Forget those instant computer predictions. Throw out those political analysts whose tales are already stale by the 11 p.m. news on election Tuesday. You're in Cambridge, and election night isn't over, not by a long shot. Here, you've got nearly a week after elections to figure out who's winning and who's losing.

The reason, of course, is Proportional Representation-that intricate (some would say Byzantine) electoral system which only Cambridge, alone of U.S. cities, still uses. Some 25 cities used it earlier in this century. They adopted it so minorities would have a better chance of election; they generally got rid of it when unpopular minorities got elected. Now, only in Cambridge can you savor the weeklong PR count.

It'll be going on again today, perhaps far into the night, at the Longfellow School auditorium-a 15-minute walk down Broadway from Harvard. Depending on the hour of day, anywhere from 150 to better than 300 people will be jammed into that auditorium, doing and watching the PR count.

The count, no matter how well organized it may be, always strikes a newcomer as something like an especially chaotic county fair: children run about, ladies gossip, politicians caucus, and loudspeakers blare.

But most of the people there know what they're doing; at least, they know why they're there: to find out which nine city councillors and which six school committeemen the City's voters elected last Tuesday. With luck, more than a little of it, they'll finish the job tonight, and go home after a count which has lasted four days and cost the city some $6500-probably not much more than it would cost under other electoral systems, since PR climinates the need for primaries and special elections. But if they don't get through the last rounds of the count tonight, counters, candidates, and "watchers" will be at it again next week.

To understand what's going on at the count, it's helpful to look first at what a citizen does when he votes under PR in Cambridge. Two long paper ballots-this year, they're blue for City Council, pink for School Committee-are handed to him at the polls. On the ballots are listed the names of the candidates. Instead of marking an "X" beside his choices, the voter ranks his preferences (1, 2, 3, etc.). In theory, he could indicate his 26th choice for a council seat, but most voters confine themselves to a maximum of four or five choices.

After the polls close, the ballots are locked up, and then brought next morning to the school auditorium. There, just before 8 a.m. the 100-odd election clerks begin a first, "unofficial" count of the "number one" votes for each city council candidate. The City has eleven wards, each with five precincts-so there are eleven little cubicles on one side of the auditorium, to allow the first precinct of each ward to be counted simultaneously, etc.

As the count of "number ones" continues throughout the day, candidates and their pollwatchers hang outside each little cubicle, keeping a running count much as they do under the usual election systems. Around five p.m., the count is over. and candidates, media, and hangers-on crowd to get the official announcement of this "unofficial" count. This "unofficial" count is merely a device to-speed up the "official" first count the next day, when the pre-sorted ballots are actually stamped for one candidate or another.

With luck, by the middle of Friday, the Election Commission will have an "official" count of number one ballots for the City Council. Then, the real fun of PR begins.

Under the system, each candidate must have a certain quota of votes in order to be elected. The quota is set by dividing the total number of votes cast by the number of seats to be filled plus one. Then, add one vote to the quotient and there's the quota. The idea behind sitting the quota at this level is to make it low enough for candidates supported by a minority group to get elected, and high enough to make the elections more than a matter of chance.

This year, about 25,000 voted in the City Council election, making the quota approximately 2500. As he almost always does, Walter J. Sullivan ran highest; he was the only candidate to make the quota from his "number ones." As an example, say Sullivan got a hypothetical total of 3750 "number ones" (slightly higher than his actual count). Thus, he is 1250 over the quota.

In the next step of the count, the 1250 surplus Sullivan votes will be redistributed by taking every third Sullivan vote from the envelopes in which they are kept, by precinct, after the first count, Each of the 1250 votes is then put into the pile of whatever candidate is marked "number two" on the ballot.

The redistribution of the "surplus" of candidates who make the quota is one half of PR's reshuffling of votes; the other half is the elimination of the candidates ranking lowest at any one time, and the redistribution of their votes to "number two" candidates.

The following shows what this redistribution can mean for one voter's ballot:

It's an unlikely combination, but let's say you voted "number one" for Sullivan, "number two" for Peace and Freedom Party candidate Cynthia F. Kline, and "number three" for incumbent councillor Barbara Ackermann. Thus, your vote would be counted for Sullivan in the first round and, since he met the quota in that round, perhaps drawn and placed in Miss Kline's pile. Since she was a weak candidate (pulling only some 250 "number ones") she would soon be eliminated. Your ballot would then go to Mrs. Ackermann's pile, and rest there, helping her to meet the quota.