Good for my filter that it is banned now for 2 years as it should be
This is the old problem.ZVQ was in such situation a year ago(even worse DD) and recovered during the day and they showed like nothing happened.
I don’t want to believe the “flexibility” of the DD values - do you have an explenation? Please compare against the screenshots above
For me this is an error which must be fixed.
It appears that DD is calculated on daily close. If today HFD recovers and closes in positive area, DD will go back to about 10%.
Although I don’t agree with your attack to HFD, I agree with this.
I also took a screenshot because I noticed that drawdown was reducing. Not just for HFD, I noticed this also while I was evaluating other Darwins.
Only February performance shows -11.91% in my screen (it was not even the worst moment), how come Drawdown shows -10.77%.
Not cool for potential investors, in my opinion.
Thank you for raising this topic! Current precision of DARWIN DD data is now explained in the following topic:
We’ll evaluate to only update “DD since inception” if the new value is lower to not lose resolution over time.
Thank for clearing this point.
But it would also make sense to show the real max. DD on the performance statistics as @FedericoSellitti pointed out.
You have the high and low data at least when you show a candle stick chart. I would also like to see that on the filter criteria. Currently it is hard or not possible to set a stop loss on the value where the real max. DD is exceeded.
I’m confused about this also, we have the data to show high and low of the h4 candlesticks in the Darwins Return / Risk tab. I’m not sure what level data these H4 candlesticks are from but they seem to be more accurate than whatever the DD stats are calculated from now…
Also, people are showing screenshots of decreasing DD stats. The DD was calculated more correctly at some point… just freeze the DD stat from decreasing. It should never decrease!
There are a couple of reasons drawdown since inception is not calculated off the candlestick data.
- There are no candles for the period when a strategy was not yet listed as a DARWIN asset.
- Candlestick maximum and minim values don’t offer enough information. We have to know the date and time as DD is a posterior minimum following a previous maximum. Inside a candlestick, we do know the minimum and the maximum but not in which order they were produced.
Currently, intraday DD is not reflected in drawdown since inception when more than 3 months have passed since the intraday DD. This is what we seek to solve with only updating the value if the new value is lower.
Yep, this is what we’re now reviewing to do.
For that I am talking about the max. DD of the performance statistics. That should NEVER decrease as any other max. value.
For some reason you guys prefer the Max. Drawdown (next to DARWIN return graphs) value to the DD since inception (upper section of DARWIN profiles) value.
But it’s the latter that’s more accurate!
In the moment the draw down is occurring, it seems like the estimation is sometimes more accurate. But even the estimation on a 3 month chart with D1 candles seems to be using open and close of the candles instead of high’s and lows. I’m talking about first a candlestick prints a new high then a later candle stick prints a lowest low since then. DD should not be less than that difference, correct?
When two candlesticks are involved we can use the high of one and the low of the other. Even if there is a big swing inside a single stick we can safely use the high to the close or the open to the low.
This is correct and it would enable us to improve the calculation of drawdown, We’d still need to solve the problem of no candlestick data before the DARWIN got listed on the Exchange.
This assumption is not correct. We can’t estimate within the same candle.
It’s certainly not impossible to “estimate” what happens inside a candle. MT4 does it either via a tic generator or via OHLC for instance.
We know the candle open happened before the candle low and the candle high happened before it closed so we can sometimes get a more accurate/pessimistic reading (without overstating) that way if the candles are a short enough time, can’t we? I know it would be inconsistent in that a swing inside a single candle might not have the same accuracy as an equity swing over 2 or more candles but over-all it would be more accurate/pessimistic and if the candles were small enough time this might still be OK, no? I think we want to show the most pessimistic values we can here, don’t we? In fact, is risking overstating DD by always assuming the worst (high happened before the low) really that bad of an option for estimating? (Especially if the candles are a short time.)
I knew you’d say that! There’s always the option of omitting such data entirely from stats, charts and filters, this would solve the account farming concern and the chart miss-match issue as well. I know some investors would prefer to at least have that as an option.
Otherwise we could (omitting pre-migration?) “produce” pre-listed Darwin candles and show/color them differently (This would also solve the chart miss-match issue). Easier said than done I know but maybe it’s not impossible to calculate just once and then record it.
On another note, MyFXbook calculates DD tick by tick. If this is too resource demanding to do for the more complicated instruments that are Darwins, how about tick by tick or second by second or even minute by minute DD calculated and recorded for the underlying account and then “estimate” the darwin DD from that? Wouldn’t that be both reasonably do-able and miles more accurate than a Darwins daily close?
You should not call something a maximum if you don’t show the maximum.
‘estimated maximum’ would be acceptable on the explanation text.
Or even ‘estimated maximum based on 6h / d1 close only data’ with a link to the precision article. Hopefully alerting investors that the intraday DD spikes might have been much larger.
Thanks for the effort put in this suggestion! The team is already working on this. We’ll keep you posted on any news.
I take advantage of this chart posted in another thread, for a friendly reminder that, in improving your service, you may also want to consider a better way to calculate drawdowns.
In that chart, the trader is losing almost 10% before recovering, but Max drawdown on his page shows -5.99%.
I’m positive that many investors would be grateful to know the real drawdown, without digging deeper into parameters like La.