2c:null 2d:["$","title",null,{"children":"Understanding Cycle Time Distribution Shapes | Troy Magennis | Observable"}] 32:T67a,md`# Understanding Cycle Time Distribution Shapes When software development work is completed, the cycle-time can be computed simply as the date-time of completion minus date-time of starting. It is the elapsed time from start to finish. You must though be clear about the definition of these boundaries, what is start and finish and to make sure it is consistently captured. Some early work [The Economic Impact of Software Development Process Cycle Time Analysis](https://github.com/FocusedObjective/FocusedObjective.Resources/raw/master/Presentations/The-Economic-Impact-of-Software-Development-Process-Choice-Cycle-time-Analysis-and-Monte-Carlo-Simulation-Results.pdf) by me. I looked at cycle time distribution shapes and development process to give contextual advice based on cycle time distribution shape and scale. This notebook aims to help make the identification and advice practical in your teams. *Note: You could go through your entire working life without needing to know this. This article is an advanced topic for research and deep understanding. Don't stress if you don't yet understand it, it probably isn't critical to living long and prospering.* ## Cycle time distribution basics Work is often the completion of multiple steps. Each step can take between a low and high amount of time. Imagine if we sum together the result of throwing a six-sided dice (values 1 to 6) representing the number of days it takes to do pieces of work. Change the number of dice rolls and see how it changes the distribution of the sum of that many rolls a few thousand times to build a pattern (distribution) of what results are more likely to others.`33:T603,md`Observe that 1 dice has equal chance of each value, and by 6 to 7 dice rolls, the distribution pattern is starting to take the shape of a Bell Curve. Additive processes will converge on a Bell Curve and this is the reason why it occurs so often in nature. For example, the heights of humans will converge on a Bell Curve because our height is the sum of all our different bones, each bone having an average for the region and sex plus or minus some variability. If our work only had sequential steps with variability that sum together, Cycle-time would also converge into this bell shape. The fact it doesn't means there is more at play in software development than just plain addition of sequential steps. ## Common Software Development Cycle-Times I cover the basics of cycle time shape in the paper [The Economic Impact of Software Development Process Cycle Time Analysis](https://github.com/FocusedObjective/FocusedObjective.Resources/raw/master/Presentations/The-Economic-Impact-of-Software-Development-Process-Choice-Cycle-time-Analysis-and-Monte-Carlo-Simulation-Results.pdf), but to recap, after examing hundreds of teams data follw some predictable shapes that I speculate the causes (warning, the paper is pretty boring reading, so grab at least one glass of wine before settling in to read it). The general hypothesis is that process choice factors will determine basic shape; And iteration length and the average cycle time of the work ill determine the time scale (x-axis). For the moment we will just focus on shape. `34:T8bb,md`I need to stress, that no process is better than another baed on distribution shape alone and there are reasons why every shape makes sense in certain development contexts. Much of that context means that shape isn't directly related to the work (or people doing the work) themselves - it is out of their control. Some of the context that causes shape to change are - - Reasons work can be queued waiting (eg. waiting for specialist resources) - Reasons work can be delayed due to (originally unknown) complexity - Rework required (how well known the work is known) ### The reason cycle-time distributions are skewed is that work can be impacted by un-related and inconsistently impacting delays. This notebook explores this working hypothesis - Cycle-time distributions are causes by - 1. The known work time with variability related to the work itself 2. 'n' number of delays with 'p' chance of occurrence probability that may or may not be related to the work itself In practical terms, it is well known and logical that software work we develop gets impacted by delays. Some work gets almost no delays, while other work gets impacted by every delay imaginable. Some delays are due to the work itself (harder than first thought, needing external help), and others are just circumstance and bad timing (hitting a long weekend). It is the combination of these delays that alters the shape. We can produce the common cycle time shapes using a pretty simple process simulation. This simulation is an initial amount of time plus a combination of up to five delays. We can alter the probability of each delay and see how different combinations play out over 100,000 simulations. Play with this simulator and observe the impact of delay probability on distribution shape. (remember, the height of the bar is 'how often' the x-axis is 'how long in time' - this is called a histogram) Try - 1. Set all the delay probabilities to 0 - does the result make sense? 2. Set Delay 1 to 50% and all the other delays to 0 - does the result make sense? 3. Add delays up to 20% one at a time. How does it change the shape? 4. Change the cycle time and delay time. Does it change the shape at all? Does it change the time (x-axis) at all? `35:T9a0,md`In the real world, work cycle time wouldn't be exact, nor would the length of each delay be equal. It is the variability that fills in the gaps and makes the observed cycle times from our industry look like smooth distributions. For now though, you can see the initial shapes forming just through this basic simulation to make the following observations. ##### Observation 1 - Work cycle time and delay time DON'T change the shape Apart from a bit of jitter due to random number generation in this simulator, altering the work time and delay time sliders DOESN'T alter the shape. Shape is depednet on number of delays and probaility of delays alone. ##### Observation 2 - As the number of delays possible increases, the peak shifts right As we add delays, the most common outcome (called the mode) shifts to the right. ##### Observation 3 - As the sum of the probabilities approaches 100%, the curve shifts to the right. This is a simplification, but until there is a sum of about 80% chance, the mode will be equal to the cycle-time. ## What does this mean for cycle time reduction? The process generating cycle-times in software development given what we have seen so far is - total cycle-time = development time + p1(delay time) + ... + pn(delay time) This means we can reduce cycle time by addressing these factors - 1. Reducing the development time at the constraint (thanks [Steve Tendon](https://tameflow.com/) for reminding me of this) 2. Reducing the average delay time if it occurs 3. Reducing the number of delays 4. Reducing the probability of delays Context will determine what factor is most bang for the buck. My thoughts are as follows: **Case 1: Cycle time distribution mode peak is at the leftmost value** Development time matters most. Pursue automation. Looks for the highly impacting but rare delays.Adding staff at the constraint will increase throughput. **Case 2: Cycle time distribution mode peak not the leftmost value** Delays matter most. Rank the delays using blocker clustering and find ways with eliminating the most frequent and largest delay time delays. Typically this is dependencies on specialist skills or external teams. Adding staff to reduce development cycle time even at the constraint will have little to no effect. There will be a blend of work types, and both cases of work will be in any systems backlog. Try and tag work types so you can determine what system changes will have the best effect. `36:Tbe4,md `### Hypothesis - Cycle time distribution shape is correlated to as a measure of process efficiency Knowing what we saw in the simulation about the shape only being effected by the number of delays (factor 3) and the probability of those delays (factor 4), it seems possible that the SHAPE of any observed process cycle time is related to the delay profile. My question is, can the shape be a measure of **process efficiency**? and if so, can we determine process efficiency without measuring the exact hands-on and hands-off time for each piece of work. Can we approximate system process efficience from cycle time distribution shape alone. I think we can. Process Efficiency is a measure representing the ratio of hands-on time versus total time for a piece of work. It is generally calculated by tracking the time work is actually being performed on each item and often this number is poorly recorded. Some electronic tools help, but the measure often suffers from when a person or team multi-task and don't indicate work is actually on-hold whilst they do other stuff (making it look like work hands-on time is larger than it is). Even given the benefit of innacurate recording of the waiting time biased towards "better" the general consensus and observed ranges of process efficiency is 5 to 15%. #### Reasons this won't work - brainstormed limitations 1. Some delays are related to work and others systemic. My current view is that the related delays due to work difficulty are in the hands-on time and given typical system process efficiency < 15% (often far less) this will be a small component and likely consistently small negating its impact on shape. 2. The ratio of delay time versus work time could be more important than the delay probability itself. We know if all delay times are equal, the delay time DOESN'T impact shape. But, what if they are all different and one is massively out of scale to the others? There will need to be limitations of some sorts, i'm not sure how common or impactful this will be. 3. Work style change. For example, a team goes from doing well known work to unknown innovative work. I actually think we will detect a change here. But, depending on how much historical data we use, the distribution will be a blend of both. I think a limitation will be stale system data. I'm wondering if we can detect these step changes in distribution using statistical tests for distribution change in different time windows. 4. In low sample counts of historical cycle times, the necessary percentile computation might be poor. To estimate the shape we need two solid percentile values. I use 30th and 70th percentile. But what conditions cause these to be incorrect and impactful. TODO: Troy to prove. 1. Create simulations with known process efficiencies 2. Estimate the process efficiency and correlate to the shape parameters estimated 3. List all of the found limitations and tests for their exposure Email me with feedback, ideas or if you want to help troy.magennis@focusedobjective.com. `37:T6f0,function histogram(data, refData = null) { const width = 640; const height = 240; const margin = ({top: 20, right: 20, bottom: 30, left: 40}); const x = d3.scaleLinear() .domain(d3.extent(data)).nice() .range([margin.left, width - margin.right]); const bins = d3.histogram() .domain(x.domain()) .thresholds(x.ticks(50)) (data); const y = d3.scaleLinear() .domain([0, d3.max(bins, d => d.length)]).nice() .range([height - margin.bottom, margin.top]); const svg = d3.create("svg") .attr("viewBox", [0, 0, width, height]) .style("max-width", `${width}px`); const bar = svg.append("g") .attr("fill", "steelblue") .selectAll("rect") .data(bins) .join("rect") .attr("x", d => x(d.x0) + 1) .attr("width", d => Math.max(0, x(d.x1) - x(d.x0) - 1)) .attr("y", d => y(d.length)) .attr("height", d => y(0) - y(d.length)); if (refData != null) { const binsRef = d3.histogram() .domain(x.domain()) .thresholds(x.ticks(50)) (refData); // to do add reference distribution line var lineCycleTime = d3.line() .x(d => x(d.x0) ) .y(d => y(d.length)) svg.append("path") .datum(binsRef) .attr("fill", "none") .attr("stroke", "orange") .attr("stroke-width", 3) .attr("d", lineCycleTime); } svg.append("g") .attr("transform", `translate(0,${height - margin.bottom})`) .call(d3.axisBottom(x).tickSizeOuter(0)); svg.append("g") .attr("transform", `translate(${margin.left - 6},0)`) .call(d3.axisLeft(y).ticks(1)) .call(g => g.select(".domain").remove()); return svg.node(); }38:T4d1,function histogramDice(data, n) { const width = 640; const height = 240; const margin = ({top: 20, right: 20, bottom: 30, left: 40}); const x = d3.scaleLinear() .domain(d3.extent(data)) .range([margin.left, width - margin.right]); const bins = d3.histogram() .domain([n,(n*6)+1]) .thresholds(50) (data); const y = d3.scaleLinear() .domain([0, d3.max(bins, d => d.length)]).nice() .range([height - margin.bottom, margin.top]); const svg = d3.create("svg") .attr("viewBox", [0, 0, width, height]) .style("max-width", `${width}px`); const bar = svg.append("g") .attr("fill", "steelblue") .selectAll("rect") .data(bins) .join("rect") .attr("x", d => x(d.x0) + 1) .attr("width", d => Math.max(0, x(d.x1) - x(d.x0) - 1)) .attr("y", d => y(d.length)) .attr("height", d => y(0) - y(d.length)); svg.append("g") .attr("transform", `translate(0,${height - margin.bottom})`) .call(d3.axisBottom(x).tickSizeOuter(0)); svg.append("g") .attr("transform", `translate(${margin.left - 6},0)`) .call(d3.axisLeft(y).ticks(1)) .call(g => g.select(".domain").remove()); return svg.node(); }39:T5a0,function histogramProcess(data) { const width = 640; const height = 240; const margin = ({top: 20, right: 20, bottom: 30, left: 40}); const x = d3.scaleLinear() .domain(d3.extent(data)) .range([margin.left, width - margin.right]); const bins = d3.histogram() .domain([cycleTime, cycleTime+(5*delayTime)]) .thresholds(6) (data); /* const x = d3.scaleLinear().domain(d3.extent(data)).nice(5); const histogram = d3.histogram().domain(x.domain()).thresholds(x.ticks(5)); const bins = histogram(data) */ const y = d3.scaleLinear() .domain([0, d3.max(bins, d => d.length)]).nice() .range([height - margin.bottom, margin.top]); const svg = d3.create("svg") .attr("viewBox", [0, 0, width, height]) .style("max-width", `${width}px`); const bar = svg.append("g") .attr("fill", "steelblue") .selectAll("rect") .data(bins) .join("rect") .attr("x", d => x(d.x0) + 1) .attr("width", d => Math.max(0, x(d.x1) - x(d.x0) - 1)) .attr("y", d => y(d.length)) .attr("height", d => y(0) - y(d.length)); svg.append("g") .attr("transform", `translate(0,${height - margin.bottom})`) .call(d3.axisBottom(x).tickSizeOuter(0)); svg.append("g") .attr("transform", `translate(${margin.left - 6},0)`) .call(d3.axisLeft(y).ticks(1)) .call(g => g.select(".domain").remove()); return svg.node(); }3a:T5de,function histogramProcessBar(data) { const width = 300; const height = 150; const margin = ({top: 0, right: 20, bottom: 30, left: 40}); const svg = d3.create("svg") .attr("viewBox", [0, 0, width, height]); const x = d3.scaleBand() .domain(data.map(d => d.value)) .range([margin.left, width - margin.right]) .padding(0.1) /* const y = d3.scaleLinear() .domain([0, d3.max(data, d => d.count)]).nice() .range([height - margin.bottom, margin.top]) */ var ylimit = simCount if (delayp1 + delayp2 + delayp3 + delayp4 + delayp5 > 10) ylimit = simCount / 2; const y = d3.scaleLinear() .domain([0, ylimit]).nice() .range([height - margin.bottom, margin.top]) const xAxis = g => g .attr("transform", `translate(0,${height - margin.bottom})`) .call(d3.axisBottom(x).tickSizeOuter(0)) const yAxis = g => g .attr("transform", `translate(${margin.left},0)`) .call(d3.axisLeft(y)) // .call(g => g.select(".domain").remove()) svg.append("g") .attr("fill", "steelblue") .selectAll("rect") .data(data) .join("rect") .attr("x", d => x(d.value)) .attr("y", d => y(d.count)) .attr("height", d => y(0) - y(d.count)) .attr("width", x.bandwidth()); svg.append("g") .call(xAxis); var labelFont = { family: 'sans-serif', size: '11px' } //svg.append("g") // .call(yAxis); return svg.node(); return svg.node(); }3b:[] 30:{"id":"c821488164dc1cdb","slug":"understanding-cycle-time-distribution-shapes","title":"Understanding Cycle Time Distribution Shapes","thumbnail":"d7d7c9d7007e34f7ab5b7d4d3f1f6d566067543327e1aba2ce369fd443c8297f","access_level":"public","create_time":"$D2019-07-29T17:46:49.200Z","update_time":"$D2020-03-09T12:06:17.999Z","view_time":null,"first_public_time":"$D2019-08-19T23:59:31.660Z","fork":false,"stars":17,"starred":false,"owner":{"type":"team","id":"5bfcd1f236454eeb","login":"troymagennis","name":"Troy Magennis","avatar_url":"https://avatars-next.observableusercontent.com/bd84c523cb40d1d089b9d8d0a4ac5e025e3593ce8554df5d578d6da6190496c5","tier":"starter_2024"},"version":697,"publish_version":697,"role":null,"authored":false,"authors":[{"type":"individual","id":"f950cac864488237","login":"troymagennis","name":"Troy Magennis","avatar_url":"https://avatars-next.observableusercontent.com/bd84c523cb40d1d089b9d8d0a4ac5e025e3593ce8554df5d578d6da6190496c5","tier":null}],"members":[{"type":"individual","id":"f950cac864488237","login":"troymagennis","name":"Troy Magennis","avatar_url":"https://avatars-next.observableusercontent.com/bd84c523cb40d1d089b9d8d0a4ac5e025e3593ce8554df5d578d6da6190496c5","tier":null,"role":"editor"}],"comments":[],"custom_thumbnail":false,"trashed":false,"subscribed":null,"forks":11,"collection":{"id":"e353f61379db4140","slug":"agile-software-development","title":"Agile Software Development","owner":{"id":"5bfcd1f236454eeb","login":"troymagennis","name":"Troy Magennis"}},"type":"notebook","body":{"title":"Understanding Cycle Time Distribution Shapes","cells":[{"id":0,"value":"$32","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":159,"value":"viewof numDice = slider({\n min: 1, \n max: 15, \n step: 1, \n value: 5,\n title: \"Number of Dice Rolls to Add Together?\", \n description: \"How many standard 6-sided dice rolls to add.\"\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":150,"value":"histogramDice(diceData, numDice)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":186,"value":"$33","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":219,"value":"viewof process = radio({\n title: 'Here is the summary. Please pick your process.',\n description: 'Choose a basic process, but ignore the x-axis (time) for the moment as this will be contextual to you.',\n options: [\n { label: 'Teams with few (or no) external dependencies
', value: 1.0 },\n { label: 'Teams who limit WIP (eg. kanban teams)
', value: 1.25 },\n { label: 'Teams who do batch iterations (eg. Scrum teams)
', value: 1.5 },\n { label: 'Teams who do large batches (waterfall)', value: 2.0 },\n ],\n value: 1.5\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":243,"value":"histogram(processData)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":257,"value":"$34","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":276,"value":"viewof cycleTime = slider({\n min: 1, \n max: 20, \n step: 1, \n value: 10,\n title: \"Work development Time\", \n description: \"How long is the 'doing' the work time? Sure, in the real world you wouldn't know exactly, but work with me\"\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":341,"value":"viewof delayTime = slider({\n min: 1, \n max: 20, \n step: 1, \n value: 2,\n title: \"Delay time if it occurs\", \n description: \"How long is each delay? Sure, in the real world you wouldn't know exactly, but work with me\"\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":286,"value":"viewof delayp1 = slider({\n min: 0, \n max: 100, \n step: 10, \n value: 20,\n title: \"Delay 1 probability %\", \n description: \"0% = no chance, 100% = certainty\"\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":289,"value":"viewof delayp2 = slider({\n min: 0, \n max: 100, \n step: 10, \n value: 20,\n title: \"Delay 2 probability %\"\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":290,"value":"viewof delayp3 = slider({\n min: 0, \n max: 100, \n step: 10, \n value: 20,\n title: \"Delay 3 probability %\"\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":292,"value":"viewof delayp4 = slider({\n min: 0, \n max: 100, \n step: 10, \n value: 20,\n title: \"Delay 4 probability %\"\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":297,"value":"viewof delayp5 = slider({\n min: 0, \n max: 100, \n step: 10, \n value: 20,\n title: \"Delay 5 probability %\"\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":408,"value":"histogramProcessBar(processSimCountData)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":466,"value":"$35","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":531,"value":"md `### Analyzing YOUR data\n\nGiven a set of historical cycle-time samples, we can estimate the critical shape parameter. The ranges of this shape parameter allow specific guidance to be given for both the cases mentioned earlier.Paste your data below (one cycle-time per line) and the shape and scale parameters will be computed. How many samples? Try at least 20 samples but as many as you can for a team still existing doing similar work mix they always have.`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":537,"value":"viewof samples = textarea();","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":617,"value":"histogram(samples.split(\"\\n\").map(Number))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":633,"value":"md `Estimated Sample Shape parameter = ${EstimatedShape}
\nEstimated Sample Scale parameter = ${EstimatedScale}`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":573,"value":"md `**Likely process and advice**\n\nShape range | Likely process | Reducing cycle time advice\n-- | -- | --\n< 1.1 | Teams with few (or no) external dependencies | Automation. Simplifying repetitive tasks\n< 1.3 | Teams who limit WIP (eg. kanban teams) | Reduce development time at constraint. Increase staff size at constraint.\n< 1.7 | Teams who do batch iterations (eg. Scrum teams) | Reduce external team dependencies. \n> 1.7 | Teams who do large batches (waterfall) | Reduce both development time at constraint AND systemic delays (Go Agile!)\n\nI'd love teams who get a shape parameter that doesn't match this table. I've collected hundreds of teams data and so far, so good - BUT I know there is a team not matching this table profile out there. Help me find it. Email me, troy.magennis@focusedobjective.com with questions.\n\n`\n\n \n \n \n ","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":585,"value":"md `### Tracking Trends\n\nIn addition to the shape parameter, the scale parameter can be computed for a set of sample cycle-times. I find an altering trend on either of these two esimates for two sets of sample periods an indictor or system change. Either work type mix has altered (and if not) then a process improvement (or decline) is indicted. \n\nTo generate these trends, segment smples by week, sprint or month (whatever period gives about 7-15 samples) and estimate the shape and scale parameters. Chart these values over time.\n\nEstimated Sample Shape = ${EstimatedShape}
\nEstimated Sample Scale = ${EstimatedScale}`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":527,"value":"$36","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":189,"value":"md`### Appendix - Internal workings of this worksheet. Ignore unless interested`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":9,"value":"d3 = require(\"d3@5\", \"d3-random@2\")","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":11,"value":"ss = require(\"simple-statistics@7\")","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":238,"value":"processData = Array.from({length: 10000}, randomWeibull(process, 5))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":28,"value":"$37","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":614,"value":"AverageTotalProcessEfficiency = (totalSimProcessEfficiency(processSimCountData)*100).toFixed(2)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":649,"value":" binsRef = d3.histogram()\n .domain(x.domain())\n .thresholds(x.ticks(50))\n (sampleEstimatedDistribution);","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":652,"value":" x = d3.scaleLinear()\n .domain(d3.extent(sampleEstimatedDistribution)).nice()\n .range([0,100]); ","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":72,"value":"$38","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":80,"value":"diceData = Int8Array.from({length: 10000}, x => diceAdd(numDice))","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":162,"value":"import {slider} from \"@jashkenas/inputs\"","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":214,"value":"import {select} from \"@jashkenas/inputs\"","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":221,"value":"import {radio} from \"@jashkenas/inputs\"","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":109,"value":"function diceAdd(n)\n{\n var v = 0;\n \n var i;\n \n for (i = 0; i < n; i++) \n v = v + getRandomInt(1,7);\n \n return v\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":305,"value":"function processSim()\n{\n var t = cycleTime;\n \n if (Math.random() <= delayp1/100)\n t = t + delayTime;\n \n if (Math.random() <= delayp2/100)\n t = t + delayTime;\n \n if (Math.random() <= delayp3/100)\n t = t + delayTime;\n \n if (Math.random() <= delayp4/100)\n t = t + delayTime;\n \n if (Math.random() <= delayp5/100)\n t = t + delayTime;\n\n return t\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":421,"value":"simCount = 100000","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":314,"value":"processSimData = Int8Array.from({length: simCount}, x => processSim())","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":400,"value":"processSimCountData = [ \n {value: cycleTime, count: processSimData.filter(x => x == cycleTime).length, efficiency: 1 }, \n {value: cycleTime+1*delayTime, count: processSimData.filter(x => x == cycleTime+1*delayTime).length, efficiency: cycleTime/(cycleTime+1*delayTime) }, \n {value: cycleTime+2*delayTime, count: processSimData.filter(x => x == cycleTime+2*delayTime).length, efficiency: cycleTime/(cycleTime+2*delayTime) },\n {value: cycleTime+3*delayTime, count: processSimData.filter(x => x == cycleTime+3*delayTime).length, efficiency: cycleTime/(cycleTime+3*delayTime) },\n {value: cycleTime+4*delayTime, count: processSimData.filter(x => x == cycleTime+4*delayTime).length, efficiency: cycleTime/(cycleTime+4*delayTime) },\n {value: cycleTime+5*delayTime, count: processSimData.filter(x => x == cycleTime+5*delayTime).length, efficiency: cycleTime/(cycleTime+5*delayTime) }\n]","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":611,"value":"function totalSimProcessEfficiency(data)\n{\n var e = 1;\n \n \n var total = data[0].count;\n total = total + data[1].count;\n total = total + data[2].count;\n total = total + data[3].count;\n total = total + data[4].count;\n\n var runninge = data[0].count * data[0].efficiency;\n runninge = runninge + (data[1].count * data[1].efficiency);\n runninge = runninge + (data[2].count * data[2].efficiency); \n runninge = runninge + (data[3].count * data[3].efficiency);\n runninge = runninge + (data[4].count * data[4].efficiency); \n \n e = runninge / total;\n \n return e;\n}","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":133,"value":"function getRandomInt(min, max) {\n min = Math.ceil(min);\n max = Math.floor(max);\n return Math.floor(Math.random() * (max - min)) + min; //The maximum is exclusive and the minimum is inclusive\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":310,"value":"function getRandom(min, max) {\n\n return Math.random() * (max - min) + min; //The maximum is exclusive and the minimum is inclusive\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":208,"value":"import {randomWeibull} from \"@troymagennis/weibull-distribution\"","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":351,"value":"$39","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":377,"value":"$3a","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":499,"value":"function getShape(p1, p2, p1Result, p2Result)\n{\n return (Math.log(-Math.log(1-p2))-Math.log(-Math.log(1-p1)))/(Math.log(p2Result)-Math.log(p1Result))\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":501,"value":"function getScale(p1, p1Result, shape)\n{\n return p1Result/(Math.pow((-Math.log(1-p1)),(1/shape)));\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":535,"value":"import {textarea} from \"@jashkenas/inputs\"","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":548,"value":"function computeShape()\n{\n var shape = 1.0\n \n var sampleDataArray = samples.split(\"\\n\").map(Number);\n \n shape = getShape(0.3,0.7, ss.quantile(sampleDataArray, 0.3), ss.quantile(sampleDataArray,0.7));\n \n return shape;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":588,"value":"function computeScale()\n{\n var scale = 1.0\n \n var sampleDataArray = samples.split(\"\\n\").map(Number);\n \n scale = getScale(0.3, ss.quantile(sampleDataArray, 0.3), EstimatedShape);\n \n return scale;\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":557,"value":"EstimatedShape = computeShape(samples).toFixed(2)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":592,"value":"EstimatedScale = computeScale().toFixed(2)","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":638,"value":"sampleEstimatedDistribution = Float32Array.from({length: 5000}, randomWeibull(EstimatedShape, EstimatedScale)) ","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":681,"value":"histogram(sampleEstimatedDistribution)","pinned":true,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"}],"files":"$W3b","stdlib":"1","titling":"auto"}} 31:"$30:body"