Trading Yesterday's High or Low ONLY

[eminiman414]

This is more of a question specifically to @NoDoji but I can't PM her so I figured I'd just start a thread and see what happens.

Essentially, I am looking to do a "study" of sorts around yesterday's high/low ONLY (for now). I am looking to trade only around those two points for the time being. I have my reasons but we don't need to get into those. My issue is I am struggling to figure out exactly how I want to go about it.

I am interested in the excursions around these two levels. I am interested in more of a "flow based approach" around these levels (sorta like an SLA type of deal....uh oh), which is in a way how I'm trading now. I don't necessarily want to pick a pattern let's say for example, a double top/bottom around these levels. I know trading these two points only will already reduce the number of trades which I like and filtering it further with a pattern I'm assuming there wouldn't be enough to go around, again an assumption. I guess you can say well just test the SLA or whatever flow based approach around those areas, kind of a duh, but I am also interested in let's say....on the first test to the tick of the PDH/L what typically happens? Do we hold and drop off the PDH? If so by how much? If we then cross the PDH how far do we trade above it? Do we cross back down essentially now chopping around the PDH? I don't necessarily want to call it MAE/MFE because you'd have to be in a position to have favorable or adverse movement. I guess I'd be interested in a more general study in which case I can then better apply whatever "tape reading"/"flow based approach" I've currently been working with.

I hate backtesting/doing these statistical type things. I either 1. mess it up somehow or 2. half way thru I go "damn it why didn't I think to incorporate tracking this or that, or I get halfway thru and go well that was dumb tracking all of that "stuff" for really no reason. Maybe I'm trying to do too much with my former tests of different ideas? For example let's take just the PDH/L. On a reversal from the PDH how often do we touch today's low, yesterday's 50% area, yesterday's low? Then I'd get thru some examples and say shit should I have tracked a trailing stop of some kind (behind swings or 50% from the turn)?

I can go on but hopefully you got the gist of it. Mainly I want to wake up every day and for the most part know exactly what I am going to do and where I am going to do it. My trading is currently pretty much 100% discretionary. Do I have rules? Sort of but not really. I am pretty open to any type of trade based on what's happening in front of me and that behavior is taking place. I've been live on and off for the last 3 months and I've definitely traded breakouts, reversals, and retracements. I've traded shakeouts and BOPBs. Essentially I "follow the waves" until something jumps out at me and I roll with it. Seems crazy and it's totally a very loose strategy but I'm up 6k in those 3 months (not live everyday either) and I am trying not to use the PnL as an indication of skill level. With that said though, since I haven't done serious statistical analysis/backtesting I've put in hours upon hours upon hours of screentime via replay and real time sessions (both live and sim). I've done between 20-30 replay sessions per week since January of this year. Of course some days I missed due to illness or vacation or just tired eyes lol, but for the most part I've hit btw 25-35 sessions most weeks. This is trading only the first 90 minutes and at times speeding the replay up so it's not terribly bad going thru all of that. I've been thru all of 2014 twice via replay. I'm only saying this because I've done work, it just hasn't been very statistical other than what my platform keeps for me. I kind of have a fuck it attitude. I know this double bottom my fail I know it may work. Do I need a statistic to tell me what I already know? I'd rather see these things happen in real time then get a probability of it's success rate. The issue with what I am doing now is I'm stressed everyday I guess in a way without having that stat or that set in stone way to trade. While I am pleased with being profitable, I'm feeling my quality of life is being affected almost by not having a set in stone plan. This week I took a step back and just watched price move around PDHs/Ls and already I felt a much more calming feel with even just that set component. In a way that's just in too. I don't want to feel anything and I know I shouldn't. This game needs to be business as usually and everyday I just don't know what kind of business I'm going to get myself into and that's what I am trying to eliminate. Any other profession I've had I never used money as how successful I am so seeing that positive PnL is almost making me feel lucky lol. Not once have I come home from work on a friday and said oh great I made $1000 this week. It was more this client did this or that blah blah.

One last thing I'd like to say is about risk. I'm starting to feel this is all very user defined. What should/could one really risk on any given trade. Sure you can use the danger point, or another price level or a line break, but what's the risk? A tick below? A point? Two points? I'd like to make it about the market and not about me but what's the true risk the market is showing me? It's all very interesting to me and I'm just venting and overthinking now but just wanted to let it out hahaha.

So yea, making a short story longer.....a study around PDH/L what's a good way to go about it lol??????

 

[k p]

This is my kind of trading... at some sort of extreme. Here is the thing that I've noticed though... first even defining a PDH or PDL is tricky. First you have to figure out what exactly is the high. I have my charts set up so that the open happens at 6pm ET. Now is the high, a high that is reached during the overnight session, or is the high that is reached during the 0930 to 1600 session? I see that often times, tracking both of these points can produce reversal trades where price hits this level within 4 ticks but can't go higher, but sometimes is the high reached during the previous RTH session, and sometimes its the ultimate high, even if that high occurred at some late hour.

Now if you take your stats of 6k in 3 months, that means 500 per week, which would be 25 NQ points, so roughly 5 points per day. Obviously since you don't trade every day this number probably varies widely, but its a good general number. I think with these very good stats, you almost don't need anyone else to tell you what to do.

Honestly, I don't think anyone else here is trading like you, so you may not find the answers you seek. Imagine if you do come up with a stat that says, ok, a double bottom at the PDL has a 75% success rate of producing a 10 point profitable trade if I put the trade on with a 3 point stop. Will you take this every time, regardless of how you see price moving?

If you are watching price move and how it moves is affecting your decision, then any stats you might develop will not take this variable into account. I think that guys only trading 1 minute or higher bars can use statistics, but if you're watching price move and following a fast chart, there get to be too many variables.

Perhaps everything you're also talking about has to do with the fact that you can't believe you're profitable, that it can be relatively this easy, and you're looking for something on the horizon that you missed which might make you fall flat on your face one day. I honestly think it can be this easy once you've worked out most of the variables and there isn't much else to do but follow the bouncing ball where ever it leads.

I spent months trying to figure out what it means to follow behavior and watching price move and how this fits into the whole statistical analysis of "setups", and not one person could provide an answer that was sufficient. Why are you stressed if you're up 6k in 3 months? I think finding the answer to this would be much better than finding out if you need more statistical analysis in your method.

 

[k p]

As an illustration, I start by plotting my overnight levels and previous day levels on my 5 min chart that then transfer down to my faster charts. This way, I can scroll through weeks/months of data and just visually see where stuff happens. Yesterday was a perfect example where price first bounced of beautifully from the ONH, then the ONL, and then the PDH, even though it ultimately broke through.

But would you really trade a statistical analysis of these levels by just seeing how far penetrations happen, if it penetrates? The fact that sometimes price comes within 3 or 4 ticks but doesn't even touch the level is a problem because now you can't have a standing limit order at the level which could fill. So any statistical model with have you buying too soon, to make sure you're able to enter the trade, and then possibly taking on too much risk to have the stop beyond the danger point. And since you already know that its about how price moves as it approaches this level, I'm not sure if a statistical model can be built around how price moves, other than the intuition that comes from how price moves.

A statistical model would I think be "price gets to with x ticks of the a level, doesn't penetrate by more than y ticks, trade is entered z ticks away, etc." I just don't see this being any more valuable for your style of trading! Intuitively you know how price should behave during a rejection or penetration of the level, but how you build this into a statistical model I have no idea.

Edit: And oh ya... I'm not NoDoji, but perhaps my ramblings will help bring your thread to the front of the list so she sees it!

 

[eminiman414]

Essentially the goal is simplicity. I'm not necessarily committed to keeping my current style. The PDH/L to me aren't tricky to define at all. There's even indicator that plots it for you lol. It to me is the most objective levels to trade or trade around.

I'm looking initially just to do a study. Not so much defining the exact plan to enter and exit and all that jazz.

Oh and price must touch the level exactly. At least that's what I'm going to work w initially. Again simple and objective. Price either touches the level or it doesn't. Black and white.

It's also not about the money. Sometimes I overtrade, sometimes I under trade, sometimes I don't trade at all, and sometimes it's just right. Yes I'm up but at what costs internally.

Again it's a study. To see if it will just make my trading world "easier". There's also the target aspect. Often I see price run 50+ points trading off a pdh/l so why am I taking 5-10. Yea I've caught bigger runners than 10 but even just scrolling thru some charts quick many times the move off the pdh/l runs for at least 20. I'd rather trade less risk a little more to catch 50 points than enter 8 times in 90 minutes and net 10 points.

 

[k p]

Ya.. I hear you. Lots of guys over the years seem to say that after a while, they switch to trading a longer bar interval for less stress and more time off. This I think only comes from having developed the skill to backtest a strategy and have the balls to stick with the plan. When you got multiple contracts on the line and price drops 10 points, you might be sweating buckets, but knowing that you're going for a 50 point move anyway, this might all be with the risk parameters of the plan.

Ok.. hopefully someone else provides some better feedback.

 

[nakachalet]

This is more of a question specifically to @NoDoji but I can't PM her so I figured I'd just start a thread and see what happens.

Essentially, I am looking to do a "study" of sorts around yesterday's high/low ONLY (for now). I am looking to trade only around those two points for the time being. I have my reasons but we don't need to get into those. My issue is I am struggling to figure out exactly how I want to go about it.

I am interested in the excursions around these two levels. I am interested in more of a "flow based approach" around these levels (sorta like an SLA type of deal....uh oh), which is in a way how I'm trading now. I don't necessarily want to pick a pattern let's say for example, a double top/bottom around these levels. I know trading these two points only will already reduce the number of trades which I like and filtering it further with a pattern I'm assuming there wouldn't be enough to go around, again an assumption. I guess you can say well just test the SLA or whatever flow based approach around those areas, kind of a duh, but I am also interested in let's say....on the first test to the tick of the PDH/L what typically happens? Do we hold and drop off the PDH? If so by how much? If we then cross the PDH how far do we trade above it? Do we cross back down essentially now chopping around the PDH? I don't necessarily want to call it MAE/MFE because you'd have to be in a position to have favorable or adverse movement. I guess I'd be interested in a more general study in which case I can then better apply whatever "tape reading"/"flow based approach" I've currently been working with.

I hate backtesting/doing these statistical type things. I either 1. mess it up somehow or 2. half way thru I go "damn it why didn't I think to incorporate tracking this or that, or I get halfway thru and go well that was dumb tracking all of that "stuff" for really no reason. Maybe I'm trying to do too much with my former tests of different ideas? For example let's take just the PDH/L. On a reversal from the PDH how often do we touch today's low, yesterday's 50% area, yesterday's low? Then I'd get thru some examples and say shit should I have tracked a trailing stop of some kind (behind swings or 50% from the turn)?

I can go on but hopefully you got the gist of it. Mainly I want to wake up every day and for the most part know exactly what I am going to do and where I am going to do it. My trading is currently pretty much 100% discretionary. Do I have rules? Sort of but not really. I am pretty open to any type of trade based on what's happening in front of me and that behavior is taking place. I've been live on and off for the last 3 months and I've definitely traded breakouts, reversals, and retracements. I've traded shakeouts and BOPBs. Essentially I "follow the waves" until something jumps out at me and I roll with it. Seems crazy and it's totally a very loose strategy but I'm up 6k in those 3 months (not live everyday either) and I am trying not to use the PnL as an indication of skill level. With that said though, since I haven't done serious statistical analysis/backtesting I've put in hours upon hours upon hours of screentime via replay and real time sessions (both live and sim). I've done between 20-30 replay sessions per week since January of this year. Of course some days I missed due to illness or vacation or just tired eyes lol, but for the most part I've hit btw 25-35 sessions most weeks. This is trading only the first 90 minutes and at times speeding the replay up so it's not terribly bad going thru all of that. I've been thru all of 2014 twice via replay. I'm only saying this because I've done work, it just hasn't been very statistical other than what my platform keeps for me. I kind of have a fuck it attitude. I know this double bottom my fail I know it may work. Do I need a statistic to tell me what I already know? I'd rather see these things happen in real time then get a probability of it's success rate. The issue with what I am doing now is I'm stressed everyday I guess in a way without having that stat or that set in stone way to trade. While I am pleased with being profitable, I'm feeling my quality of life is being affected almost by not having a set in stone plan. This week I took a step back and just watched price move around PDHs/Ls and already I felt a much more calming feel with even just that set component. In a way that's just in too. I don't want to feel anything and I know I shouldn't. This game needs to be business as usually and everyday I just don't know what kind of business I'm going to get myself into and that's what I am trying to eliminate. Any other profession I've had I never used money as how successful I am so seeing that positive PnL is almost making me feel lucky lol. Not once have I come home from work on a friday and said oh great I made $1000 this week. It was more this client did this or that blah blah.

One last thing I'd like to say is about risk. I'm starting to feel this is all very user defined. What should/could one really risk on any given trade. Sure you can use the danger point, or another price level or a line break, but what's the risk? A tick below? A point? Two points? I'd like to make it about the market and not about me but what's the true risk the market is showing me? It's all very interesting to me and I'm just venting and overthinking now but just wanted to let it out hahaha.

So yea, making a short story longer.....a study around PDH/L what's a good way to go about it lol??????

someone just posted this but wasn't posted to you directly.

E N J O Y

.................................................................

 

[eminiman414]

Some quick stats. More work to be done.

Sample size in days=186
01/05/2015 to 09/22/2015

% / #
How often does the market break higher than yesterday's high? 55.38 / 103
How often does the market close higher than yesterday's high? 30.11 / 56
How often does the market break lower than yesterday's low? 47.31 / 88
How often does the market close lower than yesterday's low? 23.12 / 43
How often does the market stay within yesterday's high or low? 9.68 / 18
How often does the market break both yesterday's high or low? 12.37 / 23
How often does the market break either yesterday's high or low? 90.32 / 168

Average range across all days was 65 points. Didn't set a number as to what qualifies a break (yet) so based on excel a break would be 1tick, but 90% of the time the PDH or L was touched. Yesterday's high or low was never matched the following day with it being the current day's highest high or lowest low. I did track the extents of the breaks and some other things but still working.

 

[k p]

Hey emini... was just watching a video and I thought of you and this post. Someone suggested I look up stuff by a guy named FT71 (Futures Trader 71). I stared to watch a random video on statistics, and just fast forwarding through, and I notice that in this video, he is showing an example with exporting data into excel to try and answer the question "How often does the day session equal or exceed the overnight low or high in the ES." This seems to start 18 minutes in. Not sure if there is any good stuff in the beginning cause I'm just scanning it, but when I saw this bit of info, I thought it was very much in line with what this thread was about. Hope you find it helpful.

 

[eminiman414]

Thanks for the post. I've watch all his videos. Some good info in his vids

 

[dartmus]

Yesterday close had a decent pb up. Thus far today price hasn't challenged yesterdays lower tail. It's still naked like the kind of tail that statistically forms short term counter trend swingL's. The attempt to rally found resistance at a prominent ceiling. Confluence formed by a combination of both of yestday's sess's opens being v near each other.

But it's broken above that area and challenging the upper tail. Uppertail is not naked. It's left side is shaded by an earlier bar but bears still see this as a desirable area to short. Shorting above yesterdays open in a downtrend is one strategy fund managers use to beat avg's. Even if they have to do it a day late.

Bulls would love to capture that entire upper tail before the reg sess opens. If we open near the H bears might be reluctant to sell this H given that lower tail would then likely still be unchallenged. If yesterdays uppertail isn't shadowed by todays bar and bears manage to cover the right side of yesterdays lower tail then direction will be clear. But given blsh shbl it's logical to be forward looking and anticipating events that will lead to bo of the ends rather than waiting for bo out of the ends. Yesterdays opens and closes are imo clearly adequate in this area. Particularly when combined with something I've seen disparaged on another thread today. Bill Wiliam's alligator distilled to it's simple essence aka NoDoji's mobile resistance. Simple multi timeframe analysis of an ema.