Yesterday we explained a trading algorithim on more desks that may explain why swing traders, retail traders, long-term investors/traders are getting stopped out more on their positions. Today we will explain another part to the algorithm that eliminates one possible adjustment.
So as we discussed, say we are long at 30 with a stop at 29.84. One way to combat the algorithm that is dropping out and causing more prices to be touched, triggering your stop, is to buy the dip. When you see an algo drop out trying to get a trader to hit out at a bad price, then bid for the stock at these bad prices. Isn’t this just an opportunity for all of us to get the stock cheaper? Instead of complaining about the danger of hitting out at worse prices, shouldn’t we embrace this new opportunity to buy at much better prices?
In a perfect world this would be a nice adjustment. But the problem is you cannot get the stock at this cheaper price. Why? If you bid 29.84 after the algo has dropped out, then immediately the algo steps in front of you. If you bid higher then the algo’s response, then it bids higher than you. The algo can enter orders faster than you can blink your eye so you cannot win this game of beating them to the bid. The algo allows you to hit out at much worse prices but will beat you to getting these much better prices.
So what do we do? I would love to hear your thoughts and will be back soon with Part III to this blog series.
Mike Bellafiore
Author, One Good Trade
26 Comments on “Why I am Getting Stopped Out More? (Part II)”
nice post
HFT sabotage mission headed by Sean Connery.
It’s a dilemma… the only solution would be setting the stops lower, but then if it’s a real reversal you will close at a much worse price.
Hide your orders.
Nice if you can tell us in the beginning how to spot an algo
This is the same logic I ran through when I read your first blog post. It’s always going to be faster than me. You can keep pushing their bid higher, to keep them somewhat honest, but you might take on too much size and have to stop…then with no one to keep it honest, it can start all over again. I’ve used this tactic before, but it’s very hard to keep up in liquid stocks.
This is the same logic I ran through when I read your first blog post. It’s always going to be faster than me. You can keep pushing their bid higher, to keep them somewhat honest, but you might take on too much size and have to stop…then with no one to keep it honest, it can start all over again. I’ve used this tactic before, but it’s very hard to keep up in liquid stocks.
look what they did to rimm at 10:46 this morn
got some people
What I have noticed is these algos tend to run on more illiquid stocks. That has happened to me more than once, you try to cover, ask runs up, you run up, runs up some more, and then you get absorbed by a “hidden” level 2 order. Stick to liquid stocks and you should be fine.
Trade naked baby…
Trade instruments wich are more liquid or trade on a higher timeframe.
Cheers,
Markus
maybe its time to revert to the low liquidity strategy of fading highs and lows.
trade stocks in play where you can see order flow and where algos are not able to do anything because big guys are acting in this stock
I am long term trader and I operate on much longer timeframes than intraday traders.
In addition, I use mental stops all the time. Only way to avoid getting killed by HFTs is
by trading on higher timeframes. Nine out of 10 times HFT algo will outsmart you,
if you trade intraday on 1-min, 5-min etc. timeframes.
What about leaving a bid out there not for the purpose of getting hit, but to control how far down the algo wants to take the stock? For example, if you long ABC at 30, have a bid at .85 or .90. If the algo is going to cut in front of you, then in theory it shouldn’t touch your bid at .90, thus giving you conviction to stay in the play. If for some reason you do get it, then that’s even better. Then you would just manage your risk accordingly…
this is part of the game right now. We can use it the same way we use the volume or candle stick patterns to be more confident on our trades, just trade after the algo, not before.
that use to be a solution. then i found out when i put a hidden order on NSDQ they actually sell that information to HFT’s so they can still cut in front of me. i was always confused as to why when i entered hidden bids/offers on NSDQ that algo’s would respond to my orders. now i know!!! not sure why it is legal for them to share my hidden orders in exchange for subscription fees. doubt the SEC will ever look into this.
In that case one could work with program stops? Programm your software to release the order. It is a little bit slower than having it on the exchange, but then it is really hidden. What do you think?
You are right on the long term aspect, and the mental stops I disagree on the HFT outsmarting short term traders. Look at it from the point of view of the HFT developer: His core strengths are lack of fear and speed. Using that you would concentrate on “outpacing” naive traders not outsmarting the adaptive ones (there are significantly fewer of them). I found quite a lot of interesting suggestions here. I have a few additional suggestions of my own. First is to break up the trade in smaller packs and scale in. The second is really consider doing what Devin suggested. Use the community to spot algos and sabotage them. Consider how you and others can gain by breaking the neck of HFTs in single situations. The more traders know about them the better. Consider posting whenever you notice algo like behaviour. (I ‘ld leave that to the experienced traders though).. Last idea is use a double stop: one normal stop placed far enough to avoid catastrophe in the worst case, and a close mental one, which only triggers after the price has stayed there for a certain time (or volume) threshold.
Idea is to trade with smaller position and than add to them as you get confirmation of breaking new levels and moving stops toward.
That is nasty from NSDQ. Always someone with his hand in your pocket. I use IB Smart order routing with the hidden function and so far so good. They use Island for that purpose. I trade quite a bit after hours when it is not liquid and I don’t get the feeling anyone knows what I am up to….but who knows.
If the algo steps in front of your bid, then either you will get filled at the cheaper price or the algo will get filled and the price action will miss your order – so you don’t load up any more, but your stop isn’t hit either (which is good).
that is not really what happens. either you dont get the fill, or you have to pay a higher price. whether it is a fill for an entry or an exit… you do not want someone cutting your bid. (though that is a natural part of price action to some extent)
I agree with not wanting someone to cut your bid, that is natural.
what I meant is: in this case if you are already long and do not want your stop hit, not getting filled on this order is fine because that means the price didn’t go underneath your order. I’m saying you don’t badly want a fill for an entry or exit here, you just want to avoid getting stopped out.
that makes sense but the bids are not there for size. a few hundred shares and then a lot of clear air in the book… so the other side of this is that your stops are hit easier as well. it’s a nice combination lol.
We can see a tag that says whether an hidden nsdq order that just executed was a buy or sell(A for ask and B for bid). Are you saying they can see a hidden order when it just sitting on the book unexecuted?!? Many of us suspect they can see that.