A Lot of Money and Risk | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Reliable Broker with nano lot, micro lot broker, leverage 1:400, 1:500 35 replies

calculating lot size based on money to risk, not % 4 replies

Buying 1 lot GBPUSD: Is it 1 lot GBP or USD? 5 replies

Risk Managament and Lot SIze Calculations 4 replies

BUY run profit 1 lot, SELL LIMIT 2 LOT down again n BUY 1 lot - EA Reqd 5 replies

Attachments: A Lot of Money and Risk
Exit Attachments

A Lot of Money and Risk

  • #21
  • Apr 12, 2010 3:38pm Apr 12, 2010 3:38pm
I plan on using stop limit orders. BY stale, you're saying they haven't been cancelled yet, and won't be honored?

What do you mean by lookback trick?

thanks for all this info.. Really informative.

Quoting takef586
As far as trading in tight distances is concerned, you would have to be extremely smart to overcome the trading costs. When your machine sees the price change and wants to follow the market, you will often get filled at a price much worse than results from backtesting, because lots of quotes are stale. If you trade on limit orders, you will be rejected many times on the bottom or top ticks by the banks that apply the lookback trick. Be careful and generate enough trades on small positions before going ahead.
  • #22
  • Apr 12, 2010 4:08pm Apr 12, 2010 4:08pm
When you go to trade on these aggregated platforms, you discover all sorts of things: there are prices you never see and some prices you see are not there... The broker will not tell you much at first, because he wants your commisions, however ask him about the lookback right (say YOU want one...) so that he might start telling you how it works - normally there is a max time span of 200msec or so when they can reject your trade... The problem with measuring the spreads is, you do not know what is real ( ecn prices) and what is a stream (a price that needs to be "approved" later on). On my platform I often see discount prices which are obviously fake, because one bank or another has some latency problem every now and then (so keeps on showing a price which is dead). In other words, a multibank platform is a complex aggregation of prices and quasi prices, and you cannot trust each one the same way. There can also be problems with latency betwen your algo and their matching engine - for example on the platform where I trade there is no ECN part, so even if I use limit orders, they are actually market limit orders generated on my server in response to a matching price and so it is important to host your trading server on the same network as your platform provider...
  • #23
  • Edited 6:19pm Apr 12, 2010 6:00pm | Edited 6:19pm
Quoting glennl2000
I've spent probably the better part of the last year researching and building automated strats for FX. I'm definitely not a pro. ......
Glenn,

Obviously nobody can comment on the efficacy of your system, without a lot more info. As Darkstar says, simply fitting optimized TP and SL points is unlikely to provide a real edge. However, after reading your post (#12) it sounds like there's a vast amount more to it than that, and you have your back and forward testing results to give you an indication of likely return, drawdown, etc, within appropriate levels of statistical confidence. The only comments that I would add are: (1) many people underestimate the sample sizes needed to attain that confidence; and (2) if you are making small, frequent intraday trades, make absolutely sure that your testing factors in worst-case spread.

Looks like takef586 is giving you the info you need on broker pricing, spreads, fills, possible manipulation, etc, so I won't attempt to add anything.

This may be stating the obvious, but whether you trade a 5k or 500k account can only be your personal decision. Only you are in a position, i.e. knowing your financial situation, appetite for risk, understanding your likely emotional response to loss, etc, to answer those kind of questions. Of course automated systems allow greater detachment from psychological factors. Personally - if it was me - I would start small to test the water first, and then add more later as results kept pace with those attained in testing, buoying my confidence. I would also size positions significantly below what you believe to be safe, as (if I may generalize) real-life markets tend to be less predictable than any testing would indicate. Preservation of capital should be a prudent trader's topmost priority. But having said all that, I'm conservative by nature, and a 125k risk might be trivial to (how can I put this delicately? ) a millionaire whose primary motivation is "the thrill of the hunt".

One question: you say "My goal is to be successful for 1 month, and return to normal risk afterwards". I assume you somehow anticipate that your system will encounter more favorable market conditions in the first month, than thereafter? Otherwise there's no mathematical justification for doing that, it merely comes down to a personal preference.

Good luck; I wish you all the best,
David
  • #24
  • Apr 12, 2010 6:02pm Apr 12, 2010 6:02pm
Thanks a lot for taking the time to write all of this. I'm going to check into EBS costs, which may be out of my range, but I'll take a look nonetheless. I can get access to ebs, hospotfi, lava, fxall, currenex, etc. Obviously, costs may be prohibitive, but which would be your top 2 choices, other than EBS which would be 1 of course.

thanks takef!

Quoting takef586
When you go to trade on these aggregated

platforms, you discover all sorts of things: there are prices you never see and some prices you see are not there... The broker will not tell you much at first, because he wants your commisions, however ask him about the lookback right (say YOU want one...) so that he might start telling you how it works - normally there is a max time span of 200msec or so when they can reject your trade... The problem with measuring the spreads is, you do not know what is real ( ecn prices) and what is a stream (a price...
  • #25
  • Apr 12, 2010 6:24pm Apr 12, 2010 6:24pm
Quoting glennl2000


1. Trading eur/usd during periods of average to slightly below avg vol.
2. Making no more than 3 trades per day
4) Using 10-1 leverage ( 5M positions)
5) Profit Target of 5 pips per trade
6) Fixed stop of 40 pips ( initially 8% of capital).

Using 10:1 leverage on a 500k account if you open a 5M position you will receive a margin call if it even lets you open the position.

5M or 50 standard lots is 500 per pip, with a stop of 40 pips that is 20k at risk which is 4% of the account not 8%.

Of course I am assuming that this is with an account based in USD.
  • #26
  • Apr 12, 2010 6:28pm Apr 12, 2010 6:28pm
Hey Glenn,

Is price the only data input you're using for trading decisions? Just curious.
  • #27
  • Apr 12, 2010 6:55pm Apr 12, 2010 6:55pm
Magnum,
Yes my mistake, that's 4%. Why would I receive a margin call or not be allowed to open the position? Yes, it's a USD account.

Quoting magnumfreak
Using 10:1 leverage on a 500k account if you open a 5M position you will receive a margin call if it even lets you open the position.

5M or 50 standard lots is 500 per pip, with a stop of 40 pips that is 20k at risk which is 4% of the account not 8%.

Of course I am assuming that this is with an account based in USD.
  • #28
  • Apr 12, 2010 7:06pm Apr 12, 2010 7:06pm
Current price is not the only input. I use fuzzy logic to help determine weekly inputs (out of a set of 10) that should be used as inputs to my neural net. The inputs generally include some technical indicator values, prices of course, volatility, and some linear statistics. Some values are included twice for different timeframes. Once I receive the output, I'll confirm it with some more TA, and decide whether to trade, ignore the signal, or add it to a pending queue for further confirmation. My system analyzes tick data and aggregated bar data. It's entirely written in c++ if anyone cares, but I'm sure no one is interested in the technical details.

Quoting Scotty B
Hey Glenn,

Is price the only data input you're using for trading decisions? Just curious.
  • #29
  • Apr 12, 2010 7:07pm Apr 12, 2010 7:07pm
Quoting glennl2000
Magnum,
Yes my mistake, that's 4%. Why would I receive a margin call or not be allowed to open the position? Yes, it's a USD account.
because at 10:1 you need 500k to purchase 5M. That leaves zero available equity to cover price swings against you. Since there will be at least 1 pip spread plus commission this will place your equity in a negative balance.
  • #30
  • Apr 12, 2010 8:37pm Apr 12, 2010 8:37pm
ahh of course.. In that case I'll buy 5M with 20x leverage.

Quoting magnumfreak
because at 10:1 you need 500k to purchase 5M. That leaves zero available equity to cover price swings against you. Since there will be at least 1 pip spread plus commission this will place your equity in a negative balance.
  • #31
  • Apr 12, 2010 8:47pm Apr 12, 2010 8:47pm
out of interest, will you be using alpari, fxcm or oanda?
Attached Image(s) (click to enlarge)
Click to Enlarge Name: alpari.gif Size: 14 KB
Click to Enlarge Name: fxcm.gif Size: 14 KB
Click to Enlarge Name: oanda.GIF Size: 140 KB
bull in a china shop
  • #32
  • Apr 13, 2010 1:56am Apr 13, 2010 1:56am
@plasmapants
The three abovementioned names are in the category of bucket shops, i.e. equivalent to a rigged gambling ring, the interbank aggregators like Currenex are limited to institutional trading and have their own problems, but they do not force you to execute everything in one venue, so they are open to arbitrage, and thus tend to represent real market conditions

@glennl
EBS is a price discovery venue par excellence for a few major pairs like EURUSD, USDJPY, USDCHF, but you can only trade in whole pips and in one million lots. It is probably more advantageous to trade there if the bulk of your execution is on limit prices.
Out of other platforms, I would avoid the platforms of platforms ( like Lava), as they have to cope with too much latency. I would advise you to go to the platform which has the highest volume and the most up to date technology - try Currenex, but also Hot Spot. The real solution down the road could be to build your own aggregator and tap directly a few major banks (plus potentially EBS and Reuters) for liquidity - this would save you a lot of bro money, as you would only have to pay your Prime Broker, and would get tighter (potentially) prices with less latency, hence fewer execution problems. There are ready made aggregators out there, but the costs can run into 150k+ USD per year (check Apama). Having read through the part where you state your background, I would also recommend that you look around for a partner who is/used to be an interbank trader, so that you gain more insight in the market making mechanisms and you acquire relations with the bank world. For example, EBS would not even let you take their stream unless you can prove a multi year professional background in the industry.
  • #33
  • Apr 16, 2010 7:12am Apr 16, 2010 7:12am
takef586, what's the trading frequencies of you strategies? While the issues you discuss are a major concern no matter what type of strategies you trade - but to really feel the pain, I assume that you at least at some point have traded on a relatively high frequency?
  • #34
  • Apr 19, 2010 4:28am Apr 19, 2010 4:28am
My current time frame is intraday at max, and preferably I like to trade several times a day per strategy. To trade high frequency, you have to get inside the spread, which is complicated and can be expensive if you run it across various platforms. My most successful strategy traded probably around 30 times a day ,on average, but currently I am happy if I manage to do 3-10 trades a day - the market volatility is collapsing.
  • #35
  • Apr 19, 2010 8:05am Apr 19, 2010 8:05am
Quoting glennl2000
Ben,
How is your scalper determing when to trade?...
I must have missed the reminder email to view this thread.
As we can see the straight 8:1 ration didn't work. I'd like to hear your take on non fixed tp/sl

Sure if you want to drop me a mail with some stuff on, I enjoy a good msn/skype chat about methods with good potential. I'm also working on a variable tp/sl scalper, so far so good.
http://www.myfxbook.com/members/BenS...1dxw9lbYkoEVT5
  • #36
  • Apr 21, 2010 7:44pm Apr 21, 2010 7:44pm
As takef points out below, I won't be using any of those. I'm in the process of co-locating at equinix. I'll be doing quite a bit of testing. As I've mentioned before, I've written my trading system in c++, however I have embedded assembly code as well for performance reasons. I'm going to want to put it all through the wringer for a bit before firing all cylinders. I hope it won't all be for nothing ( or worse than nothing!).

Quoting plasmapants
out of interest, will you be using alpari, fxcm or oanda?
  • #37
  • Last Post: Apr 21, 2010 8:05pm Apr 21, 2010 8:05pm
Dang I wish I had da bwains to understood any of dis.
"If The Fool persists in his Folly he will become wise." - William Blake
0 traders viewing now