SMART Value for ECC Recovery is both increasing and decreasing - what does this mean?

Question

I'm seeing the counter for Hardware_ECC_Recovered both increasing, and decreasing in the SMART stats for a hdd and I'm confused.

I'd expect it'd increase each time a error is detected (and corrected) by ECC, but why on earth would it ever decrease?

/var/log/messages-20191208.gz:Dec  4 20:40:58 MushaV3 smartd[4411]: Device: /dev/sdb [SAT], SMART Usage Attribute: 195 Hardware_ECC_Recovered changed from 23 to 24
/var/log/messages-20191208.gz:Dec  7 16:40:58 MushaV3 smartd[4411]: Device: /dev/sdb [SAT], SMART Usage Attribute: 195 Hardware_ECC_Recovered changed from 24 to 23
/var/log/messages-20191215.gz:Dec 10 03:40:58 MushaV3 smartd[4411]: Device: /dev/sdb [SAT], SMART Usage Attribute: 195 Hardware_ECC_Recovered changed from 23 to 24
/var/log/messages-20191215.gz:Dec 11 13:40:59 MushaV3 smartd[4411]: Device: /dev/sdb [SAT], SMART Usage Attribute: 195 Hardware_ECC_Recovered changed from 24 to 25
/var/log/messages-20191215.gz:Dec 11 20:40:58 MushaV3 smartd[4411]: Device: /dev/sdb [SAT], SMART Usage Attribute: 195 Hardware_ECC_Recovered changed from 25 to 24
/var/log/messages-20191222.gz:Dec 16 04:10:58 MushaV3 smartd[4411]: Device: /dev/sdb [SAT], SMART Usage Attribute: 195 Hardware_ECC_Recovered changed from 24 to 23
/var/log/messages-20191222.gz:Dec 16 04:40:58 MushaV3 smartd[4411]: Device: /dev/sdb [SAT], SMART Usage Attribute: 195 Hardware_ECC_Recovered changed from 23 to 24
/var/log/messages-20191222.gz:Dec 21 21:40:58 MushaV3 smartd[4411]: Device: /dev/sdb [SAT], SMART Usage Attribute: 195 Hardware_ECC_Recovered changed from 24 to 25

Full Smart output for this drive: (I'm planning to replace anyway).

MushaV3 /home/tim/server/media/music # smartctl -a /dev/sdb
smartctl 7.0 2018-12-30 r4883 [x86_64-linux-4.19.57-gentoo] (local build)
Copyright (C) 2002-18, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
Model Family:     Seagate Barracuda 7200.12
Device Model:     ST3500413AS
Serial Number:    Z2A8WLS9
LU WWN Device Id: 5 000c50 035a9b8b6
Firmware Version: JC4B
User Capacity:    500,107,862,016 bytes [500 GB]
Sector Size:      512 bytes logical/physical
Rotation Rate:    7200 rpm
Device is:        In smartctl database [for details use: -P show]
ATA Version is:   ATA8-ACS T13/1699-D revision 4
SATA Version is:  SATA 3.0, 6.0 Gb/s (current: 6.0 Gb/s)
Local Time is:    Sun Dec 22 15:15:54 2019 GMT
SMART support is: Available - device has SMART capability.
SMART support is: Enabled

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

General SMART Values:
Offline data collection status:  (0x82) Offline data collection activity
                    was completed without error.
                    Auto Offline Data Collection: Enabled.
Self-test execution status:      (   0) The previous self-test routine completed
                    without error or no self-test has ever 
                    been run.
Total time to complete Offline 
data collection:        (  600) seconds.
Offline data collection
capabilities:            (0x7b) SMART execute Offline immediate.
                    Auto Offline data collection on/off support.
                    Suspend Offline collection upon new
                    command.
                    Offline surface scan supported.
                    Self-test supported.
                    Conveyance Self-test supported.
                    Selective Self-test supported.
SMART capabilities:            (0x0003) Saves SMART data before entering
                    power-saving mode.
                    Supports SMART auto save timer.
Error logging capability:        (0x01) Error logging supported.
                    General Purpose Logging supported.
Short self-test routine 
recommended polling time:    (   1) minutes.
Extended self-test routine
recommended polling time:    (  83) minutes.
Conveyance self-test routine
recommended polling time:    (   2) minutes.
SCT capabilities:          (0x103f) SCT Status supported.
                    SCT Error Recovery Control supported.
                    SCT Feature Control supported.
                    SCT Data Table supported.

SMART Attributes Data Structure revision number: 10
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x000f   115   099   006    Pre-fail  Always       -       90061503
  3 Spin_Up_Time            0x0003   100   100   000    Pre-fail  Always       -       0
  4 Start_Stop_Count        0x0032   100   100   020    Old_age   Always       -       850
  5 Reallocated_Sector_Ct   0x0033   100   100   036    Pre-fail  Always       -       0
  7 Seek_Error_Rate         0x000f   075   060   030    Pre-fail  Always       -       38654999
  9 Power_On_Hours          0x0032   028   028   000    Old_age   Always       -       63471
 10 Spin_Retry_Count        0x0013   100   100   097    Pre-fail  Always       -       0
 12 Power_Cycle_Count       0x0032   100   100   020    Old_age   Always       -       849
183 Runtime_Bad_Block       0x0032   094   094   000    Old_age   Always       -       6
184 End-to-End_Error        0x0032   100   100   099    Old_age   Always       -       0
187 Reported_Uncorrect      0x0032   100   100   000    Old_age   Always       -       0
188 Command_Timeout         0x0032   100   098   000    Old_age   Always       -       3
189 High_Fly_Writes         0x003a   099   099   000    Old_age   Always       -       1
190 Airflow_Temperature_Cel 0x0022   076   053   045    Old_age   Always       -       24 (Min/Max 16/24)
194 Temperature_Celsius     0x0022   024   047   000    Old_age   Always       -       24 (0 12 0 0 0)
195 Hardware_ECC_Recovered  0x001a   025   015   000    Old_age   Always       -       90061503
197 Current_Pending_Sector  0x0012   100   100   000    Old_age   Always       -       0
198 Offline_Uncorrectable   0x0010   100   100   000    Old_age   Offline      -       0
199 UDMA_CRC_Error_Count    0x003e   200   200   000    Old_age   Always       -       0
240 Head_Flying_Hours       0x0000   100   253   000    Old_age   Offline      -       65743 (213 204 0)
241 Total_LBAs_Written      0x0000   100   253   000    Old_age   Offline      -       1432413407
242 Total_LBAs_Read         0x0000   100   253   000    Old_age   Offline      -       1198286283

SMART Error Log Version: 1
No Errors Logged

SMART Self-test log structure revision number 1
Num  Test_Description    Status                  Remaining  LifeTime(hours)  LBA_of_first_error
# 1  Extended offline    Completed without error       00%     51176         -
# 2  Extended offline    Completed without error       00%      1753         -
# 3  Extended offline    Completed without error       00%       111         -

SMART Selective self-test log data structure revision number 1
 SPAN  MIN_LBA  MAX_LBA  CURRENT_TEST_STATUS
    1        0        0  Not_testing
    2        0        0  Not_testing
    3        0        0  Not_testing
    4        0        0  Not_testing
    5        0        0  Not_testing
Selective self-test flags (0x0):
  After scanning selected spans, do NOT read-scan remainder of disk.
If Selective self-test is pending on power-up, resume after 0 minute delay.

Answer 1

According to general sources such as Acronis and smartmontools the normalised value of this attribute is based on the frequency of ECC errors rather than just the total number. So it will tend to go up and down over time and based on SMART conventions higher is better.

In reality it is not easy to decode the raw data as the format varies by manufacturer and they do not always disclose details. However, I found this article which speculates about Seagate’s implementation based on what is known

Answer 2

It is perfectly normal for the Hardware_ECC_recovered value to go up and down. This is one of several "rate" parameters, that characterize how often has some (prossibly worrysome) event happened recently. In this case, the drive measures the fraction of sectors that needed ECC correction applied after reading. Because the drive purposefully forgets about history (to give you a value about the reading performance "right now"), it is normal for the value to move up and down. The technical background of this parameter fluctuating in many drives is that hard drives try reading a sector before being certain that the head is perfectly aligned with the track. If it can read the sectors anyway, the user gets better performance. If it fails to obtain valid data, it retries on the next revolution. If the head is positioned " nearly right", it is quite likely that ECC is able to correct the bits that were not correctly read. In my experience, the ECC recovered rate gets worse on random accesses (like indexing all your files) and improves on sequentially reading large files (like copying DVD imagebor watching movies). Finally, I want to point out that the hard drive manufacturer also tells you to not worry: The threshold to be worried is 0, and your value is around 25, with 15 being the worst one ever seen. Lower values are worse, so you are still in the "green" range. And even if that value drops to zero, it is an indicator that the drive degraded over time ("old age") (so it might be wise to replace the drive when convenient), and not a warning of imminent total failure of that drive ("pre(dictable) failure").