Stop Loss Orders

Hi

I would like to see a mechanism where stops can be placed (optionally) to secure traders against significant losses. I work during the day and am not able to be on the site to react to market announcements like OCR cuts so would appreciate the chance to put an order in and know that should I get it wrong I wont loose it all.

While this may accelerate falling prices it could also create some interesting dynamics like short squeezes.

Thoughts?

In addition it might also reduce the margin that needs to be kept on orders if a stop was in place, this could improve the attractiveness of long term stocks

Having the same problem of Patch in not being able to be online during the day I too, would appreciate being able to place a stop loss order sometimes.

Yep - I'd love to see a stop-loss function added. I doubt that it will, however, because it will be iPredict and it's magic market-maker wearing the losses. I don't know if they could afford that.

From a trader's point of view, it would be ideal. I've lost money on dramatic price shifts, and was either away from the computer (most days) or unable to sell quickly enough -- try dumping 1000 shares 25 at a time, and you'll know what I mean.

A more liquid market might help (like 50 share bundles, instead of the current 25 that the market-maker offers), giving traders more fluidity to buy in and sell out. But who knows, iPredict might not be able to afford it.

The only way to get a stop-loss mechanism is for the traders to demand it. I'm not sure if that would, or should, happen.

I'm a little unclear on precisely what a stop-loss feature is, and how it would work?

With a stop lossif you were long 200 on a stock at 60 cents, you could, say if in the event that the stock dropped to 40 cents, have an order set to execute a "sell" of your 200 for the current market rates. Rather than logging on the next day to find that the stock bottomed out at 10 cents.



Edited 1 time(s). Last edit at 05/21/2009 02:45PM by jmvm.

I'll talk to the guys about it

Ah, I see. That would be very attractive to traders. But I can't really see how it's fair to iPredict (assuming that it will be buying your stock off you)!

Market maker losses are bounded, so that's not an issue.

There is one caveat: if you let it be known you've got a big stop-loss order, a manipulator might pull the price down by shorting hardout, and then cover by buying your stop-loss order. If the right circumstances exist, this is certainly possible.



Edited 1 time(s). Last edit at 05/21/2009 07:06PM by bkd.

The more I think about this, the more substantial the manipulation issue I mentioned above seems.

If stop-loss functionality is included, it's imperative that traders cannot see where the stop-loss orders kick in, or how large they are.

Otherwise, the following algorithm will harm every human trader it targets, and not improve the accuracy of predictions at all.

Input:
Current stock price p
A stop-loss order for S stocks at some price s<p

Algorithm:
Compute X:= sum( bid_quantity ) over all human bids with s<bid_price<p
If X<S, then
{ short stock until price falls below s;
cover stock by buying up the stop-loss order, and buying from the market maker as necessary after that }
Stop

This manipulation is certain to make money, provided it is executed quickly enough (not an issue if someone wants to automate it). Whoever places the stop-loss order loses money. It's not clear how allowing this will make the predictions better.

I'm not bashing stop-loss orders here—my balance would be more than its current 40c if they were available—I'm just pointing out that they must not be published.



Edited 1 time(s). Last edit at 05/22/2009 01:42AM by bkd.

New stock idea: BKD.HIRED.BY.IPREDICT.2009.

Heh. I'm actually looking for a job... but this stock would have to be coupled with IPREDICT.SHUTDOWN.2009; if hired, I'm bound to screw something up so badly that public outcry shuts the site down.

I don't often get to trade during the day (my work have blocked the site!) so I'd appreciate a stop-loss facility (as long as it couldn't be manipulated). By the way, there's an impressive number of traders online for this time of day - whether it's because of the new features or the new stocks like the mayoralty ones I don't know.

I had not considered the issue raised by bkd but it is a good one, a large stop will be a bulls eye for anyone wanting to beat other traders and get in some great out of price buys.

But the orders could be in the system but hidden so traders could not see them and iPredict will execute them at market rates for the traders. The market maker will not wear the loss any differently to normal trading.

I suggest stop losses be made available on a few stocks only as a test case to see how often they are used and how various traders find them. Maybe on a stock more frequntly traded by those with makret participation skills like the OCR stocks. Then they could be expanded to the whole site if they were popular enough.

Interesting discussion.

Stop loss would have to work by placing orders if and only if the market price went below some level - otherwise a sell order below the current market price would be matched and filled immediately. So it wouldn't be possible to see others' stop loss orders - only after the fact (and even then it probably wouldn't be obvious).

Stop loss orders won't introduce any extra loss for the market maker - whether the price is driven up or down by stop loss or regular trading doesn't matter to the market maker - and if stop loss drives the price inefficiently low, presumably other traders will step in and profitably correct it - leaving the market maker's position unaffected by stop loss anyway.

I could see it all causing issues even if stop loss isn't visible. Take the following for example:

Lets say I decided I was in a stock with 200 at .6000 cents and wanted to exit and maybe take profits.... I sell these at market rate for whatever reason.

Trader1 has a stop loss for his 500 shares at 55 cents, and because I've taken profits and the price is now at .5430 he gets an auto execute of his stop loss and the price tumbles to .4300 cents.

Trader 2 has a stop loss for 45 cents .... and so and and so on....

Now I might not be a manipulator, I might be just trying to exit a trade but if stop loss is really popular for a stock and they were in place all over a stock, I could theoretically look at this stock 30 seconds later to see it trading at 10cents now, just because the stop loss mechanisim executed for so many people.

Also If two people had a stop loss for 55cents and theres a third short seller pushing the stock down, who gets preference for selling first if it's done at market rates?

Sounds like a bit of a headache and I'd probably not use it myself for fear of the above.... but I'll keep an eye out for good 10 cent deals



Edited 1 time(s). Last edit at 05/25/2009 10:09AM by jmvm.

jmvm,

A good point, but I wouldn't worry about it too much.

Once traders will get used to it, you'll probably start seeing a lot of big buy orders down the limit book some way. Then, if one of these cascades happens, they get bargain stocks. Their buy orders will partially arrest the cascades.

I agree, stop loss squeezes are just another market dynamic that traders can get in on if they place orders just below where they think others would place their stops

Patch Wrote:
-------------------------------------------------------
> I agree, stop loss squeezes are just another
> market dynamic that traders can get in on if they
> place orders just below where they think others
> would place their stops

Not exactly an area of research based prediction.

Oracle,

Au contraire. Putting in a big buy order well below current price is an awesome research-based thing to do: you're confident something's gonna happen, but you don't think there's enough money in it, so you hold out for a bargain.

It's always confused me looking through the limit book on the API that we don't see much more of this, except presumably for people looking to cover very large short positions.

# # #

Basically, I'm completely in favour of stop-loss functionality, subject to my one concern above: that other traders don't get to know the trigger point or the quantity. Admin seems to be interested.



Edited 1 time(s). Last edit at 05/25/2009 02:58PM by bkd.

I thought they were an integral part of any happy stock market. Surprised they are already not on the future feature list.

bkd

A part of the entry from Patch is "place orders just below where they think others would place their stops " then from yourself "you're confident something's gonna happen, but you don't think there's enough money in it, so you hold out for a bargain. " and this "that other traders don't get to know the trigger point or the quantity.".

Thinking, to me, implies possibly belief but not definite knowledge and combined with your proviso about not being able to "know the trigger point" strikes me as leaving one in the position of taking a course of action based on gut instinct rather than research.

Now, a proviso of my own. "looking through the limit book on the API" may indeed, enable an aspect of research that I'm not familiar with and if it does, I'll have to concede your point.

I'll look forward to your response and will appreciate learning how we go about "looking through the limit book on the API".

As to Stop Loss Orders, I do not know enough about how they operate to have finalized my opinion and have been following this thread with considerable interest. The concept does have appeal as a means of exercising some control as I, like a lot of others, am not usually able to watch what is happening just whenever I wish. At the moment I remain neutral.

Hi Oracle, good points.

The API is more or less another set of pages provided by iPredict designed to let third-party programs communicate with the market. Much more information can be found here; but to get an idea of what it's about, try this (you may need to right-click and select "view page source" in your browser) — it shows you all the active orders on a stock, not just the top ten.

# # #

Many apologies if I sounded snippy earlier: I'm a terrible comunicator.

Also: wow. I made a typo: I meant Hi Oracle, and I typed High Oracle. Yikes.



Edited 2 time(s). Last edit at 05/25/2009 05:36PM by bkd.

Hi bkd

First I didn't take your comments as being "snippy" so have no fear on that score.

As to the typo, you were quite right to correct it due to the fact that my ROI precludes me from promotion. Well, for the moment at least.

Hi admin,

Is the implementation of a "trigger" (or stop loss etc) for trades by iPredict likely?

Or from your current position is it undesirable?

Its desirable from a number of persepctives but we haven't been able to prioritise it, I'm afraid. I can't give a timeframe on this development but its certainly in the frame. I would certainly like to see this feature added.

That is all good. Thank you.

People could use the advanced trading interface to place a sell/buy order above/below their target to use as a stop-loss?

A % trailing stop would also be useful for stocks prone to large movements.

Tradrs can do that if the sell is above the current price or the buy is below the current price, but not the other way around. A sell order brlow the current highest offer to buy will cause a match to be made at that highest current buy price and trade to occur.