Isolated Collapsed But-For Delay Analysis Methodology

Jyh-Bin Yang1 and Pi-Chung Yin2

Abstract: Schedule delays are common in construction projects. Although many methods have been developed for analyzing and
measuring schedule delays for construction projects, each method has functional limitations and use prerequisites. No one method is
acceptable for all project participants under all circumstances. This study presents the isolated collapsed but-for 共ICBF兲 method, an
innovative delay analysis method for construction projects. During analysis, the ICBF method requires as-planned and as-built schedules
as well as identified liability documents with key delay events to perform its analysis approach. Based on its application to an illustrative
case and comparisons with other methods, the ICBF method is effective for delay analysis. Results provided by the proposed method can
be easily traced to the actual case in an as-built schedule.
DOI: 10.1061/共ASCE兲CO.1943-7862.0000016
CE Database subject headings: Construction management; Scheduling; Claims; Liability; Analytical techniques; Dispute resolution;
Delay time.

Introduction Yang 共2005兲 compared two available delay analysis systems, the
Primavera Claim Digger and Schedule Analyzer Professional
Schedule delays are common in construction projects. Previous 共Bala Cynwyd, Pa.兲, to discover their functionalities for delay
studies have attempted to discover delay information derived analysis. Yang illustrated the system limitations in delay analysis
from as-planned and as-built schedules, which are the basis for when delay information is limited, concluding that available com-
resolving delay disputes and claims. Many recent studies have mercial delay analysis systems are still not well developed; there-
focused on analyzing delay values accurately and systematically fore, keeping computer-based scheduling information with delay
共Hegazy and Zhang 2005; Arditi and Pattanakitchamroon 2006兲. evidence for various delay analysis approaches is necessary. Al-
However, these methods have different analytical processes and though some advanced delay analysis methods have been devel-
oped and thoroughly assessed, including commercial systems, the
require different information when employed. For a complicated
available methods cannot satisfy practical requirements for delay
delay case, different methods may generate different results based
analysis. That is, practitioners still require an alternative method
on their different viewpoints or the different information utilized.
for resolving complex cases.
No one method is currently acceptable for all project participants
This study presents the isolated collapsed but-for 共ICBF兲
or suitable for all situations.
method, an innovative delay analysis method for construction
Arditi and Pattanakitchamroon 共2006兲, while discussing how projects. The remainder of this paper is organized as follows.
to select a delay analysis method, concluded that selecting a fea- Section 2 briefly introduces the documents typically used by
sible analysis method depends on a variety of factors such as available delay analysis methods. Section 3 then reviews several
available information, time of analysis, methodology capabilities, popular delay analysis methods and examines their concepts and
and time, funds, and effort allocated for analysis. Furthermore, characteristics. Section 4 describes two referred methods by this
Yang and Kao 共2009兲 reviewed 18 delay analysis methods and study and then describes the ICBF method, including its concepts,
then focused on three dynamic methods 共snapshot analysis, win- analytical processes, prerequisites and mechanisms, liability cal-
dows analysis, and isolated delay types兲 to provide a step-by-step culation and algorithms for baseline schedule development. Sec-
flowchart that discloses the analytical processes that are useful tion 5 demonstrates the use of the ICBF method via its application
when selecting a method. These two studies concluded that avail- to a test case and compares application results with results ob-
able methods, although accepted and used by courts, are imper- tained by other methods. Conclusions are finally drawn in Sec. 6,
fect in resolving complicated construction schedule delays. along with recommendations for future research.
Available delay analysis methods are academic products not
ready for use by most owners and contractors. Therefore, com-
mercial software systems have been developed for delay analysis.
Documents for Delay Analysis
1
Professor, Institute of Construction Management, Chunghua Univer- In general, schedule information for tracing delay events after a
sity, No. 707, Sec. 2, Wu-Fu Rd., Hsinchu, Taiwan 共corresponding construction project is completed varies. Elnagar and Yates
author兲. E-mail: jyhbin@chu.edu.tw 共1997兲, surveyed project control personnel from contractors,
2
Engineer, T.Y. Lin Taiwan Consulting Engineers, Taipei, Taiwan.
project owners, and the U.S. Army Corps of Engineers, concluded
Note. This manuscript was submitted on February 6, 2008; approved
on October 15, 2008; published online on March 27, 2009. Discussion that schedule networks, daily construction reports, quality control
period open until December 1, 2009; separate discussions must be sub- reports, correspondence, progress curves, productivity measure-
mitted for individual papers. This paper is part of the Journal of Con- ment, and change-order logs are the key delay indicators. Al-
struction Engineering and Management, Vol. 135, No. 7, July 1, 2009. though these documents resemble those of construction
©ASCE, ISSN 0733-9364/2009/7-570–578/$25.00. schedules, not all documents are used for analysis by available

570 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2009

methods. The most commonly used document is the schedule ties in this method are consistent with actual project occurrences
network; other documents are evidence of delays. At least four 共Arditi and Pattanakitchamroon 2006; Sgarlata and Brasco 2004兲.
schedule network types, i.e., as-planned, as-built, adjusted, and This method uses an as-built schedule as a basis for analysis
entitlement schedules, are commonly used by available delay comparison and then removes delays from the as-built schedule to
analysis methods. Although associated with different delay meth- collapse the schedule. The difference between the as-built sched-
ods, these four schedules are usually accepted by courts. In prac- ule and collapsed schedule is the final analysis results 共Zack
tice, not all construction projects have these four schedules. The 2001兲.
proposed ICBF method uses as-planned and as-built schedules In employing the CBF method, analysts should first determine
based on their easy accessibility. the analysis viewpoint: owner or contractor. This method then
splits its analytical procedure into two streams. Based on the own-
er’s perspective, this method removes nonexcusable 共NE兲 delays
Available Delay Analysis Methods 共delays caused by the contractor兲 from the as-built schedule to
calculate the anticipated project completion date. Therefore, the
Many delay analysis methods are available 共Alkass et al. 1995; difference in time between project completion date on the as-built
Arditi and Pattanakitchamroon 2006; Gothand 2003; Hegazy and schedule and the date calculated by removing delays caused by
Zhang 2005; Mohan and Al-Gahtani 2006; Yang 2005; Zack the contractor is the duty to owner. On the contrary, based on the
2001兲. The most common methods include the global impact contractor’s viewpoint, this method removes excusable delays
technique, net impact technique, adjusted as-built critical path 共delay not caused by the contractor兲 from the as-built schedule to
method 共CPM兲 technique, as-planned expanded technique, col- calculate the anticipated project completion date. Therefore, the
lapsed but-for technique, snapshot technique, time impact tech- difference in time between the project completion date on the
nique, windows technique, and isolated delay type technique as-built schedule and the date calculated by removing delays not
共Yang et al. 2006; Yang and Kao 2007兲. The analytical ap- caused by the contractor is the duty to contractor.
proaches of available methods can be categorized as process- Notably, the duty calculated by the CBF method is simply
based, mathematical, and computer-based models 共Yang and Kao assigned to the owner or contractor. However, in practice, a delay
2009兲. can be caused by a designer/architect. This delay type during
Despite the availability of many methods, some rules are typi- construction phase is assigned to the owner because the contract
cally required that help users select an appropriate method. Bub- was already signed by the owner and contractor. If a delay is
shait and Cunningham 共1998兲 and Arditi and Pattanakitchamroon caused by a designer/architect and translated to the owner, the
共2006兲 discussed the information required for delay analysis. owner can make a claim against the designer/architect under a
Their findings were similar: different delay analysis methods re- liquidated damage clause. This situation is not discussed further
quire different documents and consume different levels of re- in this study.
sources. Nevertheless, different methods may provide different
analysis results based on the same evidence. Available methods Isolated Delay Type Method
have received extensive criticism due to the following failures: The isolated delay type 共IDT兲 method is derived from the snap-
共1兲 concurrent delays cannot be recognized or calculated; 共2兲 the shot analysis 共SA兲 method to overcome drawbacks in previous
critical path method cannot be executed in analysis and critical methods when classifying proper delay types, concurrent delays,
path changes cannot be considered; 共3兲 the relative cost of float and real-time CPM analysis 共Alkass et al. 1996兲. Therefore, the
consumption is not considered; and 共4兲 analysis is not contempo- IDT method can be regarded as a modified SA method. The docu-
raneous with delay timing 共Arditi and Pattanakitchamroon 2006; ments used by the SA method include the as-planned, as-built,
Bordoli and Baldwin 1998; Gothand 2003; Mbabazi 2005; Ng et and any revised schedules implemented during a project. For
al. 2004; Yang and Kao 2009兲. delay analysis, total project duration is divided into periods. The
To provide an alternative to delay analysis for resolving con- dates of these periods usually coincide with major project mile-
current delay and liability distribution problems, this study pro- stones or significant changes in planning or times when a major
poses the ICBF method, which is a systematic analysis that traces delay or a group of delays occur 共Mohan and Al-Gahtani 2006兲.
delay events with their relative evidence in an as-built schedule. Regarding its analytical processes, the IDT method uses an
as-planned schedule as a basis for comparison, and performs
delay analysis based on the extracted analysis section schedule
and considers the liabilities of contract parties explicitly. First, the
Methodology Development
IDT method divides the as-planned schedule into digestible peri-
ods. At the start of each analysis period, the IDT method modifies
Referred Methods activity durations and relationships before and after the start point
by inputting actual start dates, finish dates, and durations, and by
Collapsed But-For Method maintaining original durations and logical relationships. This
The collapsed but-for 共CBF兲 method is also known as the as-built modified schedule is an adjusted as-planned schedule and the
collapsing technique, as-built technique, and as-built subtracting basis for delay comparisons. Similar to the CBF method, the IDT
impacts 共Bubshait and Cunningham 1998; Zack 2001兲. Based on method performs delay analysis for two independent viewpoints;
an as-built schedule that accompanies daily construction reports, i.e., those of the owner and contractor. The IDT method inserts
minutes, and correspondence, analysis results can be traced to excusable and nonexcusable delays into the adjusted schedule to
original documentation. This original documentation accounts for reflect the impact of delays from the contractor and owner per-
the truth underlying analysis results. Due to the use of an as-built spectives, respectively. From one analysis viewpoint, delays
schedule, the collapsed but-for method is usually used as a tool in caused by another party are imposed on the adjusted as-planned
claim management 共Zack 2001兲. Courts and dispute resolution schedule. Therefore, the difference in time between project
boards in the United States favor the CBF method because activi- completion date on the adjusted as-planned schedule and the date

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2009 / 571

after by inserting the delays is the duty to the opposition. By
summarizing delay values at each analysis period, the IDT
method can allocate liability to each contract party.

Isolated Collapsed But-For Method

Basic Concept
The proposed ICBF method uses the concept of IDT method, but
starts with an as-built schedule because it reflects actual start and
finish dates and actual duration. Therefore, the ICBF maintains
the advantages associated with the CBF and IDT methods, and
overcomes their associated drawbacks. The proposed ICBF
method can be regarded as a combination of the CBF and IDT
methods. The ICBF method adopts the concepts described as fol-
lows. For each analysis period, if no delay occurred, project
completion date is adjusted to 共collapsed as兲 a reasonable date.
This adjusted date is the new baseline for determining the impact
of a delay only considering the liability of the owner or contrac-
tor. Therefore, the difference between the new baseline and an
impacted project completion date caused by an analyzed delay is
the delay responsibility for the analyzed contract party. Further-
more, in clearly identifying liability associated with analyzed ac-
tivities, the ICBF calculates the delay value under an extracting
window that can consist of several activities.

Analytical Process
Fig. 1 shows the analytical processes in the ICBF method that
uses an as-built schedule as a basis for comparison, and performs
delay analysis to interpret explicitly contract party liabilities.
First, this proposed method divides the as-built schedule into di-
gestible periods. At the start of each analysis period, the ICBF
method modifies durations and logical relationships for the activi-
ties before and after the time-point using the mechanisms de-
scribed in the following subsection. This modified schedule is an
adjusted as-built schedule and the basis for delay comparison.
Similar to the IDT method, the ICBF method performs delay
analysis for two independent viewpoints; namely, those of the
owner and contractor. The ICBF method inserts excusable and
nonexcusable delays into the adjusted schedule to reflect the im-
pact of delays from the contractor and owner perspectives, re-
spectively. Therefore, from one analysis viewpoint, delays caused
by the other party are imposed on the adjusted as-built schedule.
After comparing the adjusted as-built schedule with a modified
schedule, the difference in time between project completion date
on the adjusted as-built schedule and the date after inserting de-
lays is the duty of the opposition. Finally, the ICBF method allo-
cates liability to each contract party by summarizing each party’s
delay values from each study period.
Fig. 1. Analytical processes of the ICBF method
Prerequisite and Mechanism
The ICBF method uses the same documents as the IDT method, activities before the time-point by giving their actual start dates,
including the as-planned, as-built, and any revised schedules 共at finish dates, and durations. For activities started but not complete,
minimum, the as-planned and as-built schedules are required兲 that the ICBF method inputs actual start dates as start dates, modifies
implemented during project execution. For delay analysis, total the remaining durations using the original durations minus de-
actual project duration is divided into a number of periods. The layed durations, and maintains original logical relationships with
algorithm used by the ICBF to determine the analysis periods is their successors. If activities completed after the time-point were
identical to that used by the IDT method. The key time-points analyzed in previous scenarios, their durations are replaced by
include the time critical path共s兲 changed, major delay start and their as-planned durations and the logical relationships in the as-
end dates, and periodic times 共e.g., monthly, every two months, or planned schedule are retained; the other activities proceed from
each season兲. their actual start dates with as-planned durations. In the final sce-
At the start date of each analysis scenario, the ICBF method nario, the baseline duration used for comparison must be the
modifies start and finish dates as well as durations for completed project duration of an as-planned schedule.

572 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2009

 Fig. 2. As-planned and as-built schedules for test case

Liability Calculation Algorithm Baseline Schedule Development Algorithm

For each analysis scenario, Eq. 共1兲 determines the anticipated The project completion date of a baseline is determined by adopt-
project completion duration 共Durbase
i 兲, which is calculated by ac- ing five methodologies that determine duration, start date and
tual duration 共Duracti 兲 before an analysis time-point plus as- finish date for each activity. The first method is suitable for ac-
planned duration 共Duriplan兲 for activities completed after the tivities completed before the period start date 共Type I兲. Such an
analysis time-point. In each analysis scenario, Eqs. 共2兲 and 共3兲 are activity is put on its actual start to actual finish. The second
the algorithms for determining anticipated project completion du- method is used for activities started but not completed before the
ration caused by the owner and contractor, respectively. In those period start date 共Type II兲. Such an activity is started on its actual
equations, Durown
i and Durcon
i are durations for which the owner start date and has the duration that reserves its actual duration but
and contractor are responsible, respectively; DurEC EN
i , Duri , and removes delays that occurred during the analysis period. The third
NE
Duri are the impacted durations caused by excusable- method is used for activities started after the period start date and
compensable 共EC兲 delays, excusable-noncompensable 共EN兲 de- not analyzed during previous periods 共Type III兲. Such an activity
lays, and nonexcusable delays, respectively. Therefore, in each is scheduled on its actual start date with its as-planned duration
delay analysis period, delay liability for the owner 共Dutyown
i 兲 and 共all delays are removed兲 or starts considering the logical relation-
contractor 共Durcon
i 兲 is calculated using an apportioned duration ship with its predecessor共s兲 and maintains its as-planned duration.
minus the original anticipated project completion duration, as in The fourth method is used for activities started after the period
Eqs. 共4兲 and 共5兲. After determining the delay liability in each start date that were analyzed during previous periods, but still
analysis scenario, the proposed method summarizes project delay have delay共s兲 that occurred during this period 共Type IV兲. Such an
liability for each contract party 共Dutyown for the owner and activity starts on its actual start date with the duration that main-
Dutycon for the contractor兲 from all analyzed periods using Eqs. tains adjusted duration during previous period, and further re-
共6兲 and 共7兲: moves the delays that occurred in this analysis period. The fifth
method is used for activities started after the period start date and
Durbase
i = Duract plan
i + Duri 共1兲 were analyzed and completed during previous periods 共Type V兲.
Such an activity maintains its as-planned duration but starts on
+ 共DurEN the date considering the logical relationship with its predeces-
Durown = Durbase i + Duri 兲 共2兲
EC
i i
sor共s兲.

Durcon base
i = Duri + DurNE
i 共3兲
Case Study
Dutyown
i = Durown
i − Durbase
i 共4兲
Case Description
Dutycon con base
i = Duri − Duri 共5兲 This study uses a modification of the test case developed by
Kraiem and Diekmann 共1987兲. This case was also used by Alkass
n et al. 共1996兲 to verify the capability of their proposed methodol-
Dutyown = 兺
i=1
Dutyown
i 共6兲 ogy. This illustrative case has ten activities and an original total
duration of 23 days. The project is completed in 41 days, with
18 days of project delays. Fig. 2 shows the as-planned and as-
n built schedules. This study uses this case to demonstrate the func-
Duty con
= 兺
i=1
Dutycon
i 共7兲 tionality of the proposed ICBF method. The simulated delays
consist of nonexcusable 共NE兲, excusable noncompensable 共EN兲,

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2009 / 573

Table 1. Delay Information for Demonstration Case
NE delay EN delay EC delay
Total
Act. Dur. Start Finish Dur. Start Finish Dur. Start Finish delay
A 3 2 4 1 10 10 — — — 4
B 1 6 6 2 7 8 1 4 4 5
1 10 10
C 3 15 17 — — — 2 20 21 5
D — — — — — — — — — 0
E 1 15 15 5 21 25 2 13 14 9
1 20 20
F — — — — — — 2 26 27 2
G 1 21 21 — — — 1 20 20 2
H — — — 1 33 33 1 28 28 2
I 1 31 31 2 38 39 2 36 37 7
2 33 34
J — — — 2 38 39 — — — 2
Sum 13 13 12 41

and excusable compensable 共EC兲 delays. Table 1 lists the start on their actual start dates or their predecessor’s finish dates—day
and finish dates for each delay event. Notably, this illustrative 25 for H, day 30 for I, and day 35 for J. After completing these
case has a concurrent delay that happened with Activities A and B manipulations, the project completion date is day 37 关Fig. 3共c兲兴.
on day 10. These two activities are on the parallel critical paths, This information indicates that the contractor deserved no delays
but have separated delays caused by different project participants 共37− 37= 0兲. Table 3 summarizes analysis results.
共Activity A has an EN delay; Activity B has an NE delay兲.
Scenario 2
Analysis Process Scenario 2 starts on day 24 and ends on day 29; therefore, activi-
ties E, F, and H are analyzed. In analyzing activity status for
Under the rules for determining analysis period, this demonstra- building the scenario baseline, activities A, B, C, D, and G are
tion determines analysis periods by considering start and finish of Type I, and maintain their actual start and finish dates; activity E
activities A, C, F, and I on the longest path in the as-built schedule is Type II, starts on its actual start date, and lasts for 13 days
共Fig. 2兲. Fig. 3共a兲 shows the determined four analysis periods. The 共actual duration 共15 days兲 minus a 2-day delay during the period兲;
period in scenario 1 is from day 30 to 41; the period start dates for
activity F is Type III, starts on its actual start date, and lasts for
scenarios 2, 3, and 4 are days 24, 12, and 1, respectively.
4 days 共actual duration 共6 days兲 minus a 2-day delay during the
period兲; activity H is Type IV, starts on its actual start date, and
Scenario 1
lasts for 9 days 共10 days in the previous period minus a 1-day
In this case, activities A to G are finished before the period start
delay during this period兲; activities I and J are Type V 共analyzed
date 共day 30兲; therefore, these activities are Type I, and maintain
in the previous scenario兲, maintain their as-planned durations, and
their actual start and finish dates. Activity H is Type II and, hence,
proceed when their predecessors finish. These manipulations form
is started on day 25 and has a duration of 10 days 共actual duration
the baseline in scenario 2 and determine the project will be com-
is 11 days; however, the EN delay on day 33 is removed兲; thus,
day 34 is the finish day. Activities I and J are Type III. Activity I pleted on day 36.
starts on day 30 and has a duration of 5 days; activity J starts after By considering owner liability 共EN and EC delays兲, the dura-
its predecessor H is completed 共day 34兲 and has a duration of tions of activities E, F, and H are extended to 15, 6, and 10 days,
3 days. In this scenario, no activity is Type IV or Type V. Based respectively. These modified durations replace the durations in the
on the manipulations above, project completion date is day 37, baseline schedule. Additionally, activities I and J have adjusted
which is the baseline for further comparisons. start dates based on their predecessor’s finish dates. Fig. 3共d兲
In considering owner liability, the ICBF method assigns delays shows the owner delay-loaded schedule, which reveals that the
not caused by the contractor 共EN and EC delays兲 on the impacted owner deserves a 1-day delay 共37− 36= 1兲. If contractor liability
activity. The durations of activities H, I and J are 11 共1-day 共NE delay兲 is considered, the baseline schedule is extended to a
delay兲, 9 共4-day delay兲, and 5 days 共2-day delay兲, respectively. contractor delay-loaded schedule. However, activities E, F, and H
Furthermore, these activities start on their actual start dates—day have no NE delays in this period. Fig. 3共e兲 shows the contractor
25 for H, day 30 for I, and day 36 for J. After completing these delay-loaded schedule, which reveals that the contractor receives
manipulations, the project is completed on day 40 关Fig. 3共b兲兴. no delay responsibility.
This information indicates that the owner deserved 3 days of
delay 共40− 37= 3兲. Scenario 3
However, when contractor delay is imposed on the baseline, Scenario 3 starts on day 12 and ends on day 23; hence, activities
the ICBF method assigns the delay caused by the contractor 共NE B, C, D, E, and G are analyzed. The baseline for this scenario is
delay兲 on the impacted activity. At this time, the durations of constructed by executing the following manipulations. Activity A
activities H, I, and J are 10 共no delay兲, 8 共3-day delay兲, and is Type I, and maintains its actual start and finish dates; activities
3 days 共no delay兲, respectively. Additionally, these activities start B and D are Type II, start on their actual start dates, and last for

574 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2009

 Scenario 4 Scenario 3 Scenario 2 Scenario 1

ID Act. -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 4
1 A
2 B
3 C
4 D
5 E
6 F
7 G
8 H
9 I
10 J

(A) analysis period for different scenarios

ID Act. Dur. Start Finish -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
1 A 11 d 1 11
2 B 10 d 3 12
3 C 12 d 12 23
4 D 9d 10 18
5 E 15 d 13 27
6 F 6d 24 29
7 G 5d 18 22
8 H 11 d 25 35
9 I 9d 30 38
10 J 5d 36 40

(B) owner-delay loaded schedule for scenario 1

ID Act. Dur. Start Finish -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
1 A 11 d 1 11
2 B 10 d 3 12
3 C 12 d 12 23
4 D 9d 10 18
5 E 15 d 13 27
6 F 6d 24 29
7 G 5d 18 22
8 H 10 d 25 34
9 I 8d 30 37
10 J 3d 35 37

(C) contractor-delay loaded schedule for scenario 1

ID Act. Dur. Start Finish -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
1 A 11 d 1 11
2 B 10 d 3 12
3 C 12 d 12 23
4 D 9d 10 18
5 E 15 d 13 27
6 F 6d 24 29
7 G 5d 18 22
8 H 10 d 25 34
9 I 5d 30 34
10 J 3d 35 37

(D) owner-delay loaded schedule for scenario 2

ID Act. Dur. Start Finish -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
1 A 11 d 1 11
2 B 10 d 3 12
3 C 12 d 12 23
4 D 9d 10 18
5 E 13 d 13 25
6 F 4d 24 27
7 G 5d 18 22
8 H 9d 25 33
9 I 5d 28 32
10 J 3d 34 36

(E) contractor-delay loaded schedule for scenario 2

ID Act. Dur. Start Finish -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
1 A 11 d 1 11
2 B 10 d 3 12
3 C 9d 12 20
4 D 9d 10 18
5 E 12 d 13 24
6 F 4d 21 24
7 G 4d 18 21
8 H 9d 25 33
9 I 5d 25 29
10 J 3d 34 36

(F) owner-delay loaded schedule for scenario 3

ID Act. Dur. Start Finish -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
1 A 11 d 1 11
2 B 10 d 3 12
3 C 10 d 12 21
4 D 9d 10 18
5 E 7d 13 19
6 F 4d 22 25
7 G 4d 18 21
8 H 9d 20 28
9 I 5d 26 30
10 J 3d 29 31

(G) contractor-delay loaded schedule for scenario 3

ID Act. Dur. Start Finish -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
1 A 8d 1 8
2 B 8d 1 8
3 C 7d 9 15
4 D 9d 9 17
5 E 6d 9 14
6 F 4d 16 19
7 G 3d 18 20
8 H 9d 15 23
9 I 5d 20 24
10 J 3d 24 26

(H) owner-delay loaded schedule for scenario 4

ID Act. Dur. Start Finish -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
1 A 10 d 1 10
2 B 7d 1 7
3 C 7d 11 17
4 D 9d 8 16
5 E 6d 8 13
6 F 4d 18 21
7 G 3d 17 19
8 H 9d 14 22
9 I 5d 22 26
10 J 3d 23 25

(I) contractor-delay loaded schedule for scenario 4

Fig. 3. Analysis result for different scenarios

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2009 / 575

Table 2. Analysis Results by the ICBF Method
Contractor Owner
Complete Delayed Complete Delayed
Analysis Analysis Target date for
Scenario period activity comparison Date Day Date Day
1 30–41 H, I, J 37 37 0 40 3
2 24–29 E, F, H 36 36 0 37 1
3 12–23 B, C, D, E, G 30 31 1 36 6
4 1–11 A, B 23 26 3 26 3
Sum 4 13

10 days 共no delay is removed兲 and 9 days 共no delay is removed兲, lays on an individual’s perception, delay liability caused by the
respectively; activities C and G are Type III, start on their actual contractor and owner is obtained for each scenario. The ICBF
start dates, and last for 7 共a 5-day delay is removed兲, and 3 共a method concludes that the contractor causes 4-day delay 关the
2-day delay is removed兲 days; activity E is Type IV, starts on its owner has the responsibility of 14-day delay 共18− 4 = 14兲兴 and the
actual start date, and lasts for 6 days 共13 days in the previous owner causes 13-day delay 关the contractor has the responsibility
period and then a 7-day delay is removed兲; activities F, H, I, and of 5-day delay 共18− 13= 5兲兴. Therefore, the contractor would
J are Type V, maintain their as-planned durations, and proceed claim 9 days 共14− 5 = 9兲 for compensation.
when their predecessors finish. The baseline schedule of scenario If no concurrent delay occurs, the summation of delay respon-
3 is finished on day 30.
sibilities caused by the contractor and the owner equals to the
Similar to the processes used in the previous two scenarios,
value of total project delay. Otherwise, the summation value
owner and contractor liabilities are considered to determine their
minus total project delay will equal to the concurrent delay. In
impacts on duration. Fig. 3共f兲 shows simulation results obtained
by considering delays caused by the owner; project duration is this test case, the summation of delay responsibilities caused by
extended to 36 days. Therefore, the owner deserves a 6-day delay the contractor 共5 days兲 and the owner 共14 days兲 is 19 days.
共36− 30= 6兲. Furthermore, Fig. 3共g兲 shows simulation results ob- Therefore, the concurrent delay 共1 day兲 can be calculated by
tained by considering contractor delays. The contractor receives a using the summation value 共19 days兲 minus total delay days
1-day delay. 共18 days兲. This proves the capability of determining concurrent
delay.
Scenario 4
Scenario 4 starts on day 1 and ends on day 11. During this period,
only activities A and B are delayed and analyzed. Notably, the Comparison of Analysis Results with Those of
final baseline schedule obtained by the ICBF method is the same Other Methods
as the as-planned schedule, in which the project ends on day 23.
In dealing with owner liability, a 1-day delay for activity A and a
To elucidate the differences between the ICBF and other delay
3-day delay for activity B are imposed on the baseline schedule.
analysis methods, the same demonstration case is used. Table 3
Fig. 3共h兲 shows analysis results, indicating that the owner de-
shows analysis results obtained by different methods. Some
serves a 3-day delay 共26− 23= 3兲. Furthermore, a 3-day NE delay
for activity A and a 2-day NE delay for activity B are imposed on analysis results were obtained from Alkass et al. 共1996兲.
the baseline schedule to explore the liability of contractor. Fig. The global impact technique, which summarizes all delay val-
3共i兲 shows analysis results, indicating that the contractor deserves ues 共38 days兲 and compares these values with delay
a 3-day delay 共26− 23= 3兲. evidence—18-day delay 关as-built schedule 共41 days兲 minus as-
planned schedule 共23 days兲兴—is the difference between entitle-
Summary ment completion date and the as-built completion date caused by
Table 2 lists the analysis results for each scenario, in which the project acceleration 共Alkass et al. 1996兲. The net impact tech-
analysis period, activity and calculation results are summarized. nique considers delay concurrency, and determines delay values
Based on the comparison of the target completion date at baseline based on the difference between completion dates of the as-
and adjusted completion date by considering the impacts of de- planned 共23兲 and as-built 共41兲 schedules. The final result is

Table 3. Comparison of Different Methods

Delay liability
Method Contractor Owner Final result
Global impact technique — — 18 delayed days, not specified for owner or contractor
Net impact technique — — 18 delayed days, not specified for owner or contractor
As-planned expanded technique 9 16 7 days for compensation by contractor
Collapsed as-built analysis 18 11 7 days for compensation by owner
Isolated delay type 4 13 9 days for compensation by contractor
Isolated collapsed but-for method 4 13 9 days for compensation by contractor

576 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2009

18 days; however, this is not apportioned to the contractor or technique and collapsed as-built analysis method provide a differ-
owner. ent delay liability distribution for the contractor and owner. The
The as-planned expanded technique compares an as-planned right for claiming from the owner or contractor differs because
completion date with an adjusted as-built completion date derived the project was delayed by 18 days. Delay liability in these two
from the as-planned schedule based on individual liability. After methods is not distributed to the owner and contractor accurately.
the NE delays are added to the corresponding activity, project Based on above comparisons, the IDT and ICBF methods have
duration is 32 days. Analogously, if the EN and EC delays are the same analysis results. Therefore, the ICBF method is a good
applied to their corresponding activities, project duration is alternative for executing delay analysis, while as-planned and as-
39 days. Therefore, the differences between the as-planned built schedules, and identified liability documents related to key
completion date 共23兲 and adjusted as-planned completion dates delay events are available.
共32 and 39兲 are analysis results for contractor liability 共9 days兲
and owner liability 共16 days兲. By considering the net impact from
both parties, the contractor would claim 7 days for compensation. Advantages and Disadvantages of ICBF Method
The collapsed as-built analysis compares the as-built comple-
tion date with a modified as-built completion date derived from Although no rule is available for determining which method pro-
the as-built schedule by considering removing individual liability. vides the most accurate result, the proposed ICBF method has
After removing the NE and EN plus EC delays from their corre- advantages similar to those of the IDT, including: being a system-
sponding activities, project durations are 30 and 23 days. There- atic and dynamic analysis method, and resolving concurrent de-
fore, the differences between the as-built completion date 共41兲 lays. Furthermore, the proposed ICBF method has a clear
and modified as-built completion dates 共30 and 23兲 are the delay descriptive analysis process with a definite baseline schedule de-
values for the owner 共11 days兲 and the contractor 共18 days兲. Con- velopment algorithm. This information is useful for consistent
sequently, the owner could make a compensation claim 共liqui- delay analysis and is a basis for developing a computer-based
dated damages兲 from the contractor for 7 days; the contractor can delay analysis system. Additionally, delay analysis via the ICBF
debate this claim as this delay is concurrent with owner delay. method is based on the as-built schedule, with which a contractor
The isolated delay type executes delay analysis based on ex- generally has daily construction reports or similar evidence to
tracted schedule fragnets by taking time-based snapshots of the trace delay types, actual activity start and finish dates, and delay
as-built schedule. Generally, the rule for determining analysis pe- events. Analysis results by the ICBF method would provide a
riods is arbitrary; that is why the isolated delay type method re- more convinced outcome than some other delay analysis methods.
ceives certain criticisms. In this test case, only the finish dates of Compared to the IDT method, the ICBF method derives its
major delay activities 共activities A, B, C, E, and I with larger methodology considering the suitability for computerization and
delay values兲 are considered. Furthermore, this study tried to se- the consistency for computation. Most construction projects are
lect certain separated time points that divide total duration complicated and usually have hundreds or thousands of activities.
equally. Because the start analysis document for the IDT and If they have simply delay situations, the ICBF method can pro-
ICBF methods are varied, this study, herein, tried to use different vide definite processes and rules to obtain clear results manually.
analysis time periods compared with the ICBF method used. If they suffer complex delay situations, to do delay analysis
Based on above considerations, three snapshots were utilized. The manually is impossible. This study had constructed a descriptive
analyzed periods are day 1 to 11 共activity A was completed兲, day procedure that would be implemented in a computer system and
12 to 27 共activity E was completed兲, and day 28 to 41 共the project would be helpful for delay analysis. Therefore, the ICBF method
was completed兲. is better than the IDT method in implementation.
In this test case, the project was completed on days 23, 30, and The ICBF method has attempted to provide an alternative for
39 for periods 1, 2, and 3, respectively. If the NE delays are helping delay analysis; however, it still has some limitations de-
considered, the project was completed on days 26, 31, and 37 for scribed as follows. 共1兲 It requires the as-planned and as-built
periods 1, 2, and 3, respectively. Therefore, the contractor should schedules, and classified delay information with evidence; other-
take responsibility for 3, 1, and 0 delay days in each period, 4 wise, it cannot work. 共2兲 It cannot resolve complicated delay situ-
delay days in total. Notably, in period 3, comparison results reveal ations; e.g., project acceleration, delays in appending or missed
that the contractor is responsible for −2 delay days 共this might activity, and delays due to third party. 共3兲 It does not have a robust
show that the contractor has a 2-day acceleration effort兲. Under rule for determining analysis periods. 共4兲 It does not have an
the situation in which the project exceeded project completion algorithm to distribute delay liability for encountered concurrent
date, the negative delay value means zero delay responsibility. In delays.
considering the EN and EC delays, the project was completed on
days 27, 38, and 40 for periods 1, 2, and 3, respectively. Hence,
the owner has responsibility for 4, 8, and 1 delay days for periods Conclusions
1, 2, and 3, respectively, and 13 delay days in total. This analysis
result differs from the result 共owner is responsible for 16 days and While schedule delays occur frequently in construction projects,
contractor for 6 days兲 obtained by Alkass et al. 共1996兲—that be- how to best identify the liability of contract participants has re-
cause the analysis periods in this and their study differ. ceived considerable attention. Although many methods have been
Table 3 summarizes the test results when using different delay developed and used for analyzing and measuring construction
analysis methods. All methods provide different analysis results schedule delays, no one is acceptable to all project participants
with various distributions of delay liability for the owner and and suitable for all delay situations. This study presents the ICBF
contractor. The global impact technique and net impact technique method, a novel innovative delay analysis method, which adopts a
produce the same results 共by chance兲; i.e., the project has 18 process-based analysis approach to resolve concurrent delays and
delay days. These two methods could not provide accurate results liability distribution problems, provides systematic analysis, and
for delay sources and delay liability. The as-planned expanded easily traces delay events using an as-built schedule. The concept,

JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2009 / 577

analytical process, prerequisites, mechanism, liability calcula- DurECi ⫽ impacted duration caused by
tions, and baseline schedule development algorithms for the ICBF excusable-compensable delays in the i period;
method have been thoroughly illustrated. The proposed method DurENi ⫽ impacted duration caused by
maintains the advantages of CBF and IDT methods. The efficacy excusable-noncompensable delays in the i
of the proposed method is proven with a test case. period;
Schedule delays in the construction industry are normally de- DurNEi ⫽ impacted duration caused by nonexcusable
rived from complex delay situations, result in various delay phe- delays in the i period;
nomena, and involve many project participants. Therefore, a Durown
i ⫽ adjusted completion dates considering
systematic and accurate delay analysis method is necessary to owner’s duty in the i period;
resolve delay problems. Based on research results, this study pro- Duriplan ⫽ as-planned duration in the i period.
vides following suggestions for further study. Dutycon ⫽ delay liability to contractor; and
1. Extending the liability calculation algorithms in ICBF Dutyown ⫽ delay liability to owner.
method to scrutinize different delay types 共EC, EN, and NE兲
can help delay analysts to obtain accurate responsibilities for
different viewpoints. This study has provided a fundamental References
in resolving this problem. Further study will focus on deriv-
ing existing equations to calculate delay liability for different Alkass, S., Mazerolle, M., Tribaldos, E., and Harris, F. 共1995兲. “Com-
delay types. puter aided construction delay analysis and claims preparation.” Con-
2. Evaluating the performance of the ICBF method and other str. Manage. Econom., 13共4兲, 335–352.
similar methods for diverse and real cases can improve their Alkass, S., Mazerolle, M., Tribaldos, E., and Harris, F. 共1996兲. “Construc-
acceptability in the construction industry. However, illustra- tion delay analysis techniques.” Constr. Manage. Econom., 14共5兲,
375–394.
tive cases, covering all delay situations or real delay cases,
Arditi, D., and Pattanakitchamroon, T. 共2006兲. “Selecting a delay analysis
are hard to retrieve. Developing and collecting these cases method in resolving construction claims.” Int. J. Proj. Manage.,
are essential for further development and evaluation. 24共2兲, 145–155.
3. As the analysis results generated by different methods vary, Bordoli, D. W., and Baldwin, A. N. 共1998兲. “A methodology for assessing
developing a solid method that resolves various delay prob- construction project delays.” Constr. Manage. Econom., 16共3兲, 327–
lems and generates accurate results acceptable to both con- 337.
tract parties is required. Increased delay problems bring an Bubshait, A. A., and Cunningham, M. J. 共1998兲. “Comparison of delay
urgent demand for such a method. A well-structured method analysis methodologies.” J. Constr. Eng. Manage., 124共4兲, 315–322.
should consider easily accessible documents or evidence, Elnagar, H., and Yates, J. K. 共1997兲. “Construction documentation used
thereby increasing the acceptability of delay methods. as indicators of delays.” Cost Eng., 39共8兲, 31–37.
Gothand, K. D. 共2003兲. “Schedule delay analysis: modified windows ap-
4. Most available delay analysis methods are not implemented
proach.” Cost Eng., 45共9兲, 18–23.
in popular project management systems or supported by Hegazy, T. M., and Zhang, K. 共2005兲. “Daily windows delay analysis.” J.
those systems. This is a barrier to applying these methods for Constr. Eng. Manage., 131共5兲, 505–512.
resolving real delay problems. Developing easy-to-use sys- Kraiem, Z., and Diekmann, J. 共1987兲. “Concurrent delays in construction
tems embedded in, based on, or supported by available com- projects.” J. Constr. Eng. Manage., 113共4兲, 591–602.
mercial project management systems can advance the Mbabazi, A., Hegazy, T., and Saccomanno, F. 共2005兲. “Modified but-for
application of delay analysis methods. method for delay analysis.” J. Constr. Eng. Manage., 131共10兲, 1142–
This study presents a novel innovative delay analysis method 1144.
as an alternative for resolving construction delay problems. Even Mohan, S. B., and Al-Gahtani, K. S. 共2006兲. “Current delay analysis
with the availability of highly effective methods, complete delay techniques and improvement.” Cost Eng., 48共9兲, 12–21.
Ng, S. T., Skitmore, M., Deng, M. Z. M., and Nadeem, A. 共2004兲. “Im-
information must be retained regardless of the method used.
proving existing delay analysis techniques for the establishment of
delay liabilities.” Constr. Innovation, 4共1兲, 3–17.
Sgarlata, M. A., and Brasco, C. J. 共2004兲. “Successful claims resolution
Acknowledgments through an understanding of the law governing allocation of risk for
delay and disruption.” CM ejournal, CMAA, 具http://cmaanet.org/
The writers would like to thank the National Science Council, ejournal.php典 共Dec. 12, 2007兲.
Taiwan, ROC, for financially supporting this research under Con- Yang, J. B. 共2005兲. “Comparison of delay analysis software for construc-
tract Nos. NSC95-2221-E-216-048 and NSC96-2221-E-216-027- tion projects.” Proc., 33rd Annual Canadian Society for Civil Engi-
MY2. The writers are also thankful to the reviewers for their neering (CSCE) Conf., Toronto, Canada, CT-115-1–9.
valuable suggestions and comments. Yang, J. B., and Kao, C. K. 共2007兲. “A knowledge map for delay analysis
development.” Proc., Fourth International Conference on Construc-
tion in the 21st Century, Gold Coast, Australia.
Yang, J. B., and Kao, C. K. 共2009兲. “Review of delay analysis method-
Notation
ologies: A process-based comparison.” The Open Construction &
Building Technology Journal 共under review兲.
The following symbols are used in this paper:
Yang, J. B., Kao, C. K., and Lee, Y. Y. 共2006兲. “System requirement
Duracti ⫽ actual duration in the i period; analysis of a construction delay analysis system.” Proc., 23rd Inter-
Durbase
i ⫽ anticipated project completion duration in the national Symp. on Automation and Robotics in Construction, Tokyo,
i period; 102–106.
Durcon
i ⫽ adjusted completion dates considering Zack, J. G. 共2001兲. “But-for schedules—Analysis and defense.” Cost
contractor’s duty in the i period; Eng., 43共8兲, 13–17.

578 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2009