Skip to content
All library documents

Why ADX and Directional Indicator Calculations Differ Across Platforms

Article Quant Q&A · Author: Antony
Read the original

Summary

The document describes a replication problem: calculating 14-period ADX, positive DI, and negative DI in a spreadsheet produces early values that differ substantially from trading platforms. The author has tried a formula found in an encyclopedia and asks what could explain the mismatch, using a stock chart comparison as context.

The sole answer suggests that the discrepancy may come from different calculation conventions: the encyclopedia formula may use a newer approach, while the spreadsheet or platforms may use Wilder’s original method. It notes that software may offer both variants. No formulas, comparative data, or validation procedure are provided, so this is a plausible explanation rather than a demonstrated diagnosis; matching initialization and smoothing conventions would need to be checked for the specific platforms.

Key ideas

  • ADX and directional indicator values can vary across implementations.
  • The answer attributes the mismatch to newer calculations versus Wilder’s original method.
  • The question concerns early values from a 14-period calculation.
  • The document provides no formulas or platform-specific verification.

Tags

Technical indicatorsStatistics
Full text
# How to fully replicate ADX + DI Indicators in Excel?


# How to fully replicate ADX + DI Indicators in Excel?












For black box testing, I was hoping that I could replicate the ADX + DI+ and DI- indicators that are provided in trading platforms such as ThinkOrSwim, ScottradeElite etc.

However, I noticed that even after applying the formula I found on WikiPedia, I noticed that initially calculated data tends to be off by a large margin when compared to ThinkOrSwim or ScottradeElite.

Example spreadsheet is based on HIMX @ Google Drive. In the spreadsheet, you will see the ADX + DI oscillators that I calculated. Here is the screenshot of my calculated ADX + DI

Below is the screenshot based on ScottradeElite and ThinkOrSwim for HIMX

The period for the ADX & DI are 14 periods.

I have been spending countless hours, trying to figure out why my calculated results would deviate from the existing trading platforms. However, I have not been able to spot mistake. Can anyone help me out by pointing out what I am missing? Thanks a bunch!

## Answer by cuong (score 1)

https://quant.stackexchange.com/a/17885

I think wiki calculated upon new way and your excel calculated upon Wilder way (Wilder book).

I have same issue when use both method. Look like new Metastock 5 software offer both method.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.