First Off, I am part of a community that runs around 100 tests a week on multiple different ea's. Your daddys computer is probably capable of running around 5 versons of metatrader all by itself. And if you knew anything, you could do spread analysis. Oh too big of a word, rephrase, you take the absolute probable low of a setting and the absolute probable high, and then you take the spread, adjusting the testing to the closer side of your positve outlook. Again, your backtesting is useless, the only thing it is used for is to test the EA to make sure it doesnt have any bugs in it. Run it live for a week, then backtest that week, check your results. Second, by defualt setting I assume you mean adjusting it too that perfect magical number right? You sound like one of those people, well, this well help you out. Go to google and search for curve fitting. Learn it. Then you can understand how funny that smart elec comment is. Third, im by far being negative. Im simply pointing out some things that need to be said. Negative would imply something derogatory or offensive to someone. If you felt that way, then you need to do a check up from the neck up bub. My apologies, you know, what I said I wasnt going to post on this thread anymore. This is my fault. I apologize. By all means, continue the backtesting.

- secXces

It all depends on how the EA is programmed. If you always calculate it on previous bar and don't use small take profit/stop loss. The Back tester is ok.

Or it is useless. Did anyone test it with tick data?

Bluto,
Thanks for the work done on Goblin it is truly a great EA. I use it in the 5 min chart while I am monitoring the market. I find I am "trading" the EA depending on market conditions instead of trading the market. Works great.

My only question. Could someone alter the bipolar edition so the buy side buys every so many pips down with no idicator and the sell side sells every so many pips up with no indicator. If they are hedging they can both be working at the same time. Limits would be set at the same distance as the pip entrys and the most recent limit hit would close every long position and start at .1 lot entry again when triggered.
This EA would simply be shut off during news and can be set to enter every so many pips based on the current market movement.
Does that make sense?

Initially I have used separate winner/loser counters for the buy/sell sides. However, with my settings I hardly ever seem to get both active at the same time. So for global money management purposes it seems more reasonable to count global winners/losers, where only the last closed trade (collection of orders) is used for determing the next step in a progression.

Although I find it a good idea to have bipoler trade both directions at the same time, this feature seems only useful if there are wide (or no) stops and it can take quite a while to complete a cycle.

I am working on better control over the expected value (the "edge") of the Goblin strategy and I have been collecting some statistics about my winners/losers. The interesting result is that with max trades of 6, most of the winners are at trades 1-3 and trades 4-6 win rarely. This suggests that the standard lot size management of orders is suboptimal since the risk must be increased drastically for the last orders to squeeze out some more winners. The result is that losers hurt more than necessary.
An optimal lot size management weights the orders in such a way that the higher probability trades generate more profit in relation to the lower probability trades and the total lot size is reduced to lower the costs of a loser.

any ideas are good. On my stradigy the last orders are there to get you out with minimal losses.

I been thinking that orders 4-6 could open at a wider pip interval would help.

over all a pip setting less than 30 will crap out average once a month.

one of the settings I use to get longer back tests is 1.5 risk 10k accnt. pips=30 tp=15to18

Dave

I'm sorry, my appologize too. I dont know what to say, you're a senior, I'm just a junior playing with MT4. I'm too used to trade by myself following my trading plan. Thats what I thought the robot will behave, if the plan is correct, it will run lively just like what I did on my live trading. I've run through the forward test after a week and put on backtest for the same week. Cant belive I saw a big difference on that. The slippage, requotes, everything is countered the preferred settings on the backtest. My bad to say those words out. I'm sorry for the *****ie behaviour. But I will hold on to my word as well, I'll run backtest and forward test simultaneously just to fine tune the entry.

Regards,

David

My reasoning goes as follows:
- Find a setting that gives a reasonable winner/loss rate with stable profits
In my case 5 pips, tp 15, sl 10 (will be 15 after last trade), max trades 6
on 15 min EURUSD
- Use sl to make sure losers generate predefined amount of loss, reduce sl without sacrificing too many winners
- Collect statistics about winner/loser distribution to allow lot size optimization
Example: Winners trade 1: 21, 2: 28, 3: 23, 4: 15, 5: 12, 6: 6, Losers 14
- Compute (estimate) expected value per trade, which is sum of profit/loss weighted by probability and divide by risk to get expected percentage per trade (1-3%)
- Try to optimize expected value by rearranging lot sizes of individual orders
- Compute optimal capital allocation per trade using the Kelly formula, this gives you the percentage of capital to use per trade and therefore the factor to multiply the order lot sizes with
- Backtest with computed "optimal lot sizes" to confirm expected capital growth

The described procedure tries to optimize capital growth given that your Goblin settings have an "edge".

I have found the best calcualtion to maximize profit and still be safe is to set the MaxTrades to 6. Then find the average daily range of a pair and multiply it by two. Then take that number and divide it by the MaxTrades number to get the pip size spacing between trades you should have to be safe.

So if GBPUSD is an average 150 pips...... then you mulitply that by 2 to get 300.... then divide that by 6 and you get a pip spacing of 50.

So when you get the 300 pip run..... you are only at .64 lots if you started at .01. Plenty to keep trading thorugh NFP.

For the standard 15 pip spacing..... you would be at 10,450 standard lots just on that 20th trade..... EEKS!!!!

Janus is right - must reign in the leverage......

be safe

Also -

i think you should AVOID the yen pairs........ you don't want to wake up Sunday to find out that China decides to float their currency.

Why?

Take a look at the USDJPY in 1998. Asian economic crisis cause the pair to drop 2600 pips......

in three days.

Yen is the barometer for all Asian currencies including the Yuan.

Need to be around another day to trade.