You can start with the persistence model and use the previous day midpoint or perhaps the close price as your benchmark. That would be the benchmark method your formula needs to beat and if you can't beat it suggest the market is a random walk and/or your methods have no additional predictive value. You can use a metric such sum of squared distance between your actual value and your predicted value to see which method is best.

T A S software can project a dynamic profile for the next time frame if things fall in to the algos eye sight. value high low and poc. you might be able to take the rth open and over lay yesterdays profile. just set them side by side drawing a level line through the opens... it is nuts i know but you have to start some where..

The following user says Thank You to forgiven for this post:

I was suggesting 2 different methods, either using the midpoint of the prior day or the close of the prior day. Theoretically, the last price should have the most information, i.e. so that's going to be the best bet of your midpoint. However, I don't know if some information from the volatility might be missing from that that the midpoint from the prior day might capture. It is a great question though! Right, you can test it by simply shifting your data and taking the difference from the actual midpoint and summing the value. The formula I gave was the mean squared error but you could use the absolute difference or another error measure if you desired.

I coded up a quick verification in Easylanguage and using the close of the prior day is significantly better than using the midpoint for predicting the midpoint, assuming I didn't make any mistakes (big assumption). The code is below and I've linked to an instructive video if you want to learn more.

The average error in points is around 11 points for using the close over the past couple years I tested. This would mean on average using the close, the midpoint will occur within +- appx 6 points. That's your accuracy level. Using the prior midpoint, I get some ridiculous number which means I probably made a mistake. I updated to calculate using the simple MAD (mean absolute deviation) and get 8 points for the close and 14 for the prior midpoint. That would put you on average +-4 off using the prior close. PS: If someone wants to explain the differences in mean squared error vs absolute average error -- it would be illustrative for me!

Update: Found the mistake. You must wait until CurrentBar=2 because there isn't a previous value for the first bar. It did not change the findings though.

remember i said this is nuts so do not lola me to bad...you have yesterdays RTH profile on the ES ...part 1... this morning the ES RTH opens and the chart is blank just a * for the open ,,,that is part two. you move yesterdays profile in place where today RTH OPEN an yesterdays RTH close are on the same price level ...part 3 ...now yesterdays profile is your projection profile for to day... look for the poc ...that is the project poc for to day. like i said it sounds nuts but you have to start some where...what you are saying is day 1 will look like day 2 just at different levels.

The following user says Thank You to forgiven for this post:

Tested the floor pivots (h+l+c)/3 and using the prior close still had a lower error. I'd expect the open to have an even lower error but can't test it easily. You could try to use machine learning to do better but the tradeoff is the introduction of model complexity. But, there is hope, I have came up with a model that beats the simple model nearly every single day for the past 10 years or so but only by a tiny amount.

No laughing - it's much aprreciated!
(If you saw the formulas I have THEN you'd laugh!
As the saying goes:
"Crazy- yah crazy like a fox!" LOL

So one shifts yesterday's profile with the o/n gap - that is if Close to Open is minus 3 then slide it down 3 pts, ie a -3 pt adjustment.
So today's est POC is yesterdays adjusted POC, correct?

--- Part 2---
I haven't use Market profile (MP) much. I did start at one time, but concluded that:
A: I would need to read the book to use it properly
B: if everyone was using it then it, then it was so widely known that it probably had little value for me.
(About point B I'm not saying this is true "little value" - I'm just relating where I am coming from in the question below.)

Question
1. Is the POC usually equal the midpoint? , that is if someone told you what today's POC would be before the start of the day, would that level be equal today's RTH midpt?

I read the first paragraph and asked myself: "how am I - a non-programmer - going to test this?" and then reading on found you went off and did it. Thank-you. Your professionalism is noted and appreciated!

Summarizing your findings:
1. yest close is within 6 pts of today's
2. yest midpt is within 4 pts of today's

No laughing - it's much aprreciated!
(If you saw the formulas I have THEN you'd laugh!
As the saying goes:
"Crazy- yah crazy like a fox!" LOL

So one shifts yesterday's profile with the o/n gap - that is if Close to Open is minus 3 then slide it down 3 pts, ie a -3 pt adjustment.
So today's est POC is yesterdays adjusted POC, correct?

--- Part 2---
I haven't use Market profile (MP) much. I did start at one time, but concluded that:
A: I would need to read the book to use it properly
B: if everyone was using it then it, then it was so widely known that it probably had little value for me.
(About point B I'm not saying this is true "little value" - I'm just relating where I am coming from in the questions below.)

Questions
1. Is the POC usually equal the midpoint? , that is if someone told you what today's POC would be before the start of the day, would that level be equal today's RTH midpt?

step 1 take take yesterdays RTH run the fib extentions in both direction to the 127% of the RTH both ways. step 2 run the fib ruler from yesterdays RTH high low to the 127 extentions both ways.. mark of the 50% retracemente of both swings . then set yesterdays range on top or bottom of the 50% retracements. now you have projected range and path of price both ways... half of the range would the mid point. you are saying you will have a brake out of range to the 127 ex then a 50 % pull back then another swing in the direction of the trend ... that is the way markets tend to move...stair stepping,, it would just be a place to start,,,

The following user says Thank You to forgiven for this post:

Based on these I made the summary for the week just finished (ES RTH)

Note POC isn't the true MP POC of the prior day- which I don't know.
ongap= o/n gap and is the amount used where is says "AdJ .."
Abs= absolute of Diff

so as an example: the pivot point for Monday based on Friday (using the classic PP calculation on Friday's RTH numbers) is: 2290.50
and its difference to Monday's midpt of 2259.75 is 30.75

The Adj PP is 2264 and its difference to Monday's midpt of 2259.75 is 4.25

So the total of the abs differences of the Adj PP is 23.50 which divided by 5 would be 4.70

step 1 take take yesterdays RTH run the fib extentions in both direction to the 127% of the RTH both ways. step 2 run the fib ruler from yesterdays RTH high low to the 127 extentions both ways.. mark of the 50% retracemente of both swings . then set yesterdays range on top or bottom of the 50% retracements. now you have projected range and path of price both ways... half of the range would the mid point. you are saying you will have a brake out of range to the 127 ex then a 50 % pull back then another swing in the direction of the trend ... that is the way markets tend to move...stair stepping,, it would just be a place to start,,,

you have time based an volume based profiles the larger the range the less center the pocs are for the most part ,,the smaller the range the more centered the pocs are. i would google range projection ...it is about 16 points on the es i think. you would have to project the range and direction to get the mid point ..you will have your work cut out for you...you might get some insight from JS services they put out a market map on bloomberg and for banks...it cost 1000 a month but they have it now on ninja and bookmap ,,they have a free trail two weeks...the way they are mapping may help you in your research...hope it helps

The following user says Thank You to forgiven for this post:

Try Open+.2. It should improve it ever so slightly. Be sure you can view all the decimals. This one outperforms over most of the entire history I've looked at, most.

aquarian1

From the limited cals in the prior post, the Open as an estimate of the midpt is the winner at 3.25.

(Naturally, this isn't good enough for the perfectionist in me so ideas to improve this anyone??)

For quite some time I worked with PP and different versions of them.
I was looking at matches to H and L

Some days one would match some-days others but not enough to make it worthwhile

(Perhaps your floor pivots are different from regular pivot points? - they seemed to be the ones I calculate - which I called "classic" as opposed to Woody, Camila, etc...)

..........
peace, love and joy to you
.........

The following user says Thank You to aquarian1 for this post:

Trading: The one I'm creating in the present....Index Futures mini/micro, ZF

Posts: 2,285 since Nov 2011

Thanks: 7,247 given,
4,436
received

aquarian1

Thanks Ron,

For quite some time I worked with PP and different versions of them.
I was looking at matches to H and L

Some days one would match some-days others but not enough to make it worthwhile

(Perhaps your floor pivots are different from regular pivot points? - they seemed to be the ones I calculate - which I called "classic" as opposed to Woody, Camila, etc...)

Try Open+.2. It should improve it ever so slightly. Be sure you can view all the decimals. This one outperforms over most of the entire history I've looked at, most.

Do you adding a constant of 0.2 will improve it?

Is this constant the same amount as the average daily trend rise over the period you examined?

BTW the benchmark to exceed is 2.90 (current error of my default formula). This isn't good enough and I want to get it to 1.5.

I don't want to muddy the thread, but I do appreciate the ideas from all of you so here goes:

More important than an low average error is to know, if possible when the estimate of the midpoint is about correct and when it is way off.

lets say we have information that influences our formula(s) accuracy so under certain conditions we know are formula(s) are going to be way off. Then under those conditions we do not use our formula

I'm make up some stuff here:
w = a very large o/n gap (for ES >5pts)
x = an open at the low end of the estimated range
y = an open at the high end of the estimated range
z = a strong multiday trend (either up or down)

In any case, I guess I am rambling here, but I have notice that my estimated mid can be close many days and then way off on one.

..........
peace, love and joy to you
.........

The following user says Thank You to aquarian1 for this post:

Today is a good example of the need for several formula for predicting the midpt
It was clear from the o/n it would be a strong down day (at least from the open)

We had started a downtrend on Friday and once the open happen the selling was unleashed.

In this case estimates based on a continuation of Friday, or - the open would give a poor estimate.

Are you familiar with investorIQ ? He does some forward projection based on back testing. He usually puts out a YouTube video everyday. Might give you some ideas.

Today is a good example of the need for several formula for predicting the midpt
It was clear from the o/n it would be a strong down day (at least from the open)

We had started a downtrend on Friday and once the open happen the selling was unleashed.

In this case estimates based on a continuation of Friday, or - the open would give a poor estimate.

So I am thinking of various formula

I am completing my weekend updating and here is an example:
(Abs Diff means to the actual midpt)