I say this only to remind you that the ultimate outcome of this discussion could be a more complex algorithm – a process, and not a formula.
For example (and here is an idea I have not reflected on in any way!): The algorithm could first determine the trading "style" of the system, and exempt from APD risk flagging any system that trades with a certain style (i.e… long-term trades, no averaging down into losses).
This was intriguing, because it would then show systems comparable to others in that style, and not to every single one, based on the APD.
I think diversified portfolio stock futures or forex portfolio, futures/stock/forex (high, medium, low) leverage, trading style, medium-term, short-term as ranked on the best lists, would be a very interesting avenue to explore.
"Unfortunately, there has been a trade-off, because the APD – while warning people about certain trading styles – also catches in its net a certain class of systems which may not be as risky as the APD indicates (and may, in fact, on a long-term basis, be less risky than those systems not flagged by the APD).
To me, the onus is on people who say “System XYZ” is unfair. I have yet to see convincing proof of APD’s “unfairness” - demonstrated via several current, relatively popular C2 systems, using ironclad testing methods that prove this unfairness. But I do see a lot of speculation and rants.
For example, someone mentioned that it does not apply to multi-instrument trading, which is patently absurd. IF someone does 30YR TBonds at 9am and ES emini at 1pm, that has nothing to do with APD’s effectiveness. Or if holding multiple instruments, yes it diversifies you, but it is QUITE possible that they both go against you at the same time.
Second, the concept of hedging. When people claim this, quite a few here have argued that the concept of hedging is simply the same as being flat. I would not like to see a loophole allowing people to hedge just to AVOID a low APD. There are spreading systems that are long one thing and short another similar instrument (like a covered call). If C2 had the concept of a spread or hedge trade, then the trades could have a combined APD. But barring this…
In general, I think suppressing APD for some systems grants a loophole, and denies potential subs that info. Again, many stats from Realism to slippage/commission calcs have been argued broadly. But they remain, and subs decide what to do, warts and all.
The real thing that APD and other stats do, is force vendors to practice better money management. I constantly see systems, that will start with 2 contracts, and go up to 8 or 16 trying to hammer down a large loss. Or a system with generally small losses but also some very large losses Sometimes this is rare, but even one careless trade can blow a sub out of the water. Some of the money management on C2 is atrocious, and statistics flag such bad behavior.
The concept of only applying APD, or any other stat on a rolling basis is also questionable. Few systems here have a long life anyway. The stats should all do the same thing, not a few on the whole life of systems, and a few just on recent activity.
Finally, making the APD a complex calculation also detracts from being a simple one for traders to understand. Most traders “get” Profit Factor, unlike Sharpe, which few traders likely grasp well its meaning.
Yes, it is intruiging, but I have no hope that we will shortly arrive at anything that is acceptable for everyone. It will take months, if not years. Look what happened in this discussion: We have now 100+ posts and still no agreement over something so trivial as the mouseover text. And even if we succeed in reaching consensus about a new statistic, it is very likely that it will be displayed in addition to APD, as opposed to replacing it. And that is also the wise action to do, at least initially, until there is clear proof that the new statistic is better than APD in every way.
So I predict that APD will stay here for a long time. Therefore I really urge people to vote about the mouseover text. This is something we can improve now, today. After this we can still discuss all kind of alternative formulas.
I’m agree about the the mouseover text change of jon
I vote we implement a double-bold font option, that only Ross is allowed to use.
Apparently he doesn’t understand the purpose of bold. Or, maybe he does, and thinks every word of his post deserves bold treatment.
In any case, he definitely needs more options.. I suggest the tag be <douche></douche>.
or for yourself:
<ninny></ninny>
hehe So the game in democracy is over.
Well… I think the logic of the APD’s show is perverted. With all respect to Joe SixPack systems vendors asked legitimate questions. If we have APD please, please, please show us a study that proves following assumptions.
"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."
There isn’t timing in APD i.e. profits/loss distributions. I’d like to see a proof for the assumption.
"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."
For those as well. Give me the proof.
"In summary: a high APD is good but a low APD is not always bad. "
It means 50/50. What is the high and what is the low? It’s known for Sharpe.
The definition gives you no clues about boundaries of profitable system. Only baseless guessing. It was shown numerous times that APD doesn’t work as indicator. It cannot be practically used. Only one purpose of APD is self-convincing to not trade. If we want to save Joe SixPack just put red bold blinking “do not trade”. It will have the same effect. Otherwise the Joe should at least understand business standards and respected trading statistics. APD isn’t the business standard, it has no studies it has nothing related with trading except understandable paranoia of “do not trade”.
So my opinion is still the same. APD should be removed from C2, not masking it under new description. Put red bold blinking “do not trade” there instead of APD 
Eu
Have you not listened to a word that’s been said in 150 posts?
""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."
There isn’t timing in APD i.e. profits/loss distributions. I’d like to see a proof for the assumption. "“
Start by looking at any high win % system
”"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.“
For those as well. Give me the proof.”"
Start by looking at any long-term trend following system
You have a lot of catching up to do.
(sigh)
There is the same perverted logic. Why I have to prove anything?
I’m asking legit question that has an answer for any trading indicators. Where is your answer? How you define the APD boundaries? In other words what is “low” and what is “high”? Without the knowledge your description is useless as APD.
Eu
Things are not always as they appear, that is the reason for studies and proof. Perhaps some think an exception should be made for APD. Why, to make some feel good about who knows what. Just so our hearts are in the right place is all that matters to some, but that does nothing to help subscribers or C2.
It was a mistake to accept APD so quickly, it is now time to correct the mistake.
This was just suggested to me as an alternitive to the APD.
The Ulcer Performance Index: http://www.tangotools.com/ui/ui.htm
I’m not familiar with it, anyone have any comments.
You are such a noob.
Go back and read the history, EU. I don’t want to rehash all the theory that has been put forward in the last two years, two years of which you’ve been around.
It means 50/50. What is the high and what is the low? It’s known for Sharpe.
Right, at the heart of the problem. I’ve backtested on 1000 scripts running all day. A good one from that was an APD somewhere between 0.2 to 0.3 and above. Less than 0.2 nearly always had pf’s below 2, and sharpes between 1 and 2.2.
In many ways in many things, that is what makes life interesting.
You’re forgetting no one knew a good sharpe ratio, either when it was introduced.
Wealth Lab is the only solution available for us to see what a good 5 year or 1500 bar backtest would show, and I’m saying the good ones are between 0.2 and 0.3 and higher. Mostly clustered in that 0.2 and 0.3 are systems with sharpes between 1.25 and greater with pf’s of 1.9 and higher.
If you’re looking for a good ratio, I think 4 years as the most profitable chartscript ever backtested at wl4.wealth-lab.com which is what the [LINKSYSTEM_35438029] is based on at 0.3 is a decent long term APD.
Since there’s no one here with a publicly available system confirmed by a 3rd party, and based on its history, it’s somewhere around 0.3.
If you’re lower than 0.2, I probably wouldn’t trade it. Most systems won’t last long enough to see what their real APD is, so when they’re killed below 0.2, I guess that’s what their long term APD really was, going with the rest of the 85% of systems here.
You’ve totally lost sight of what we are trying to achieve. I am not the inventor of APD, I am not defending it and I don’t have to prove anything, the main premise of this whole discussion from 150 posts ago is that the existing description is inaccurate and misleading to people discovering it for the first time, that is what we’re trying to correct and I believe we’ve collectively come up with something that does it effectively. Thanks for your help.
I liked it better than the current version, provided we’re still showing the distrubtion, but it is a timing statistic.
The APD will definitely spot a holy grail system, but it will take a long time for us to find one.
Anyway, we need to put a sentence somewhere that says:
"When analyzing the APD statistic of a system, consider first its style with respect to position sizing and the length of trades. A higher APD will indicate superior timing capabilities of the vendor. These may be temporary due to the system being too young, but over time, a higher APD will exhibit both a correspondingly higher Sharpe Ratio and higher Profit Factors."
That’s what’s needed I think to get people thinking about the statistic the way they’re supposed to. Especially newbies to the site that give credence to system statistics younger than a month.
I’m through debating what the APD tells you, because I’ve explained it to you, and even Ross with his sampling does show what I’m saying as well. Furthermore, just thinking about what components make up the APD stat will lead you to this same conclusion.
The debate about what a good one is I hope I’ve answered for you, since I’m the only vendor that has the APD stat programmed into his system performance stats page. I think Ross’s guess at 0.4 was a bit optimistic, but probably around 0.2 to 0.3.
"“When analyzing the APD statistic of a system, consider first its style with respect to position sizing and the length of trades. A higher APD will indicate superior timing capabilities of the vendor. These may be temporary due to the system being too young, but over time, a higher APD will exhibit both a correspondingly higher Sharpe Ratio and higher Profit Factors.”"
OK, let me see if I can disect what you’re saying, as there’s an awful lot of definitives in there, you’re saying 'A higher apd WILL indicate superior timing… will it, or could it?
And if a higher one indicates that then does a lower one indicate poor timing? Not necessarily, but people might conclude that if you definitively say high apd = good timing., just as the existing description currently implies bad apd = bad system, period. You might want to avoid that kind of rigid statement.
‘over time a higher apd WILL exhibit both correspondingly higher Sharpe and PF’… will it? or could it?
and higher than what?
You’re basically saying if one is high all 3 WILL be high, but if that’s the case it would already show that in those statistics so there wouldn’t be any need to state it, would there? I’m not sure referencing other statistics is useful, I think it’s better to interpret the APD figure in itself without reference to other stats.
Personally I’m still not convinced a high APD indicates superior timing, all it indicates to me is that the profit is high relative to the drawdown, that is afterall exactly what’s being calculated, and what’s stated in the proposed new definition.
