Oanda Spikes

BTW: Here comes a statement of one of my live accounts. You can easily see the mistrade in EUR/USD at 25th of november (a drop of several 100 pips), it was canceled some hours later:

http://www.weeklypips.com/charts/tra...s_081207_1.png

And in this thread I complained about a mistrade, this time no spike but a wrong fill. Got my money back as well...
http://www2.oanda.com/cgi-bin/msgboa...=003585#000000

Regards - Xaron

Well. Glad to see some proof here
Meanwhile, as an off-topic question, what time frame are you trading?

Hi guys,

Do you know if this stuff happens on demo also ? I'm considering using Oanda for a live account again and I need the tight spread for my stretgies. I've not noticed stuff like this on the demo. It would really be a shame if it is so.

Also, I didn't quite get this, do I actually need to call them or write them and ASK for a refund if this happens (and what can I do to prove it if so) ? Or is it automatic ?

Thanks a lot.

@jlryan87: I trade the 5m chart up to weekly.

@Jurrasic: Spikes might happen in demo as well but as far as I know there are no rollbacks there, it's too much effort for Oanda as they have to do it manually.

Regards - Xaron

really ????

but how to determine a false spike ?

hmmmmm............ let's see if i can take advantage of this........put my SL near these danger areas when the trade doesn't seem to be working out.............

findcount my friend, why so sceptical?

Ok, a false spike is something that is obvious. If it only had happen at Oanda's place and not at other platforms they will correct their mistake. Please note that there can be always some pips deviation as there are different sources for pricing.

The other thing that might happen are wrong fills. Sometimes (very rarely) you get a fill outside the bid/ask range. That had happened once to me (I'm there since 2006 and did thousands of trades) and Oanda gave my money back as well. But for these kind of wrong fills you have to contact them as they don't monitor it automatically, I guess.

Regards - Xaron

I personally trade 1m. I was previously trading 4hr, great profits until one big loss wipe out part of my NAV. The graph patterns look similar in all time frames. The problem is no TA strategy can correctly predict the direction every time, no matter which time frame I use. It only takes one big loss to wipe out all previous profit. I notice that a sudden swing of more than 100pips is very frequent in 1m time frame, and so trading longer time frame is prone to getting whipsaw. What's your take on this, Xaron, since you are trading longer timeframe?

i'm also looking at 30min charts nowadays and find things more 'clear-cut' than 1hr charts sometimes especially recently.

Xaron, do your Oanda daily charts have a lot of gaps nowadays ? i find there's more gaps on many pairs compared to the past.........how come ?

Well there's your answer then - take them trades where TA shows good win to loss but with adequate risk to reward also. Eventualy the real win and loss is a multiplication of these two factors so one can't ignore risk to reward in favour of just good win to loss - they should be adapted TO the the system together and verified to work by sufficient testing (which for large time frames or very big/small risk to reward even on small TFs can take a long time and I believe is also something to consider).
Gaps ? That's strange.. Could it be near the weekends maybe ? The hide weekend data and times can be adjusted in preferences if that's the case.

I've only gaps due to weekends because I've disabled the sunday.

@jlryan87: I just trade smaller for longer timeframes. I always risk about 1-2% per trade. 1m is too fast for me.

Regards - Xaron

its inevitable. this is the cost of being a low end retailer in this market we have to pay huge transaction fees.

but if im forced to choose, Oanda at least they are bit more honest.

I have been hearing the risk:reward ratio for a long time, but I didn't use it in my trading (I'll explain why later). I take the ratio as SL:TP, of which a smaller value denotes a better trade. The problem is the market is so random, that I have no way of knowing how far a trend can go before hitting SL. A good risk to reward ratio, let say 0.5, probably means 50pips for SL, and 100pips for TP. Chances are you'll be getting a lot of SL's hit before you achieve your 100pips TP.

Another thing is you need at least a win:loss ratio of 1:1 to be profitable. During consolidation, you won't be getting this ratio anyway. Another problem arises, you won't know when consolidation occurs. Consolidation doesn't necessarily occur after a big move in price, and it can happen anytime.

What's your take on this, Jurassic?

jlryan87, You should realize that what you're asking me is specifics of a SYSTEM that is used. The whole TP:SL and RISK:REWARD ratios issue together is adapted TO the system - so the system dictates ussualy what they are. To have a system you should first choose the specific set of entry rules that define it - for example, suppose I want to trade channels on 1m so I will have a number of rules that I more or less can think of from experience with charts and adjust during testing like - take channels that are at least 20 pips wide, if the channel has an incline then don't take trades where the incline is towards your SL (if you enitialy place an SL outside the channel but then the inside reaches it because the channel incline is towards your SL), take trades only within the 10% range of channel weidth from the border line (in the opposite direction of course), put SL 10% of channel weidth outside the border line and if it is inclined then you can track the SL behind the border line thus gradually decreasing it, make sure to notice potential S/Rs in range of your TP and so on and so forth all the rules.... just one example. Now this partial example of a SYSTEM dictates to you the TP target more or less, in this case I'm testing the full system with a 1:2 risk to reward (TP range is thus between about 20%-40% of channel weidth). Now after I've tested anough I have the win:loss and I can estimate the average risk:reward (too much work to record the risk:reward every time, just win or loss is easy, I can guesstimate the average r:r more or less easily for this one). Now I know both factors TOGETHER determine the end result, so if I see the average win:loss is 1:1 it's still good because the r:r is about 1:2+. So again, the system itself GAVE me the reasonable sizes of SL and TP through enitial estimate and then maybe some adjustment during testing to optimize the system for best results.

Hope that explained my point and is usefull to you, good luck with your trading

Jurassic, thanks for taking your time in explaining. I really appreciate it.
I knew this stuff before, it's just that I find it hard to break the chaos barrier (more on this later). If I wanna stick to the r:r ratio, then I might be over optimising the system to achieve the expected earnings, and at the same time, violating the entry and exit rules. Let say entry when line 1 crosses line 2, and exit when line 2 crosses line 1. Assume a position is opened some time ago when line 1 crossed line 2, and TP is not yet reached but line 2 has already crossed line 1 now. By then I only have 2 options:

1) Adhere to the entry & exit rules --> Exit position
OR
2) Adhere to r:r ratio --> Hold position until either TP or SL get hit

Theoretically, backtesting will show either (1) or (2) is more profitable. But in reality both (1) and (2) can be both having a loss, it's just that one is losing less only. To be profitable, we need a mix of (1) and (2), sometimes applying (1), sometimes applying (2). The problem is we don't know when to apply which. This is what I call chaos barrier.

You might propose that we can use one more indicator to confirm the trend. But another indicator comes along with another chaos barrier too. The barrier just got more complicated with an additional indicator.

I don't know whether what I have written above makes any sense to you but anyway, good trading to you

Well I suppose that if you test it enough then you should get your answer. Sometimes it's better to take a smaller loss then a larger one when the odds seem to turn against you. If you use certain indicators though, stick to them while you test them, doesn't make sense adding others until later or just as secondary guides or else you mess up your testing - that also requires patience, like trading.

What I described above was for a rather strict system on small TF. On bigger TF you could trade as you choose with less specific rules (though stuff like MM should defenitly NOT be neglected). On a bigger TF it becomes more of a cross between TA and fundumentals. You could place SLs behind S/Rs, you could stick to good r:r or you could trade with no SL at all (very bad idea unless you trade only EUR/USD as the major pair and you're an expert in TA and fundumentals - definetly not me), it's more vague and statistics is not very usefull to judge it and people keep arguing about it on the forums. But that's just not for me, I personnaly prefer to have relative certainty by finding small TF almost purely TA dependent methods through carefull study of price action and trying to notice repeatetive patterns I can take advantage of and maybe occasionaly ride the big trend on E/U with proper mm.