My comments on Gore's legal strategy
We need a fair count
The action is under (3)(c) ... rejection of a number of legal votes sufficient to change
or place in doubt the result of the election." Specifically, there
are a large number of legal votes that have never been counted once that
we MUST count in order to understand voter intent.
What was shown:
- This election is a statistical dead heat. The vote count is less
than the margin of error of the tabulating equipment right now (.1%
typically under ideal conditions). And we know that! Nassau alone was 200
votes on re-running the same cards! We don't know which count was right!
Right now, the margin is so close that we are, in effect, flipping a coin to
determine the will of the people. We must find out the true will of the
- We did NOT have a fair count because the VOTING MACHINES were
- We showed that Florida uses two types of voting systems: optical and
- We showed that the undervote rate is significantly higher in
punch-card sytems than optical systems
- In effect, a significant percentage of the voters were voting
with a visual handicap. In an punch card system, you can't see
if you overvote or undervote. In an optical system, you can
see exactly how you voted. That's why there is a 5X difference in
undervotes. Our intuition is confirmed by the statistics.
- Evidence that the outcome of the election could be placed in doubt
- Punch card systems are primarily used in democratic counties.
The uncounted votes here are mostly from democratic counties.
The large number of uncounted votes from democratic counties can shift
the outcome of this election.
- There are other actions affecting the vote count. In an election this
close, if our argument is reasonable, it should be granted. We can
project (using techniques just like the Miami Herald study showed) that
if all the votes are counted, Gore picks up a convincing number of votes
compared to Bush
- Due to voter error, operator error, machine error, fraud, tampering,
butterfly ballots, etc., the vote count in an election is just a
statistical approximation of the true will of the people. It is the
latter that matters.
- To determine who really won, we need to equalize out as many variables
as we can especially to resolve a close contest.
- We must have a fair count. We need to see who really won.
The general rule you must apply in a close election with mixed voting machine
types with vastly different characteristics:
- In any CLOSE election where different voting machinery with
significantly DIFFERENT undervoting statistics is used to place a
vote, we MUST do a manual recount to ATTEMPT to EQUALIZE counting accuracy
to COMPENSATE for the different number of undervotes.
- If we don't do that in Florida, we give an advantage to the counties with
optical systems. And those counties are primarily Republican. That's not
fair. That's going into an election where the voting machines are biased
- So if you have machine A which undervotes 5 times as much as machine B,
you need to work 5 times harder to "read" the results of machine A
to have a fair count. In our case, we can partially do that by applying the
Broward counting standards to all punchcard ballots which recovers only 25%
of the data (we needed to recover 80% to be even with the optical cards).
- We don't handcount the uncounted optical cards too because we are trying
to close the gap, not widen it. Even with handcounting (with the Broward
standard) all the uncounted punch cards, we still don't close the gap, but
we get close.
The remedy you should ask for is simple and fast:
- We want a fair count where ALL voters in Florida are treated
equally, regardless of party.
- We'd like to fully level the playing field. We'd like the
statistics adjusted as if everyone used the same voting equipment.
- We can get there statistically, by statistically counting those
legal votes that have never been counted once. Here is the result of
that.... (show study if we counted 80% of the undervote in counties with
punch card systems using the statistics of the county). This can show that
we overturn the election significantly (so we meet (3)(c)).
- If we must count ballots by hand, then as mentioned above, we can only
partially equalize the effect of the voting machines since the best
handcount policies recover only 25% of the undervote instead of the 80% we'd
need to "be even" with optical machines.
- We propose we count for the first time ALL uncounted punch card ballots
in the state of Florida from ALL counties which used a voting mechanism
that has significantly higher undervote statistics than the baseline system
(optical). In this case, we'd hand count all uncounted ballots
counties using punch-card systems.
- The process is the following:
- Collect all previously uncounted punch card legal ballots from all
counties that have not been included in the certified count
- Mix them up randomly
- Draw 10 groups of 1000 cards
- Assign each group of 1000 to 10 independent committees of special
masters to hand count these cards using the Broward (Texas) standard.
This is the standard that recovers maximum voter intent (25%), but
still does not get to the fair recovery rate (80%).
- Linearly extrapolate those results to all uncounted ballots
- NOTE: By having 10 independent groups, we have several advantages:
- We can verify that the counts reported by each group are the same
(giving us assurance that each group counted correctly)
- By looking at the standard deviation of the counts, we can
determine a margin of error for the extrapolated count
- We can count the ballots faster since the groups work in parallel
and have only 1000 ballots each
- If the margin of victory is less than the standard deviation
(error), we repeat the count more ballots until the margin
of victory is substantially higher than the margin of error.
David Boies did not pursue this strategy. Under the law, he's justified in
his current strategy of seeking selective recounts. The reason is simple: the
legal system itself is unfair. Sauls shouldn't have ruled against him.
Miami-Dade should have figured out a way to add more people to count faster,
etc. The butterfly ballot case was tossed due to poor legal representation, not
because of the merits (the lawyers argued for a re-vote and invalidation of the
election, rather than asking solely for a re-allocation).