The votes may have been reversed because the user was determined to have been engaging in voting fraud.

According to the latest documentation on how the decision is made to preserve votes for deleted users, reputation is not checked when the system decides to field a voluntary deletion request to employees for potential vote preservation, only the user's votes are:

• There are two thresholds:
  1. Number of votes cast by the user being deleted
  2. Number of people affected significantly by those votes

The exact values of those thresholds don't particularly matter; they're pretty low, but not so low that you could hit them easily while still hiding fraudulent activity.

• If either threshold is exceeded, deletion is held up until someone reviews it. Otherwise, deletion proceeds and any votes are discarded.
• If, during review, it becomes apparent that the user is or ever was involved in voting fraud, the votes are discarded as they would normally be.
• Otherwise, the votes are preserved.

Based on the posts I've seen (and the fact that I've also been affected by the same user's deletion), it seems that both the numbers mentioned in the first bullet are high, enough to cross the "low" thresholds, thus fielding the account to system administrators for potential vote preservation.

The third bullet states that even if the user would qualify for vote preservation, their votes can still be invalidated if it is determined that the user has been engaging in voting fraud. Based on the public evidence, this is one thing that may have happened here.

It's also possible that the user wasn't deleted voluntarily, but by a Stack Exchange employee (as it was network-wide); note that if a moderator or employee deletes a user, their votes are always invalidated, and vote preservation only comes into play if the user voluntarily requests deletion.

A couple clarifications:

• Note that while the text above does imply that users who have engaged in voting fraud are forever precluded from having their votes preserved if they later request deletion, this is in fact not the case: users who engage in small serial voting in their early years then later go on to become constructive users can still potentially have their votes preserved if they later request deletion.

• The idea that reputation is the main criterion that the system uses to potentially have votes preserved can be traced back to this (speculative?) edit to an FAQ, which just got copied as-is into the help center page. The help center page's statement that votes are preserved if the user "has a very high reputation score" is incorrect, per the above quoted answer and this recent clarification. (Update: The help page has since been edited to correctly state the criteria for kicking deletions into staff review.)

The votes may have been reversed because the user was determined to have been engaging in voting fraud.

According to the latest documentation on how the decision is made to preserve votes for deleted users, reputation is not checked when the system decides to field a voluntary deletion request to employees for potential vote preservation, only the user's votes are:

• There are two thresholds:

  1. Number of votes cast by the user being deleted
  2. Number of people affected significantly by those votes

  The exact values of those thresholds don't particularly matter; they're pretty low, but not so low that you could hit them easily while still hiding fraudulent activity.

• If either threshold is exceeded, deletion is held up until someone reviews it. Otherwise, deletion proceeds and any votes are discarded.

• If, during review, it becomes apparent that the user is or ever was involved in voting fraud, the votes are discarded as they would normally be.

• Otherwise, the votes are preserved.

Based on the posts I've seen (and the fact that I've also been affected by the same user's deletion), it seems that both the numbers mentioned in the first bullet are high, enough to cross the "low" thresholds, thus fielding the account to system administrators for potential vote preservation.

The third bullet states that even if the user would qualify for vote preservation, their votes can still be invalidated if it is determined that the user has been engaging in voting fraud. Based on the public evidence, this is one thing that may have happened here.

It's also possible that the user wasn't deleted voluntarily, but by a Stack Exchange employee (as it was network-wide); note that if a moderator or employee deletes a user, their votes are always invalidated, and vote preservation only comes into play if the user voluntarily requests deletion.

A couple clarifications:

• Note that while the text above does imply that users who have engaged in voting fraud are forever precluded from having their votes preserved if they later request deletion, this is in fact not the case: users who engage in small serial voting in their early years then later go on to become constructive users can still potentially have their votes preserved if they later request deletion.

• The idea that reputation is the main criterion that the system uses to potentially have votes preserved can be traced back to this (speculative?) edit to an FAQ, which just got copied as-is into the help center page. The help center page's statement that votes are preserved if the user "has a very high reputation score" is incorrect, per the above quoted answer and this recent clarification. (Update: The help page has since been edited to correctly state the criteria for kicking deletions into staff review.)

The votes may have been reversed because the user was determined to have been engaging in voting fraud.

According to the latest documentation on how the decision is made to preserve votes for deleted users, reputation is not checked when the system decides to field a voluntary deletion request to employees for potential vote preservation, only the user's votes are:

• There are two thresholds:
  1. Number of votes cast by the user being deleted
  2. Number of people affected significantly by those votes

The exact values of those thresholds don't particularly matter; they're pretty low, but not so low that you could hit them easily while still hiding fraudulent activity.

• If either threshold is exceeded, deletion is held up until someone reviews it. Otherwise, deletion proceeds and any votes are discarded.
• If, during review, it becomes apparent that the user is or ever was involved in voting fraud, the votes are discarded as they would normally be.
• Otherwise, the votes are preserved.

Based on the posts I've seen (and the fact that I've also been affected by the same user's deletion), it seems that both the numbers mentioned in the first bullet are high, enough to cross the "low" thresholds, thus fielding the account to system administrators for potential vote preservation.

The third bullet states that even if the user would qualify for vote preservation, their votes can still be invalidated if it is determined that the user has been engaging in voting fraud. Based on the public evidence, this is one thing that may have happened here.

It's also possible that the user wasn't deleted voluntarily, but by a Stack Exchange employee (as it was network-wide); note that if a moderator or employee deletes a user, their votes are always invalidated, and vote preservation only comes into play if the user voluntarily requests deletion.

A couple clarifications:

• Note that while the text above does imply that users who have engaged in voting fraud are forever precluded from having their votes preserved if they later request deletion, this is in fact not the case: users who engage in small serial voting in their early years then later go on to become constructive users can still potentially have their votes preserved if they later request deletion.

• The idea that reputation is the main criterion that the system uses to potentially have votes preserved can be traced back to this (speculative?) edit to an FAQ, which just got copied as-is into the help center page. The help center page's statement that votes are preserved if the user "has a very high reputation score" is incorrect, per the above quoted answer and this recent clarification.

The votes may have been reversed because the user was determined to have been engaging in voting fraud.

According to the latest documentation on how the decision is made to preserve votes for deleted users, reputation is not checked when the system decides to field a voluntary deletion request to employees for potential vote preservation, only the user's votes are:

• There are two thresholds:
  1. Number of votes cast by the user being deleted
  2. Number of people affected significantly by those votes

The exact values of those thresholds don't particularly matter; they're pretty low, but not so low that you could hit them easily while still hiding fraudulent activity.

• If either threshold is exceeded, deletion is held up until someone reviews it. Otherwise, deletion proceeds and any votes are discarded.
• If, during review, it becomes apparent that the user is or ever was involved in voting fraud, the votes are discarded as they would normally be.
• Otherwise, the votes are preserved.

Based on the posts I've seen (and the fact that I've also been affected by the same user's deletion), it seems that both the numbers mentioned in the first bullet are high, enough to cross the "low" thresholds, thus fielding the account to system administrators for potential vote preservation.

The third bullet states that even if the user would qualify for vote preservation, their votes can still be invalidated if it is determined that the user has been engaging in voting fraud. Based on the public evidence, this is one thing that may have happened here.

It's also possible that the user wasn't deleted voluntarily, but by a Stack Exchange employee (as it was network-wide); note that if a moderator or employee deletes a user, their votes are always invalidated, and vote preservation only comes into play if the user voluntarily requests deletion.

A couple clarifications:

• Note that while the text above does imply that users who have engaged in voting fraud are forever precluded from having their votes preserved if they later request deletion, this is in fact not the case: users who engage in small serial voting in their early years then later go on to become constructive users can still potentially have their votes preserved if they later request deletion.

• The idea that reputation is the main criterion that the system uses to potentially have votes preserved can be traced back to this (speculative?) edit to an FAQ, which just got copied as-is into the help center page. The help center page's statement that votes are preserved if the user "has a very high reputation score" is incorrect, per the above quoted answer and this recent clarification. (Update: The help page has since been edited to correctly state the criteria for kicking deletions into staff review.)