APD is meaningless.
AHHWAHHAHH!!!
APD is meaningless.
Glad to see everyone is beginning to come around to exposing the sham statistic called "APD" - I was the first at C2 (that I know of) to attack it (see post in "December 10-14, 2007" on Results II: Stocks-Options forum).
My vote, naturally, is that APD is meaningless.
. . .how can a statistic mean one thing in certain instances and something else with others?
Definitely sounds meaningless or at best needs clarification.
Gilbert
Thank God I am not a statistician or an accountant.
According to the APD, Pegasus should not be on the "best systems" list. But then again, neither should any of the other systems be on the "best systems" list.
Useless stat in my opinion.
Pegasus
So we debated for a week whether we want one or two sugar in our coffee, and then everybody voted to have icecream.
LOL.
As I tried to show in the theory, APD is adaptive and comparable to the Sharpe Ratio. Thatâs why I canât vote to dismiss the APD as useless.
Both are measures of excess return, but while the Sharpe ratio had nice formulas to go with it, followed by a nobel prize, I can see the APD ratio having the same effect, especially on investment managers like myself to really keep a tally on how much excess return above our losess we are generating.
Weâve noticed that long term holds when they are timed properly have outsized APDâs, as was the case with several buy hold systems with hold times out past 2-3 weeks. As long as they were timed properly, and MVP-3 is a perfect example of decent entry timing for long term holds, the shorter term systems did not fare well in this.
However, theyâre still relevant. Iâd definitely not dismiss this outright, because it would be foolhardy not to say that there is no value to the statistic.
It hasnât been calculated because of one thing: data. Now that we have another element to add to the EMH, we need to start applying it to other managers. The current investment management side is base strongly on 10 year performance, without regard to risk. We are only left with annual standard deviations, and I think were we to apply APD outside of an amateur context that we might find some of the best on C2 outperform numerous asset managers with billions under management. The problem? We donât have the data, which is the only limiting factor in its applicability.
I have a way to backtest all of the top systems on wealth Lab, and I think that would be a great place to start ranking the APDâs to really get a feel for what a good one is.
Iâve provided the example of my own systemâs backtest to show that 0.4 would be nice, but more likely, 0.3 on this stat is a decent long term value.
Thereâs more to the APD that needs to be discovered, in my opinion, and as somebody on his way to the CFA charter, I donât think you should ignore me on this. I go way down on all these measures of performance and I would find them quite a bit more useful than even the sharpe if we had a better context to asses a decent value.
We know 1 on the Sharpe is high, but I think 0.3 is high for the APD.
"There is nothing one can do to or with such an arbitrary measure as the APD ratio except sweep it aside. In the absence of evidence, there is no way to consider any idea, on any subject. An arbitrary idea must be given the exact treatment its nature demands. One must treat it as though nothing had been done or said. The reason is that, cognitively speaking, nothing has been done or said. One cannot allow into the realm of cognition something that repudiates every rule of that realm."
Posted in February of 07 by one of the first to see the true reality of the APD.
How could anything be used without evidence? Because it makes us feel good? Because it looks or sounds good?
There is nothing one can do to or with such an arbitrary measure as the APD ratio except sweep it aside. In the absence of evidence, there is no way to consider any idea, on any subject. An arbitrary idea must be given the exact treatment its nature demands. One must treat it as though nothing had been done or said. The reason is that, cognitively speaking, nothing has been done or said. One cannot allow into the realm of cognition something that repudiates every rule of that realm."
Itâs a very whimsical statement that doesnât really say anything.
I donât find it arbitrary. What other evidence do you need besides the APD distribution for the systems described in their stats?
1 Like
4% between 0.2 and 0.4, 4.4% between 0.4 and 1, and 6.5% above 1. The problem with that distribution is that weâve had a lot of new systems come on in the last two months. Some havenât built up enough time to say anything about their long term performance.
Weâre saying approximately 15% of systems have an APD of 0.2 or better. I think thatâs enough proof for me to say itâs tellling you something. 85% lie below 0.2.
Need any answers, as it shows a negatively skewed distribution of systems, which is what one would expect, and the outliers doing well usually revert to the mean. The ones that donât out past 1 or 2 years are good systems.
"APD is adaptive and comparable to the Sharpe ratio."
The Sharpe ratio includes the contribution of Money Managment to the trading solution, APD does not. If you have all the trades of a strategy, but without the dates, you can compute APD. You canât compute the Sharpe ratio.
Some people believe Money Managment is even more important than the timing strategy. How can you embrace a statistic that ignores the Money Managment?
This was spoken by Palsun. His anger against APD was summed up in his multiple strange beliefs. Particularly, that a 95% drawdown does not matter unless it is realized (tell that to brokers giving you a margin call).
He also railed against C2 and Matthew and the trading establishment as well.
I am really glad to see you quoting him! It thoroughly establishes you as two steps below a newbie.
Oh, Palsun, Palsun, Palsun, PalsunâŚto be remembered as the guy that was exposed by the APD.
No kudos for Futurm for pointing out this ridiculous post.
Both have the same aspect to them, Keith. As Iâm noticing I can increase the APD by decreasing the position size in my backtest on WL to avoid using margin. So, because of that fact, I can say that it doesnât ignore money management quite the way youâre describing.
"Itâs a very whimsical statement that doesnât really say anything."
It simply says there was no evidence for using the APD. All there really was is look good feel good for some!
"I am really glad to see you quoting him! It thoroughly establishes you as two steps below a newbie."
So you create and embrace negitives in your mind. Surprise!
Your constant droning against APD belies the facts. it is a simple calculation, based on a systemâs numbers. It is the same every time you calculate it. It does a very good job flagging systems that tend to run up large drawdowns compared to net profits, exactly as averaging down and Hold&Hope systems do
Are you blind or just sightless?
Now, where is your evidence that there is no evidence? I donât see your tests anywhere, just your opinionâŚ
Beau,
The Sharpe ratio is computed using the standard deviations OF THE EQUITY CURVE. The equity curve is the result of trades filtered by money management.
If I make a long and short equity trade each day because my money management dictates a equal long and short exposure, I will get a resultant equity curve. If each day one trade wins two points and the other loses one point, I will make a point each day, the equity curve will go straight up, and the Sharpe ratio will be high. If I take the same trades, and plot all the losers first, then all the winners, the equity curve will go straight down and then straight up. The Sharpe ratio will be much smaller then in the first case. But the critical point is that APD will be the same in both case. APD ignores the effects of money management.
