Try My EA! Need Group Feedback...

ooh... so that is part of the v2.11 feature now?

I suppose I don't have to do anything with the inputs when applying v2.11?
Except for grid size (well, this is based on preference, whether more spicy or less spicy) and my broker's average spread, am I right?

Check the post on page 37, post 727. The way the inputs work are detailed there. The version 2.11 is the one that corrects a couple of small issues that came up.

I feel stupid about it, but I really don't know how to link to another post using this editor I only recently started really doing anything in the forums.

No prob, been there before. Top right side for each post there's "# Post 822" thing after "Quote" and "Cleanup", right click, copy shortcut.
Will check for the explanation. Thanks for the hard work!

I do feel fun in this thread though only replied once in a while but I read it most, because of the strategy, the kindness of the people contributing... keep it up! Will try v2.11 next week (after I got memory slot for my VPN...hahaha... out of memory already by running 10+ of MT4).

http://www.kitco.com/images/live/nyg...07787682455684

~dgodfrey...I think your idea has great potential, we just need to put it to test on demo to see how it goes. I also think, if you want to use historical prices\charts to test, I would choose the last FOMC meeting news, because that's the one that hedged out everybody. If you can make that one work, I think the EA would become unstoppable.

Looking at it, I think it would go one of two ways.

Looks pretty good.

It's probably a good idea, once you implement this, to test with various buffers, like you have in your calculations. That way, we can see which settings net the biggest profit over a significant period of time (like 3 months). I think the buffer will be the setting to change, not the gridsteps. I think 0.50 on that is the reasonable compromise.

I like how it allows some profiting, while keeping potential losses floating until they can be ultimately overcome by a straight move.

on another note, FF keeps taking my explorer off my signature!! ARRRRG!!! I put it back on, now.

maybe need double confirmation swinging or following trend

Yeah, I think all those trades up close to the $10 line are just hedge bait, they get taken in a lot in volatility. So much so that I think they probably lose more over time than the potential gain in a no hedge move of the trades snugged up on the profit side.

Does this EA run off grid? If not what indicators is it based on?

It's under development, one of the things we're working on now is using S/R lines as deflection points for grid origin. Details are in the thread.

@ Pipster

Kosmo is having pretty good luck using the $10 grid on a low target and doing catch and reset during volatility - hitting the target then resetting on the same $10 line.

Thanks for the quick response.

good! she needs to post it on an explorer.....(hint, kosmo, for you :nerd:blush:

OOPS! I see she already did...very nice!



~kosmo...did you do something to keep the BET version from placing new orders once price had moved up to 1220, and cycled?

A few more thoughts about the $10 grid, and some possible answers to nasty situations.

Remember from the chart on this post the very first example was of a heavy hedge. Actually it was the heaviest of the them, an example of maximum DD if it was left to run. As much as -27% in hedgeloss with GridStep = 30 and 0.1 lots by the time a reset was done.


But Kosmo has been taking some nice 1% bites and resetting the grid in volatility lately, which got me thinking of a solution.

One thing I didn't consider was that at the moment the momentum died in that surge upward, there was a very nice profit available. As much as 10% on GridStep=30.
Also from this post we found that there was a multiplying loss effect of 2.5:1 upon market reversal regardless of step size, and as it turns out on the $10 grid the $5 dividing line approximates the no profit on reversal boundary line nicely. (it's not exactly that for all gridsteps because of where they happen to land and the size of the buffer, but close).

Added NOTE: Reading back over this and looking at the charts, I may have created a sense that as long as the volatility hedge trades created by movement around the $10 line occur at prices less than $5 away, the overall P/L at the next reset will be net positve. That's not true. There certainly is no chance of a net profitable position at the next $10 line if the contrary volatile movement exceeds $5, but the tipping point at which the cycle is net negative upon completion is really a little less than $4 away from the $10 line, regardless of the GridStep setting. Still, you have to have some room to let the grid do it's stuff. As long as the price stays under $5 away on the opposing trade side, the hedge pairs generated are very manageable. And with the deciding price set at $5 away the market has a chance at least to run without immediately triggering a BE event in volatility.

So here's a TP/BE protective idea: Consider that if the price exceeds $5 away on either side in volatility, you're not going to make any money on the opposite $10 reset line if the market goes that way. As a protective measure then, why not just get out should the price exceed the $5 line (or so, depending on step size) and then reverse to the BE area? You'll be ready for the next time the grid resets and you won't pick up any more hedge pairs than you already have.

The profits, BE's and stepsizes are compared here. These numbers assume a loss of $0.50 due to spread.


As you can see with this protective view, the 0.3 stepsize makes a compelling case - over 2% profit at the $5 line and the most wiggle room to BE of any of them. On the downside though, should the price not reach $5 and reverse, there will be larger hedge losses to cover at the next reset using 0.3 as compared to others. See here.

So that's the idea, be proactive about near miss reversals, have the ability to allow BE protection. Allow it to hit the next $10 band if it will or take the money off the table at some nominal target beyond the $5 line, but don't get trapped in Max DD. Here's the view of the amended chart above using this idea.

There will be times, probably quite a few, when getting out will be unnecessary and leave you gritting your teeth on what might have been, but that's just trading. And check out the bonus that happened in this case! A nice reset for a clean run with no hedges for max profit.



And one last note: Again looking at the chart on this post, you can see that 3 out of 4 of those heavy hedges would have been setups for clean runs instead, just like here.

dgodfrey,

Thank you for the wealth of knowledge and ideas included in your recent posts. What an interesting read! Especially the latest development, the protective BE level, is something I would be very thrilled to test. On paper it looks powerful and promising.

Yes. My thinking was very simple here, and as it turns out, quite in line with your observation that the $5 level is the make or break point for our $10 runs - set the target in such a way that it will be reached before price gets to the $5 level. Having this in mind, the question we should be asking ourselves is: what the target would have to be (within the $5 band), so that it can be reached easily either in a straight line (clean, no hedge scenario) OR by the opposite $10 level, in case it does pick up some hedges on its way to the original target, and then reverses. Following your 2.5:1 hedge ratio, it'd seem that we should be aiming at a target within $2.50 from any $10 level. The rest is figuring out grid spacing, risk, and buffer so that our effort is both - feasible and worthwhile.

NOTE: Just realised that my thinking doesn't cover an expanding triangle scenario, lol. Imagine one step up, two down, three up etc. What's the solution here? Killing the grid at some loss until the next $10 level is reached?

Not sure what you're referring to, Pipster.

I did think about that, how because of volatility you could get that scenario. That would be worst case of course, with all the buy and sellstops opened between a $20 range before it gets to the next $10 band. I tried to find one of those and really couldn't, but some came close.

Setting the GridStep then, you have to consider worst case and the kind of hole you would have to dig yourself out of. With GS=30 it's over 27%DD, GS=50 yields around 17%DD and GS=75 runs about 10%DD for the perfect storm. Then you will reset and start to dig out.

Against that you have to consider the perfect runs there are, and what those mean to the account gain. With GS=30 it's around 10%, against that risk of 27% DD (notice how that 2.5:1 ratio runs all through this?), and so forth. The more you risk, the more you stand to gain - what a shock! Now if we could just know the number of perfect runs compared to the perfect storms, lol.


Unfortunately the numbers just don't really add up for a target this close with a reasonable buffer, mainly due to slippage. 1% on GS can be as close as the $3-4 range, but that turns out to be right where the tilting point is, the price that if the market should turn and head the other way it's pretty much net zero gain at the next $10 level when closing out everything. Fortunately if you don't pick up any more hedging than $2.50 volatility gets you you can still make money on the reset. I guess with a 1% target there might be a sweet spot - you either get your 1% or breakeven at the next reset (unless it's the line you just came from, what you've been doing lately).

I guess that's market justice being meted out - there's no safe place in the grid. No guts no glory lol.

I'm almost through with the code for the $10 resets and patching up the hedgeloss hole with the gains. That was easier than I thought. I'm in the middle of a rewrite, the whole program crashed and won't initialize now, got me scratching my head. After the rewrite I'll put in the BE trigger and we should be good to go.

Some changes - no mention of NoOfSteps - there won't be more than fit to $9 away, controlled by GridStep and Buffer. Less trades to close and open ought to make it faster when cycling. No MaxLoss function, since you don't really have the runaway loss situation and you can stop it whenever you like. No slippage function - that's just an open invitation to orphaned trades and corrupted grids. No BE or TP lines, those just add to the overhead and are kind of irrelevant to what this version does. And Buffer will be absolute - Buffer=100 will start the grid $1 away from the $10 line. Probably some other stuff. I've tried to clean it up and make it more efficient.

Hoping I can get it done and released this week.

What would the worst case DD be and the perfect run gain for GS $1 (effectively net 0.9 after spread)/$1 buffer/0.1 lot? Considering gold's volatility, expanding the GS still feels like the right way to go to me.

What if it'd be 1% gain per step? Am I right that two steps would give 3% well before it reaches step three? Wouldn't that be worth a try? Here I go again with my 'wider GS, higher risk, tight target' theory, lol. Or maybe it should be the other way round, i.e. tight GS and higher risk; then the target could be even tighter (in $ distance). I can't stop thinking that we've been somehow overlooking the risk per step factor vs. target in our calculations so far.

I like where you're going with the changes for the new version. Are you going to keep Sensitivity though, or is Buffer going to take over its function?

You have it right with the second choice, the trade density implied by GS=30. But that's assuming that we hold the lotsize constant. it's really more about how many lots you trade. The thing you really have to look at is the average entry price of all the trades in the series, and the quantity of lots. If I trade at GS=90 and lotsize=0.3, or GS=30 and lotsize=0.1, I'll be in nearly the same place after a few dollars movement, because both the average entry price and the number of lots traded will be nearly the same. There is something to be said for the finer granularity of the smaller GS, but the main reason for the wide disparity of results in the previous posts is that I was holding the lotsize to be constant for all GS, and that puts GS=30 with a lot more money on the line in the moves.

So with the above said, you can see in a move between lines, have GS=100 will result in an average price about $5 away from the closing price (just like in all the GridSteps) . Then the question is how much exposure do you have? With only 8 trades if you hold lotsize constant, you will only be looking at a result of $400 or so on a $10k account. With GS=30 and the lotsize the same 0.1, the average price is nearly the same, but now there are 26 trades executed for almost $1200 profit with the same 0.1 lotsize. And given that the 2.5:1 rule of thumb holds true throughout, I would expect a Max DD of around $1k for the GS=100 with 0.1 lots.

All these numbers are just off the cuff, but I hope I'm being clear.

Looks like it's going to make a stab at 1230, hope I'm getting a reset there on this version.