Before we move on to the next stage let’s just summarise what we have so far on the narrative to which all those bar one who have commented think is better than the existing description.
“The APD Ratio (“Average Profit to Drawdown”) measures how much profit is made for each dollar of drawdown suffered while a trade is open.
A high APD means that on average, the system makes trades with high returns in comparison to drawdowns.
A low APD can result from averaging down or allowing open losses to run in the hope they recover to achieve relatively small profits. This is considered a high risk strategy.
Other possible causes of low APD are the system applies a hedging strategy, or the system has many small losing trades and less frequent but large winning trades. These strategies do not necessarily imply a high risk.
In summary: a high APD is good but a low APD is not always bad.”
The next stage is as follows:
As we have shown that you can have a low APD but not be trading a high risk strategy it would be wrong to use APD to determine system ‘riskiness’ as it currently states. On a side note it’s incorrect anyway to say it measures riskiness when the idea is to have a higher number, in other words if it measured riskiness then the higher your APD the more risky your system would be. That is not the case here.
So in addition we need to change the ‘Where this system stands’ part below the description so that in the following example;
12.2% of systems have less-risky APD stats than Corporate Investments’ 0.31 (and 87.8% have more-risky APD stats) we should change ‘less-risky’ to better, and ‘more-risky’ to worse.
Finally below that on the ‘General distribution of APD stats on C2’ the ratings should be in terms of the APD reading itself as to whether it is a good number or not rather than as a measure of riskiness so where it says
’…of 0.20 or worse (high risk)’ that should change to read (low) and
’medium risk’ to ‘medium’,
‘low/medium risk’ to ‘high’
‘low risk’ to ‘exceptional’,
so now they are measuring the relative rating of your APD stats rather than judging the riskiness of your system, this then also means we can totally remove the last line of the existing description “An APD of 0.40 and higher is good, 1.0 and higher is exceptional” because it would now be redundant.
So to summarise, we have incorporated a lot of feedback into this final version and I think we have covered the vast majority of reasonable concerns, start to finish the pop up would read in it’s entirety as follows:-
The APD Ratio (“Average Profit to Drawdown”) measures how much profit is made for each dollar of drawdown suffered while a trade is open.
A high APD means that on average, the system makes trades with high returns in comparison to drawdowns.
A low APD can result from averaging down or allowing open losses to run in the hope they recover to achieve relatively small profits. This is considered a high risk strategy.
Other possible causes of low APD are the system applies a hedging strategy, or the system has many small losing trades and less frequent but large winning trades. These strategies do not necessarily imply a high risk.
In summary: a high APD is good but a low APD is not always bad.
Where this system stands: 10.7% of systems have a better APD than Broadsword Macro’s 0.41 (and 89.3% have worse APD stats)
General Distribution of APD stats on C2:
85.1% of systems have an APD of 0.20 or worse (low)
4.0% have an APD of between 0.20 and 0.40 (medium)
4.4% have an APD of between 0.40 and 1.00 (high)
6.5% of systems on C2 have an APD of 1.00 or better (exceptional)
I agree with that. I would suggestion though: Perhaps change
"85.1% of systems have an APD of 0.20 or worse (low)"
into
"85.1% of systems have an APD of 0.20 or lower"
and similar for the other statements. That is a more neutral formulation.
how about ‘have an APD of less than 0.20 (low)’
That’s fine.
"
Other possible causes of low APD are the system applies a hedging strategy, or the system has many small losing trades and less frequent but large winning trades. These strategies do not necessarily imply a high risk.
In summary: a high APD is good but a low APD is not always bad.""
Disagree completely. Again, if other stats do not do this, neither should APD.
In addition, this remains theoretical. And further, it has not been shown the systems that even fit in this category would not eventually fail as a result of ringing up a lot of small max DDs.
Again, if we do this for other stats, I am more amenable for APD. Until then, there is no reason to modify APD.
I understand the small and occasional winning ones, but again, APD does not need to be singled out from the other 10 or so stats here.
GRRR - Cancel last 3 posts - this is what you get for posting after a run and your heart is pounding.
This is my point:
1) I disagree with a line of reasoning that APD must be changed in absence of analyzing and modifying the other 10 stats as to "cases where it does not apply."
2) Theoretical cases have been presented that have not really been demonstrated as being unfair to current systems, hedging or occasional large wins. Show me three of each in current, mildly popular systems that are unfairly treated. Otherwise, every stats has "flaws" in rare events, and APD needs no more description that the others.
"In addition, this remains theoretical."
And so does APD, you managed to get it used here without ever having to show any published paper or proof, totally unaware of it’s limitations. Now that it has been in use for a considerable time we are discovering various instances where the current description would be misleading, therefore we are making it more accurate. Yes you are right other statistics have their flaws also and some will benefit certain types of systems over others, but none have descriptions that are misleading or inaccurate. That is an important distinction and one that you don’t seem able to recognise.
And so does APD, you managed to get it used here without ever having to show any published paper or proof, totally unaware of it’s limitations.
Correct. Let’s see where APD will be misleading as risk indicator:
- trend following systems (as it was shown)
- “scalping” system where avg loss>avg win in few times
- pairs trading, where risk is overlapped by different positions in a pair (forex/etfs/options)
- portfolio trading where portfolio is made from overlapping instruments. Close to pair trading, but on a little bit higher scale.
The limitation practically covers most of trading areas 
Where the APD works as risk indicator?
- nowhere
I probably have to apologize for C2 community. I supported implementation of the indicator for fun with kind of knowledge that it’s useless in general. My idea was that smart people will understand uselessness of APD and for others it will always show “high risk” that isn’t bad, because a person has to dig in to APD marked system before subscribing.
Eu
P.S. In general, APD is useless.
"but none have descriptions that are misleading or inaccurate. "
Well, that is a convincing argument. How much research have you done into this, or is it something to say that sounds important?
Demonstrate to me that none of the other stats is neither misleading or inaccurate? Seriously? Please providet a comprehensive list of this below. Show me that these other stats do not treat any current C2 system unfairly, either real or in theory. Show how they are DIFFERENT than APD.
I "managed to get APD here", because I developed and used it, saw the value, and offered it to MK, and Matthew contacted me about adding its functionality to the site. Matthew and quite a few others knew the serious gaming that went on with equity curve, winning percentage and profit factor, and this was the only statistic that blew holes in the smoke and mirrors.
I will not be participating further in this exercise. This continues to be based on rants and theory. Where is the list of unfairly treated systems I requested? Or is it easy to stick with theory?
There is no need to provide a list of systems, what purpose would that serve? You’re losing sight of what we’re trying to achieve. We are trying to make the definition more accurate. No other statistic on C2 has such a rigid interpretation. The description for APD currently says "Low APD means a system allows open losses to run in order to achieve relatively small profits (bad)"
That is pretty clear, it’s not saying it COULD mean it does that, or that’s ONE of the interpretations, it’s basically saying if the APD is low then this system allows open losses to run. Period. No if’s and but’s, that’s it, no other way to look at it, bad APD = bad system.
THAT IS INACCURATE AND IT IS MISLEADING. You yourself said you could see our point. So in principle you agree, all we’re arguing about now is the best wording that would accurately describe it.
You should be thanking us, we’re making it a stronger more useful statistic which people would now have a better understanding of it’s purpose, how to interpret it and it’s limitations. You’re welcome.
I will not be participating further in this exercise. This continues to be based on rants and theory.
Well… Are you speaking about yourself? You didn’t provide any prove before implementation of APD. you provided only theory and rants. No prove. (see I can use bold as well as you ) At the moment I would agree that APD is outdated and misleading and probably should be removed from C2 as nonsense. Otherwise you has to provide a prove of usefulness as you love to requested from others. Prove it. Give us mathematical/statistical prove instead of whining.
If Bob (not you as it must be) will publish his research I would be second to check it. And first for removing APD from C2 if his research convince me. At the moment I can show on examples that APD doesn’t work as risk indicator in cases that I mentioned.
Personally you can use whatever you want.
Eu
We are trying to make the definition more accurate.
There isn’t any need in definitions, because statistically APD shows you nothing.There is major flaw that makes APD useless as risk indicator.
It’s good time to remove it from C2.
Eu
Well without absolute proof one way or the other as to it’s authenticity I think we have to respect MK’s original decision to include it in C2’s arsenal of statistics.
Even Ross agreed that he could see what we were saying about the current definition being too rigid in it’s interpretation so in the meantime I would be settle for APD remaining as it is but with a more accurate description alongside that is clear and balanced as to it’s uses and implications so as to not mislead people. Without proof either way I think that is the fair and reasonable thing to do.
Well without absolute proof one way or the other as to it’s authenticity I think we have to respect MK’s original decision to include it in C2’s arsenal of statistics.
Well… of course MK can do whatever he wants as the business owner. However I would not agree that APD is statistics and has statistical proof as any other respected trading indicator (i.e. Sharp) Only one unique future of C2 is RF indicators and it’s based on real life data, not somebody fantasy as it should be in fantasy world.
I would be settle for APD remaining as it is but with a more accurate description alongside that is clear and balanced as to it’s uses and implications so as to not mislead people. Without proof either way I think that is the fair and reasonable thing to do.
Clear definition of APD is its formula. It’s only one thing that you can put to its description. The formula covers nothing from real life trading world. It gives useless number as it might be in fantasy world. Period. So there is no need in definition until author of the formula provide a study that proves the formula shows risk indication and it should be a “must” before implementation. Uselessness of APD is visible practically and I’d like to see statistically based proof (at least theoretically) instead of whining of APD author.
Otherwise it’s better to remove the joke from C2.
Eu
I agree with everything what Jon said in this thread. I also agree with most of what Eu said. In addition:
The explanatory text of other statistics is completely irrelevant for this discussion. Statistics have no feelings and therefore the argument that they should be treated equal is a fallacy. The only thing that matters is that the description is accurate and informative. The existing C2 description is not accurate, and the short description that Ross suggested is not informative. The description of Jon is accurate and informative and therefore it is the best of these three.
Listing systems that are treated unfair by APD is futile because whatever system we list, it is obvious that Ross will reply that it is a bad system. In fact, the so-called evidence for APD is largely created by circular reasoning. I have seen many posts where the inventor says that a system is bad because its APD is low. Having done this a dozen of times, it is now said that APD is valid because it supposedly identified so many ‘bad’ systems.
But I think that by now we all have exposed our arguments three or four times. It is futile to repeat this a fifth and sixth time. I suggest we simply vote.
I vote in favor of replacing the current C2 text by the text that Jon developed in this thread.
But I think that by now we all have exposed our arguments three or four times. It is futile to repeat this a fifth and sixth time. I suggest we simply vote.
I don’t believe in democracy and it has practical proof nowadays from US government with "privatization of profit and socialization of loses"
I vote in favor of replacing the current C2 text by the text that Jon developed in this thread.
Anyway. My vote is for removing ADP from C2 stats as meaningless risk’s indicator that has not statistical base nor theoretical study as any respected trading indicator.
Eu
P.S. The discussion was futile from beginning and the vote is futile as well.
