MTF Multi-TImeFrame Repainting Pitfalls

[HODL-Lay-HE-hoo!]

I saw this code from an @a1cturner post is this a viable option (at a glance) to delay an signal from a higher TF to plot after the higher ag bar is closed?

https://usethinkscript.com/threads/chart-bubble-exception.10457/post-93159

 

[HODL-Lay-HE-hoo!]

Or maybe set hiding? Hmmm

 

[METAL]

I made a lower indicator to show clouds for the MTF Heikin Ashi indicator that @halcyonguy made. Is there a way to make it not repaint?

 

[MerryDay]

Multi-TimeFrame indicators are scripts that reference a higher timeframe besides the current timeframe of your chart.
MTF indicators repaint because they reference the current higher timeframe candle which is updating on every tick until the candle closes; at which point, it will go back and repaint all the candles with the higher aggregation price that you requested.
This means on a 1min chart, the 5min candle with continually repaint for 5 bars until it closes and does its final repaint.

If you don't want the repainting, then reference the prior candle == price[1]
The prior candle is fully formed and closed thus no repainting.

The no-win choice with MTF indicators is:

Suffer the repainting and false signals with the use of the current candle
OR
Incur the lag built-into using the previous closed candle.

Most members choose repainting; as lag will make most indicators unprofitable.
I choose to run in the other direction from MTF indicators.

 

[HODL-Lay-HE-hoo!]

Thanks, FOR CRUSHING MY DREAMS! .... joking geeez alright fine ill give up the pipe dream of making repainting indicators not repaint haha.

 

[Bingy]

Hi all,

Second time posting on here, as this forum has been immensely helpful and everyone seems cool. I'm currently closing in on developing a pretty good strategy, but as I was forward testing this past week I noticed that one of my six buy signals seemingly "repainted", but the way it did it was so strange. It took a trade, then the trade turned against me and the buy signal disappeared. It then hit my stop loss. However, the resulting losing trade didn't re-appear until another trade was taken shortly thereafter. Then the same thing occurred where that was also a small loss and as the trade turned against me, not only that one disappeared, but the one right before it also disappeared again. Super strange, especially since it's "repainting" 2 different trades. My stop loss was once again hit, but the results haven't reappeared, so now I'm wondering if I'll have to wait for the next winning trade to happen before these two losses show up again or if they'll just never reappear again?

I'm wondering if it's because I'm using higher timeframe RSI's such as a four hour RSI and a Daily RSI? If that is the case, is there any way around this?

And I'm also wondering what the common pitfalls are surrounding this phenomenon of "repainting", as I'd like to avoid it at all costs.

 

[MerryDay]

MTF Multi-TimeFrame indicators are scripts that reference a higher timeframe on the current timeframe of your chart.

MTF indicators repaint because they reference the current higher timeframe candle which is repainting on every tick until that higher timeframe bar closes.
At which point, it will go back and do a final repaint of all the candles with the higher aggregation price that you requested.

This means if you are using a 4hour aggregation calculation in your buy signal on a 15min chart;
it with continually repaint 16 bars until the 4hour candle closes and does its final repaint.

If you don't want the repainting, then reference the prior candle == price[1]
The prior candle is fully formed and closed thus no repainting.

The no-win choice with MTF indicators is:
Suffer the repainting and false signals with the use of the current candle
OR
Incur the lag built-into using the previous closed candle. When using a 4hr aggregation the lag will be the length of those 16 candles. Which will not be updated at all.

Most members choose repainting; as those kinds of lags will make most indicators unprofitable.
I choose to run in the other direction from MTF indicators.

While MTF indicators are useful in providing insight into the trend on higher timeframes; they are never meant to be used as part of the buy signal calculations.
The market was bullish over the last few months, so the effect would not have been as self-evident.
But as we head into the August/September Correction there will be more volatility with much more repainting.

 

[Bingy]

MTF Multi-TimeFrame indicators are scripts that reference a higher timeframe on the current timeframe of your chart.

MTF indicators repaint because they reference the current higher timeframe candle which is repainting on every tick until that higher timeframe bar closes.
At which point, it will go back and do a final repaint of all the candles with the higher aggregation price that you requested.

This means if you are using a 4hour aggregation calculation in your buy signal on a 15min chart;
it with continually repaint 16 bars until the 4hour candle closes and does its final repaint.

If you don't want the repainting, then reference the prior candle == price[1]
The prior candle is fully formed and closed thus no repainting.

The no-win choice with MTF indicators is:
Suffer the repainting and false signals with the use of the current candle
OR
Incur the lag built-into using the previous closed candle. When using a 4hr aggregation the lag will be the length of those 16 candles. Which will not be updated at all.

Most members choose repainting; as those kinds of lags will make most indicators unprofitable.
I choose to run in the other direction from MTF indicators.

While MTF indicators are useful in providing insight into the trend on higher timeframes; they are never meant to be used as part of the buy signal calculations.
The market was bullish over the last few months, so the effect would not have been as self-evident.
But as we head into the August/September Correction there will be more volatility with much more repainting.

Great insight. Thanks for such an in-depth response. Much appreciated.

So, just to clarify, if I'm using the 15min timeframe, with a 4-hour buy signal, I'd essentially have to make the buy signal look back 16 bars via "[16]"? and however many bars, like "[96]" for the previous daily bar ect?

I have had to scrap 3 of the 6 buy signals I was using. And since then I've been able to generate (1) promising one and I have like 4 more in the works in hopes that at least one of those sticks.

I'm not sure if you're aware, but it is in-fact possible to use a higher timeframe and for the indicator to not repaint, but it's far and few between that it's possible. I have one buy signal that doesn't repaint even if the trade completely turns against me, but it's using a "broad" range within that higher timeframe and the timeframe is high enough to where the levels don't move off relative small price action. So, in my experience it is possible, but it's super hard to achieve. They can't be used to actually enter, the current timeframe has to do that(I've now realized), but it can be used to determine the overall market conditions in those higher timeframes.

 

[MerryDay]

Great insight. Thanks for such an in-depth response. Much appreciated.

So, just to clarify, if I'm using the 15min timeframe, with a 4-hour buy signal, I'd essentially have to make the buy signal look back 16 bars via "[16]"? and however many bars, like "[96]" for the previous daily bar ect?

Since posting my last post and the time it took to get approved, I have had to scrap 3 of the 6 buy signals I was using. And since then I've been able to generate (1) promising one and I have like 4 more in the works in hopes that at least one of those sticks.

I'm not sure if you're aware, but it is in-fact possible to use a higher timeframe and for the indicator to not repaint, but it's far and few between that it's possible. I have one buy signal that doesn't repaint even if the trade completely turns against me, but it's using a "broad" range within that higher timeframe and the timeframe is high enough to where the levels don't move off relative small price action. So, in my experience it is possible, but it's super hard to achieve. They can't be used to actually enter, the current timeframe has to do that(I've now realized), but it can be used to determine the overall market conditions in those higher timeframes.

It sounds like you have a good grasp of how MTF Multi-timeframe studies work.
Here is a summary of everything discussed in this thread:
https://usethinkscript.com/threads/...-indicators-in-thinkorswim.16606/#post-131220

 

[vlalani86]

I saw somewhere someone has posted on how to avoid the repainting on the ToS. I am new to ToS, So I have no idea how to apply the below settings. I appreciate your help.

To avoid repainting, ToS recommends using the open of the next bar open[-1] for entry in strategies,
and use of the previous bar [1] for studies. Thereby, never experiencing repainting again!

 

[MerryDay]

I saw somewhere someone has posted on how to avoid the repainting on the ToS. I am new to ToS, So I have no idea how to apply the below settings. I appreciate your help.

To avoid repainting, ToS recommends using the open of the next bar open[-1] for entry in strategies,
and use of the previous bar [1] for studies. Thereby, never experiencing repainting again!

If you are referring to 'real' repainters,
No, unfortunately, they cannot be made to not repaint.
read more:
https://usethinkscript.com/threads/answers-to-commonly-asked-questions.6006/#post-57833

If you are referring to a MTF indicator, yes, your solution will work. But as you can see from reading through the above posts, it is not a great solution.
The real solution, is to make better use of MTF indicators:
https://usethinkscript.com/threads/utilizing-the-power-of-mtf-indicators-in-thinkorswim.16606/

If you are referring to indicators triggering on the current candle. That is not called 'repainting' per se, but yes, it will update with inter-candle activity, until the candle closes.
If you don't want to take your move until after the candle closes, yes, your method will work.
The majority of indicators on this forum, do use the current candle because whether the indicator's trigger happened during the candle formation or after, most members want to jump on it.