Rare events such as safety incidents or major defects are mostly zero, so conventional P/C charts drown in zeroes. The G chart monitors the time (or batch) interval between events, which follows a geometric distribution; the T chart applies the power transform t^0.2777 to the intervals to approximate normality so conventional control limits can be used. The exponent 0.2777 is empirical (about 1/3.6) and is applied automatically by the tool.
Use it for extremely low defect rates dominated by zeros: safety incidents, injuries, major equipment failures and rare quality defects. Record the interval between events rather than only counts — a shorter interval means events are clustering. Keep the event definition consistent to avoid definition drift.
Enter the event dates or sequence intervals, and the tool applies the t^0.2777 transform and plots I-MR-type limits, marking violations. An interval below the lower control limit means event frequency is rising — investigate immediately; lengthening intervals mean the improvement is working.
T = t^0.2777; control limits are computed from the mean and moving range of the transformed values. Accumulate at least 15–20 event intervals before fixing the initial limits, and recompute the baseline whenever the event rate changes.