Abstract
This study examines how often final recorded bids in Swedish property auction traces make unusually large increments relative to earlier bids. A final step is classified as an overbid under this report’s definition when it is more than twice the mean of the trace’s other positive recorded increments.
11.6% of eligible traces (61 of 524) had a final recorded increase more than twice as large as the mean of their earlier positive increases.
1. Dataset overview
The aggregate snapshot covers 892 stored auction traces and 7838 recorded bid rows from Notar.se. The available provenance establishes the stored trace population, not a complete universe of finished auctions.
- Stored traces
- 892
- Stored bid rows
- 7838
- Avg. distinct bidder labels
- 2.73
- Avg. recorded price change
- 6.95%
| Metric | Value |
|---|---|
| Average starting price | 2927775 SEK |
| Average final recorded price | 3100000 SEK |
| Average starting-to-final change | 172225 SEK (6.95%) |
| Mean earlier positive increment in eligible traces | 25132.89 SEK (n = 6005; final increments excluded) |
| Calculation contract | notar-final-step-v1 |
| Inclusive source cutoff | 2019-10-19 21:15:00 |
Note: Starting price is the percentage denominator.
Source: Reviewed aggregate snapshot reproduced from the
versioned Asight calculation.
2. Method
2.1 Eligibility and final-step ratio
One stored object_id is treated as one auction trace. The rules are applied
in order: a trace must have exactly one final-marked row; that row’s offer must
equal the trace maximum; its increment must be positive; and the trace must
contain at least one other positive increment.
Of 892 stored auction traces, 524 met all four conditions. The remaining traces are reconciled by mutually exclusive exclusion reasons in Chapter 5.
For each eligible trace, the final-step ratio is:
The final increment is excluded from its own baseline. A ratio above 2× is a deliberately simple descriptive screen: it is the operational definition of overbidding in this report, not a statistical outlier test or evidence of economic overpayment. Using the mean also makes the screen more conservative in irregular traces: a large earlier increase raises the baseline and makes the above-2× classification harder to trigger. Chapter 5 reports the corresponding median-baseline result.
2.2 Repeated offers
Repeated-offer rows are retained as separate recorded events so the published calculation reproduces the reviewed historical result. The snapshot also tests keeping only the earliest stored row for each offer under the versioned source ordering. That alternative reduces the comparison increments from 6005 to 6000 but changes 0 above-2× classifications: the headline count remains 61.
3. General bidding patterns
Figure 1
Price change and bidder participation across stored traces
A. Starting-to-final price change
- Below 0% 13.1% (117)
- No change 16.4% (146)
- Above 0% to below 5% 21.3% (190)
- 5% to below 10% 16.0% (143)
- 10% to below 15% 13.8% (123)
- 15% to below 20% 8.4% (75)
- 20% or more 11.0% (98)
B. Distinct positive bidder labels
- 1 31.2% (278)
- 2 23.9% (213)
- 3 18.4% (164)
- 4 11.1% (99)
- 5 7.1% (63)
- 6 3.9% (35)
- 7 2.2% (20)
- 8+ 2.2% (20)
Share of stored traces (%)
Two views of the same 892 stored traces: starting-to-final recorded price change and distinct positive recorded bidder-label count.
Note: Both panels use the same 0–35% scale. Counts are shown in parentheses. The panels describe separate attributes; their category rows are not paired.
Source: Reviewed aggregate snapshot reproduced from the versioned Asight calculation.
4. Final-step findings
The final-step calculation applies only to the 524 traces that meet every eligibility rule.
- Eligible traces
- 524
- Above 2×
- 61
- Above-2× share
- 11.6%
- Median ratio
- 0.7692×
Figure 2
Final increment relative to earlier increments
Distribution of final marked increment ratios relative to the mean of earlier positive recorded increments
Final-step ratio
Final marked increment divided by the mean of other positive increments.
Note: The last three bins total 61 traces above the 2× threshold.
Source: Reviewed aggregate snapshot reproduced from the versioned Asight calculation.
Average starting-to-final percentage change by distinct positive recorded bidder-label count.
| Distinct bidder labels | Stored traces | Average price change |
|---|---|---|
| 1 | 278 | -0.62% |
| 2 | 213 | +4.63% |
| 3 | 164 | +9.9% |
| 4 | 99 | +11.92% |
| 5 | 63 | +15.3% |
| 6 | 35 | +19.12% |
| 7 | 20 | +21.18% |
| 8+ | 20 | +26.16% |
All 892 stored traces are included;
starting price is the percentage denominator. Higher recorded
bidder-label counts coincided with larger average recorded price
changes. This is a descriptive association; the group sizes differ and
the pattern does not establish a causal effect.
Source: Reviewed aggregate snapshot reproduced from the
versioned Asight calculation.
5. Aggregate checks
This chapter replaces the former auction-level explorer with aggregate-only sample-flow and sensitivity evidence generated by the same versioned calculation as the rest of the report.
The ordered flow reconciles every stored trace: 892 = 4 + 12 + 252 + 100 + 524.
| Outcome | Traces |
|---|---|
| Multiple final markers | 4 |
| Final marker below trace maximum | 12 |
| Nonpositive final increment | 252 |
| No earlier positive increment | 100 |
| Eligible | 524 |
| Total stored traces | 892 |
Outcomes are mutually exclusive and applied in rule order.
Source: Reviewed aggregate snapshot reproduced from the
versioned Asight calculation.
| Threshold | Mean baseline | Median baseline |
|---|---|---|
| 1.5× | 93 (17.7%) | 125 (23.9%) |
| 2.0× | 61 (11.6%) | 78 (14.9%) |
| 3.0× | 35 (6.7%) | 46 (8.8%) |
Values show eligible traces classified above each threshold.
Source: Reviewed aggregate snapshot reproduced from the
versioned Asight calculation.
| Earlier positive increments | Eligible traces | Above 2× |
|---|---|---|
| 1 | 46 | 9 (19.6%) |
| 2-4 | 91 | 12 (13.2%) |
| 5-9 | 137 | 12 (8.8%) |
| 10+ | 250 | 28 (11.2%) |
Values show the mean-baseline result by comparison-history length.
Source: Reviewed aggregate snapshot reproduced from the
versioned Asight calculation.
6. Conclusions
Bidder Participation and Price Change
Higher recorded bidder counts coincided with larger average starting-to-final price changes. This relationship is observational, not causal.
Overbidding
Under the report’s definition, 11.6% of eligible traces—61 of 524—were classified as overbids because their final recorded increase exceeded 2× the mean of their earlier positive increases.
Limitations and further research
The 11.6% result describes only the 524 traces eligible for this comparison, not all 892 stored traces. The analysis cannot show whether another bidder would have continued, whether the winner paid more than the property was worth, or whether bidder participation caused the observed price changes. The 2× boundary is analyst defined, and the sensitivity tables show how the result changes with other thresholds, baselines, and comparison-history lengths.