As Built Critical Path

The "Critical Path" extends from "Today" (the data date) to the end of the project. Due to the dynamic nature of daily CPM activity, the critical Path can switch daily for one chain of activities to another and back again the next day. Each day, there will be a critical path (but not necessarily the same path.)



I am saying that the only thing that the critical path has in common from one day to the next is that ON THE DATA DATE, a certain set of activities were both critical and active. This is the only thing that matters. Note what was critical on any date. Add all of the dates together (one at a time) and you have a list of critical, active activities. This is the only truth.



The daily critical ’path’ may hop from one chain of activities to another without a logical relationship between then to ’model’ it. The ’critical path’ may actually hop between two or more activities as they progress. It does not matter. Each and every day, something is critical and active. All of the rest is either unknown or planned.



Don’t focus on one single critical path. Instead, focus on what was critical and active on any given day. The difference between the two is the difference between classical CPM theory and the truth.



Ron Winter.

Jaco - Yes, you can compute the CPM for completed activities but do not forget that the CPM is computed using activity remaining durations. The remaining duration for a completed activity is zero. This does not tell you much about the critical path in the past.



Philip - You are correct ahout ’hindsight’ but what about ’Today’? You can certainly tell what is critical today. Tomorrow will eventually come and when it is ’today’, you will be able to determine the critical activity for that as well. My point is that for any given date, you should be able to state what was critical for that date - even for dates in the past. I have not said anything about the Critical Path - only the critical activity for any given date.

Hi Gerry and others.

I have been watching this discussion with keen interest.



My opinion is, on the question that Gerry originally asked, is that there is no thing such as an as-built critical path, simply because if something is complete, how can it be critical? ie Hindsight is 20/20 vision. Anybody trying to establish the critical path in retrospect did not do his/her planning properly in the first place. Further, the critical path will not always remain in the same chain of activities. If Gerry is working on a claim, any tampering with the original and as-built programmes would make such programmes nul and void, as back-up material in the claim.



Regards



Philip

Hi Ronald



I was looking at your paper and was wondering why can’t you calculate Float on activity’s that is completed. Remember you can change the dates to Non dates (numericals) even though computer systems don’t calculate float in activitys that is in the past you as a person can calc it manual.



I know actual should not have float but to do an as Build Critical Path you must calculate the float manual.



You can do a forward and a Backword pass on actual start & finish dates



Cheers

I have no qualms on explaining myself. Take a look at my Published Articles Section http://www.ronwinterconsulting.com/published.htm of my website. There is a listing there for, "AACEi 2004 Annual Conference: How to Manually Determine the As-Built Critical Path." It is a later version of the paper that I delivered at the conference. They let me live after presenting it, so it can’t be too far off.



Ron Winter.

I believe that the secret to understanding the type of network that you describe is to look at a re-computed CPM on a day-to-day basis. Start from the latest date, move the data date back one day, adjust the remaining durations of each active activity to reflect the actual work days remaining and re-compute the CPM. Then note which activity (or activities) was critical and active that day. Repeat this process, moving the data date to one day earlier until you reach the start of the project.



This is the algorithm that we use for a software product that we call “As-Built Critical Path.” It is a part of the Schedule Analyzer Forensic series. You can see more at http://www.ronwinterconsulting.com/forensic_main.htm. I also presented a paper on performing this algorithm manually at this year’s AACEi Annual Convention. You can find it somewhere under http://www.aacei.org/.



Ron Winter

Wow! Just a fascinating thread! And from my point of view, very serendipitous, as I’m currently co-authoring an article which incorporates the very idea that Vladimir argued: that every project has a CP, whether or not CPM is used.



I went back and read the whole thread. Here are my thoughts:



1. I one hundred percent agree with Vladimir, David and others who feel that there is such a thing as an ABCP. I think part of the problem lies in the mis-definition of the CP (by many authors and authorities!) as the path with zero float. We all know that, during implementation, a path might have either positive or negative float (supercritical). Even during planning, the longest path can have float due to calendar-based schedule constraints (No Earlier Than). Is such a longest path still "worthy" of being called critical? Absolutely!



The reasons for calling the longest path "critical" are not limited to (1) the fact that it has no float, so that if it slips, it becomes longer -- two other reasons, equally important, are that (2) the longest path determines the length of the project, and (3) if you want to shorten the project, you have to shorten the longest path.



Both (2) and (3) may have relevance for the post-project analysis: (2) in the sense of the ABCP, and (3) if the question ever becomes why were actions not taken that could have shortened the project (or avoided delay) by addressing the DRAG on the activities on the "currently estimated CP".



Thus I completely agree with Vladimir: EVERY project has a CP, whether planned using CPM or not, because it is that CP that ultimately determines project length. In fact, the ABCP may be called the "true" CP; during planning, what we are really dealing with is the "estimated" CP. Yes, our estimated CP may move around, as either scope changes or delays occur, or as duration estimates turn out to be wrong.



Surely it is HUGELY important, from both liability and knowledge database points of view, to know when, how and why our estimates of the CP changed? And whether, at any moment, our estimates were based on best available data, or simply on incompetent performance (either by contractor or client)?



In another thread, there was a question about whether a computerized "expert system" would ever be able take over the planning function. I expressed doubt, at least for the near future. But let us think for a second about a "perfect" planning system -- would not that system be, in fact, a crystal ball, allowing us to perfectly foresee what would happen (scope changes, client delays, fragnets, resource shortages, etc.) and incorporate them into the plan so as to have the best (i.e., most profitable) project?



And finally, would not the plan that the crystal ball created be, in fact, the ABCP?

[Duplicate entry deleted.]

Hi Gerry



Please Explain



Is it not the case that the piecemeal working and intermittent effort upon the activity (especially the kind of activities that Jaco mentioned earlier) seriously compromise the insertion of logic into an *as-built* programme.

Dear all,



This has become a very interesting thread. I would like to return to something that Roger Gibson said earlier. Assuming that Gerry is undertaking some sort of a delay analysis, or is preparing a claim, then that is why he is considering an as-built critical path approach.



In so doing, Gerry also needs to consider what Roger described as the “windows” approach. I concur with Roger, in that it is my preferred method of analysis as well. The reason why is this…..



In producing an as-built critical path programme, you can see what was the longest critical path chain throughout the project as it was at the end of the project. This is fine, but it lacks one vital consideration. The problem is that it considers the project only after it is completed looking back, but ignores what was known about at the time. Let me give a very simple example. Let’s say a house is being built and there are problems with building the foundations. The design is late and this causes a 3 week delay. This is on the critical path and an extension of time of 3 weeks is quite properly given. All quite straightforward and simple.



However, the design also incorporates a swimming pool in the garden, which is a contractor designed item. At the time of granting the extension of 3 weeks due to the late foundations, the swimming pool excavation is on programme and all seems okay with the pool. But, the contractor takes far too long to design the pool and order the water treatment equipment and this turns out to be the eventual cause of the house being completed 10 weeks late.



So if a retrospective as-built programme was drawn, it would show the as-built critical path going through the swimming pool, but not through the main house foundations.



Would you say that this as-built critical path programme was correct? It does, quite correctly, show that the eventual critical path went through the swimming pool, but some would argue that it would ignore the true cause of early delay – the house foundations. This is where a “windows” approach is preferred by some courts, because the windows approach, as Roger has stated, considers what was critical AT THE TIME. This is the important point. An as-built critical path can only consider what eventually turned out to be critical, but doesn’t consider what was known about at the time.



So, Gerry will have to bear this in mind when (if) he produces his as-built critical path. There are drawbacks to this system, as there are with most.



Gary France

Chairman

Planning Engineers Organisation

Hi Gerry



Yes it is Practical to do an As Built Critical Path on these Items.

Your Critical Path Should be something like this

From Back to Front.

1) Building Handover

2) Commision BMS - SAT -BMS FAT BMS - Development BMS -

2) Commision HVAC

2) Commision Fire Detection

2.1) Commision Permenet Power (LV) - Delivery MCC - Delivery UPS etc

2.2) Commision Permananet Power (HV)- Delivery Transformer -Delivery Switchgear etc

2.3) Power Supply Available

3) Installation of Controls Instrucments

4) Installation of HVAC - Delivery HVAC-Order HVAC

5) Access to Install HVAC

Your other activitys should run Concurrent.



Please note this just to give you some ideas.



Cheers

I think Uri is right in general to make As built critical path Schedule. But if you want to keep As built Schedule for future reference of planning then we should consider only those delay points which generally happens in project to identify the critical path. I mean to say that if the delay something like unavailibity of fund or vendor’s bankrupt comes on critical path we should ignore that to say that is not the critical path even though that will be longest path. Then we can say second longest path is really ctritical for future reference of planning but for cliam purpose the longest path will be considered.

There are two issues here:

1. The legel issue

2. The programming issue.



When analysing delays, the first place to look at should be the CONTRACT. Only when there is a reference in the contract to actual delays to the date for PC, the As Built programme becomes relevant. The Critical Path then would be the longest path in the as built programme.



In most cases the As Built Critical Path is not relevant, since the analysis is carried out RETROSPECTIVELY. The as built programme is used only as a record of actual dates, and not as a real programme with float, links etc.

Gerry, re your question: "Can a true as built critical path be created?"



I hope so, otherwise I have wasted much of my working life!! (not an entire impossibility, I grant you!!) ;-)



Cheers,



Stuart



www.rosmartin.com

If you keep a history of project schedule versions you will know all activities that were considered critical in the past.

They may belong to as built critical path or not. In any case both information is useful for the project performance analysis and lessons learned.

Does anybody think that this information is useless?

Yes, critical path is most commomnly used term in planning and also it is the longest path based on the correct logic whereas "As Built Path" is not commonly used and not depenable.

Daya,



Jaco has defined it correctlt; it is the longest path that is established. However, this is often termed as the ’As Built critical path’.





Roger Gibson

Dear Jaco,



What you have said is correct, The longest path is the critical path. As Built Critical Path is based on the actuals happening at the site i.e. based on the actual update; when the project is in progress due to the rate of progress of work experienced at site the original critical path(or the baseline critical path) could be altered and may be at the end of the project a new longest path could be found than the expected one. Once the project was over we can say it as a As Built Schedule just like saying as-built drawings. The As built schedule should reflect the history of the project in genuine. Obiviously the longest path formed in the As-built schedule would be a As-built critical path but not necessarily to represent the original baselined critical path.



Regards



Daya

My answer is yes.



An as build longest Path can be produced. I would suggest to forget about the logic issues. Please note the diffinition of a critical path is the longest path. Not the longest path due to logic. With other words if the logic is wrong (We are all human) the critical path is something else than the longest path.



But why do you require this. I am not sure what type of project you are working on / or refering too.



Cheers

Gerry,



Personally, I don’t think an ’As Built’ critical path with true as built logic can be created. Just because activity B started 3 weeks after activity A started does not mean that the as built logic between these activities is a start-to-start of 3 weeks. It could be because of a number of factors, e.g. shortage of resources or awaining information.



In carrying out a retrospective analysis of actual delay my preferred approach is a series of progress snapshots, or ’windows’, which show which activity was critical at each snapshot. From this, and further investigation and interrogation of the network, progress and other records an ’As Built’ critical path is established.



Roger Gibson

Gerry,

Whilst it is acceptable to change the logic by adding links in creating an As Built Critical Path, it is important to justify and explain the reasoning for each change.
Another important ’verification’ check of the As Planned network are the activity durations. I have often seen eot/delay calculaions based on original programme durations that were excessive and therefore resulted in eot/delay calculations being longer than reasonable or realistic.

Roger Gibson

Gerry / Stuart,



I agree, there is nothing inherently wrong in adding retrospective links. In fact, you must do it because chances are that the original programme logic will have changed over time. Apologies if I didn’t make that clear.



Gary France

Chairman

Planning Engineers Organisation

Gerry,



Gary is right that you have to take care when you add in retrospective logic – hence my caveat about ensuring that there are no inconsistencies with the original logic.



In principle, however, provided that your logic is sound and reasonable, there is nothing inappropriate in inserting retrospective logic (subject to the points that Gary and I have made).



Cheers,



Stuart



www.rosmartin.com

Uri



Now we are getting somewhere!!!



But come on - if one can work out what was critical for every day of a project how can that be meaningless??? I think you’re being coy, and simply unwilling to abandon your previous position.



The identification of what was critical in retrospect is a vastly meaningful conclusion. If you were to successfully present that conclusion alone to an arbitrator / judge, you would be carried from the room atop shoulders with clients and lawyers slapping you on the back as you departed. Same for the site, where EOT’s haven’t been dealt with instantly.



Linking all the days together would identify the ABCP (- = activity and + = critical).

+++++++----

-------+++++++++++++++++-------

-----------++++++++++++++++++++++++*



And there is no problem with having concurrent criticality!



Regards

David

Gerry,



Can we assume that you are putting together some sort of a claim document to demonstrate what actually happened on a project? If you are, then you have to be very careful in adding links retrospectively. You will / may have to justify each one if they are different from the originally programmed logic. This is because, by introducing new logic, you might well change the critical path from that which would have occurred in your as-built programme had you retained the original logic.



Many experts / delay analysis consultants have been criticised for adding retrospective logic that change the critical path. In effect, by adding links retrospectively, you can make your as-built critical path show whatever you want it to show. Be wary of falling into this trap. You will need to be certain that you can justify all changes, because if you do not, or cannot, than be prepared to do battle big time with whoever you send this to.



Good luck.



Gary France

Chairman

The Planning Engineers Organisation

www.planningengineers.org

Gerry,



Provided that the logic as inserted is not inconsistent with the other related activities, there is nothing wrong in adding it into as-built already-constructed activities.

This happens all the time when trying to demonstrate that a Delaying or Relevant Event (call it what you will!!) has had an impact on the CP.



Hope this helps



Cheers,



Stuart



www.rosmartin.com

Gerry,

It is essential that an as-built CP is developed if an argument arises over EOT to a Contract.



The as-built CP will sequentially trace the work activities that were executed for the project to reach completion, or up to the time under consideration. If completion is reached, or shown to be reached, later than originally planned, then you have a delay. If all or part of that delay is attributed to reasons not applicable to the Contractor, then the Contractor may be entitled to an EOT, which may equal all or part of the delay.

Subsequently the Contractor may be entitled to costs related to the EOT.



But the foundation for all of this is the as-built CP. If the Contractor cannot demonstrate that the as-built CP was pushed out for reasons not attributable to himself, then he cannot claim an EOT.



And remember that an as-built CP is not just the final one produced at the end of a project; it is the CP that has occurred from the start of the project up to the time in the schedule under consideration.



Cheers,



Stuart



www.rosmartin.com

A Complex discussion even though it seems simple.



The definition of the "Critical Path" is the longest Path through the schedule and the answer is yes.



I have observed that you will always have a critical path (more than one) in the process industry. But what I have noticed is that the path will change during the project live from planned to final (Different Activity).



The reason why is once a critical path has been identifyed the project does everything in their power to reduce the critical path once you have reduced it you will noticed that something else becomes critical. And then you would do everything in your power to reduce this.



Also you end up giving more slack (delay) on your non critical activity’s so that you can address your critical path.



So in the end of the project you end up with various critical paths (also called a photo finish)



So yes "critical path" does exist in As Build Project I sometimes use the Term Critical Chain (Everything Finishing together but all part of the Critical Path) but I know that somebody is marketing "Critical Chain method" I have not seen what their definition of a critical chain is.



If it was not for the longest path we will be able to do a project overnight.





Cheers

David B.,



1. Forgetting the word "path" for a second, and assuming excellent records exist, do you think you could work out what was critical to completion on Day 59 of a particular project that was already complete?



YES.



If your answer is Yes, then surely working it out for every day will represent the as-built critical activities throughout the job? This would be the ABCP, though perhaps another term for it might help.



WORKING IT OUT FOR EVERY DAY WOULD BE MEANINGLESS AS THE CRITICAL PATH MAY CHANGE DURING THE EXECUTION OF THE PROJECT I.E. THERE IS NO SINGLE ’OVER-ALL’ CRITICAL PATH. RATHER, THERE MIGHT BE AS MANY ’CRITICAL PATHES’ AS YOUR ANALYSIS POINTS...

Hi Planners



The point of my earlier question was that if in order to create an as built programme you need very good records and information i.e. a regularly updated programme with actual dates etc. I really don’t see the need in producing an as built programme with subjective logic which would only produce one longest path (can’t be a critical path if it is a historical document).

In a live regularly progressed programme the critical activities can change from update to update, and this is the key information that is most useful and reliable in a claims situation.

If the programme is properly administered and maintained as it progresses what better information do you need? In most cases claims do not become an issue if you maintain records good enough to produce an as built programme.



If a claim situation arises then any analysis that is produced is prepared against a programme that the client most likely reviews on a regular basis as part of the reporting process and not a programme produced purely in substantiation of a claim.



IMHO an as built critical path sounds like a chocolate tea pot.

Andrew,



Answering your questions:



’Well claims persons, if you can’t look back at a project’



To me, it would be wrong to look back, you have to "go back" in time and then look FORWARD in order to decide what was critical at the time. I do not think there need for ’a long hard think’.



I can identify critical activities, but I’d use the currect programme when the delay occurred (or became evident) and not the ’ABCP’, which to me does not exist.

Well claims persons, if you can’t look back at a project and determine what was a critical activity at a peticular point in that project then I’d have a long hard think.



EOT’s usually attach themselves to a critical activity or one which becomes critical due to the delay, so if you can’t identify them or they don’t exist what else leads to a EOT becoming due (in general)?

Hi Planners,



Some really interesting points in these threads.



I have constructed as-built programmes on a number of occasions for producing collapse analysis and that is all.



My question is, if an As built CP does exist, and I am not sure it can be reliably determined, to what use can it be put?



(Loaded question)



Regards



Paul

Uri,



Good point, but one that I think is relying on semantics.



Can I pose a question to you -



1. Forgetting the word "path" for a second, and assuming excellent records exist, do you think you could work out what was critical to completion on Day 59 of a particular project that was already complete?



If your answer is Yes, then surely working it out for every day will represent the as-built critical activities throughout the job? This would be the ABCP, though perhaps another term for it might help.



If your answer is No, please explain why?



Philip / Sigfredo - what would your answer be?



David

David Barry,



I say nay - as planning by definition refers to the future, and critical path is a planning concept, whereas AB is merely a record of what happened (i.e. past) in the project.



Garry France in his post from 7 December 2004, 18:19 provided us with an excelent example whereby prospective and retrospective analyses can lead to completely different results.



the AB in my opinion can not go together with the CP.



I think an AB v AP analysis is not in accordance with most forms of standard contracts, requiring a retrospective analysis.

Sigfredo



With the greatest respect, who made you the forum Judge? The burden of proof is not on one side of the debate. It would be more interesting if you put down arguments and logic why you think the ABCP doesn’t exist! Or is your point, simply, that since we’ve failed to convince you then it can’t exist?



My sense of the statistics you would get from this forum are that:

1 the substantial majority say Yes,

2 the courts say Yes

3 Philip and one or two others, to their credit, justify a No position

4 A number of people have changed their views from early in the forum.

5 We all agree a practical approach is better than computer jockey planning



Come on you nay sayers, let us hear the arguments for the No campaign. Oh, and laywers begone!



David

Philip,



Not one of those who believe that there is ABCP was able to convince me as well.



If I go back to one of my posts, I proposed to this forum to get statistics on who believe that there is ABCP and those who do not. Moreover, I would like to know which field of specialization they working on.



Cheers,

Se

Hi andrew,

Thanks for the suppott, it is like finding the "Holy grail", let’s be practical/

Philip,



So what’s new

I have disagreeed with ABCP since the birth of the thread

Hi Andrew,



I forgige the grammar, simply fot the sake of practicbility or alternatively pragmatism, this thread was hi-jacked by the lawyers, and forgot practical things.



Regards



Philip