Notebook · not a matching file
Competitor rates should not be treated as a nightly file to copy. A rate change may show inventory pressure, strategy, or a change in demand. The useful timing is the next customer search or walk-up rental, not only the next batch. RateIndex provides a measured view of the public market. This notebook asks what that view could report if each rate change were treated as an event with context.
Evidence, not an instruction
RateMonitor already collects rates, compares them, and applies corrections inside the operator’s strategy. Copying a competitor is only one possible rule, and the platform can reject it through floors, ceilings, utilization limits, and length-of-rent plans. The platform acts at market speed. The operator sets the guardrails.
Event-based analysis asks when the competitor changed the rate, how often, for which vehicle class and rental length, and which competitors moved together. A lower one-day rate three weeks before pickup is different from a last-minute change for walk-up demand. The goal is to estimate possible intent while still producing evidence, not a price to copy.
Market speed does not mean shopping every location more often. Coverage has a cost. Some locations should follow a related market or parent location. The research question is whether a meaningful competitor event can update the operator’s view before the next search, while the operator remains free to reject the signal.
Still does not price
Observation must produce evidence before action. The operator decides what the evidence means, defines strategy and guardrails, and can reject any recommendation. A system that must follow a competitor is not providing useful awareness.
The analysis must connect each event to a specific location, vehicle class, pickup date, and rental length. An estimate of intent is useful only when it can change a real decision. RateMonitor remains the execution platform; this notebook studies the evidence that could inform it.
A direction, not a paper
The mathematics we are exploring
Model competitor rate changes as events with timing and attributes. Estimate a possible competitor strategy from the event history. The result is evidence the operator can accept or reject; it is not a direct instruction to change a price.
N(dt × dy)
The timing and attributes of each rate change. A nightly file of rate levels loses the sequence of events.
θ
A possible competitor strategy that cannot be observed directly. It should not be treated as a known fact.
a = α(p(θ | ·))
Advice based on the estimated strategy. The operator decides what to do; RateMonitor applies the approved decision.
This is a research notebook. It is not an announced product or release. The operator still decides.