kill”. If I can’t give any advantage to my client, then I dont have a right to pretend in a holier-than-thou fashion.{Like what we see in electronic media}
There is no attempt to cosmetise the results.{In fact, just the opposite is being done}
Soham
Soham,
Reason I mentioned OOS is because many a time, its difficult to have a long track record to put out in public domain. Also, a OOS / hypothethical test gives details such as Max Drawdown / Risk/Reward ratio, No of Winners and Loosers, etc which provide some food for thought.
Comparing ourselves (I am a trader myself too) to guys on TV is a no go. They change their tune depending upon what the market is saying at that point. Hence, you can see the guy coming on CNBC in moring and being bullish and coming (after a fall) on NDTV profit and saying the reverse. They are there just for entertainment, if anyone places any trust, god help them
.
In that way, I run a risk of presenting something which is not really there. Can’t be audited if someone challenges me… sorry, no go!
But again, OOS tests wont help in understanding the Drawdowns and Returns etc, Monte Carlo Analysis will…
But then, you have to take all those things with a pinch of salt. The current tools of statistics are too delicate to take care of such “blunt” applications like performance simulation
>> Can’t be audited if someone challenges me
Unless one uses something like TimerTrac or collective2, its difficult to show (even with real trades in a real account) that the trades were carried out in accordance with the signal. What a auditor at max can certify is that you have really put the trades in a live environment, but in no way can assure (as far as I know) as to whether the said trades were the result of some algorithm one uses.
Yeah, If you are having constraint’s due to Inadequate Capital, Monte Carlo may be a good way to analyse what if scenario’s when you select all results rather than picking only a select few. Of course, this can also mean that a good system shows up as bad and vice vera. Never been a holy grail in testing
Anyway, keep typing away. Good thoughts that you present.
I would agree, completely…
In fact, I have plans to move onto TimerTrac after I grow the AUM to a respectable level. But they usually want you to have atleast one Ryder based stategy{or something to that tune} Atleast the basic auditing, that I have taken live trades is a good point to start over.
I think, thats the advantage of Monte Carlo. It shows potential wrong moves and tells “hey-you-got-fair-chance-of-hitting-your-head-around-this”. Which I would say, is a plus point.