# Thread: Comparision of Estimation methods

1. ## Comparision of Estimation methods

I am trying to arrive at an estimation method that is simple and accurate for maintenance projects. I am comparing Cosmic FFP with other methods – Function points, Feature points, Use Case points, Regression Analysis, etc.
Is there any case studies or data to compare these methods against the following factors-Accuracy, Simplicity, Estimation effort, Industry recognition and Applicability for sizing maintenance requests... THANKS

2. ## Re: Comparision of Estimation methods

There is no algorithmic method that works. Predicting the future is going to be faulty.

Estimation is not an easy task. If we over estimate, good projects may not get started. If we under estimate, we may end up working long unplanned hours and have to make difficult explanations to management. People are rarely rewarded for accurate estimates that contribute to a successful project, but few forget those who's bad estimates lead to disaster. Heroes don' do estimates. [img]/images/graemlins/smile.gif[/img]

That being said...

We can always provide and estimate. And, that estimate will be wrong. (See Paul Coombs' book IT Project Estimation.)

Paul Coombs describes a "secret method" to estimation that is the best I have found. This secret method is very simple. Yet, it will still be wrong. I think we often make estimation overly complex and look for the magic formulas when none work. (This is why companies can sell expensive tools and workshops to assist and teach estimation methods.)

So what's the secret method?

Here it is:

1. Make a list of the things to be estimated
2. Estimate each of them
3. Add all the estimates together

The hardest part can be making the list of items to be estimated. No matter what method we apply, if we don't estimate the right things, we will be wrong.

3. ## Re: Comparision of Estimation methods

Completely agree with Ben on this.Just add the efforts required for managing the project as well to it.Also different projects require different estimation technigues which again varies from project to project.Web based project estimation might vary from the client server application.

4. ## Re: Comparision of Estimation methods

I have been having a fair amount of success lately with BPE - Bug Point Estimation. There was a discussion in May or June on these forums about it.

Try searching for "bug point estimation."

5. ## Re: Comparision of Estimation methods

Function point estimation is not Test effort estimation.In one of the discussion it was stated that it is overall effort and also function point is not broken down into various sub tasks to track the development.

So question remains how would one break down function point into Testing work break down structure and estimated time for each task i.e.

1. Test Planing.
2. Test case creation.
3. Test case review.
4. Number of builds that will be required and there testing duration.
5. Test bed creation.
6. Builds acceptance tests.
7. Test case execution effort (Permutations on various servers,browsers ,Localization testing etc).
a.Automation effort.
b.Performance testing effort
8. Release effort (Installation testing,Help docs review,Configuration files review etc).

How will function point and Use case point resolve this?

6. ## Re: Comparision of Estimation methods

Below is a great resource on what Function Points are all about. http://ourworld.compuserve.com/homep...comp/fpfaq.htm

We used to use a hybrid of FPC to estimate the detail that you're looking for. We broke our applications down in Features, Functions and Test Points. We used a Bottom up approach to estimate what it would take to create new scripts, maintain current scripts, run the scripts, predict defect counts, etc for new test Points that had been added to or maintained within our applications.

A Test Point is any element that a user can interact with or is calculated, etc on a screen. Test Points were combined together to create Functions and the back into Feature Sets.

For example:

Feature - Book a Flight
Function - One Way Flight
Test Point - From City, To City, Date of Departure, etc

We baselined the data to identify how long it would take to create data for each function based on the # of Test Points in the function and used those baselines to estimate regression runs, or creation of new functionality that had similar attributes as existing Features, Functions and Test Points.

I hope this helps.

7. ## Re: Comparision of Estimation methods

Using function points to determine test duration is a bit dangerous if the component for which function points were identified has no code changes. It's helpful for a test strategy but not estimation.

there's standard estimation techniques which can be understood via a few google searches

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•

vBulletin Optimisation provided by vB Optimise v2.6.0 Beta 4 (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
BetaSoft Inc.
Digital Point modules: Sphinx-based search
All times are GMT -8. The time now is 10:14 AM.