EastSouthWestNorth: The Hong Kong 7/1 March: Crowd Size Estimates (Civil Human Rights Front / Human Rights Monitor)

The Hong Kong 7/1 March: Crowd Size Estimates (Civil Human Rights Front / Human Rights Monitor)

SOURCE #1: CIVIL HUMAN RIGHTS FRONT says 530,000

SOURCE #5: HUMAN RIGHTS MONITOR says 160,000

According to Yahoo! News, the march organizers have put in a claim of 530,000 people today, which is larger than the 500,000 last year. There were several conflicting figures from the organizers through the evening, with 350,000 at 5:00pm, 400,000 at 6:40pm, 450,000 at 6:55pm but I am going with the latest number of 530,000 at 9:20pm. Incidentally, the different numbers at different times is the reason why foreign news sources are reporting different totals depending on their deadlines.

Civil Human Rights Front spokesperson Jackie Hung Ling-yu said that the front of the march departed at 2:30pm from Victoria, half a hour before the scheduled time, and took 90 minutes to reach the Central Government Office in Central. Based upon government information, the road between Victoria Park to the Central Government Office could hold 170,000 persons. The end of the march reached the Main Government Building at 8pm. Therefore, they inferred that there must have been 530,000 persons present today.

How does this calculation work? This is not rocket science. They give this type of problem to primary school kids. Let me stipulate that I will accept all the assumptions about travel time and road capacity because I have no reason to doubt them (and no reason to accept them, either!). So here is how I did it.

At 2:30pm, the front of the march departed from Victoria Park and arrived at the destination 90 minutes later at 4:00pm. At that point, the route is filled to the brim with 170,000 people. Good. So the running total is 170,000 people.

At 4:00pm, the next group of people departed from Victoria Park and arrived at the destination 90 minutes later at 5:30pm. At that point, the route is filled to the brim with another 170,000 people. Good. Now the running total is 170,000 + 170,000 = 340,000 people so far.

At 5:30pm, the next group of people departed from Victoria Park and they kept leaving until 6:30pm. The latest time for departure is 6:30pm because they needed 90 minutes to finish at 8:00pm. Thus, the last group of people all left within one hour, which means that the road was only two-thirds full. This means 170,000 x 2 / 3 = 113,000 people. Good. Now the running total is 170,000 + 170,000 + 113,000 = 453,000 people. This is illustrated as follows.

How did they arrive at 530,000? I am clueless (note: see comment further down after the Sun article where the organizers gave the explanation of their (incorrect) calculation).

The organizers also claimed that they had volunteers at various points to count the crowd. No methodology was described, and no clue was given as to how the two estimates were reconciled to arrive at 530,000.

By the way, the police estimate was 200,000 this year. Last year, the organizers claimed 500,000 whereas the police estimate was 350,000.

For the record, I really don't care whether the number was 530,000 or 200,000 or whatever. Nor did it surprise me that the estimates from different parties should diverge so significantly. The real point was that there were a lot of people out there, enough to influence the September elections and thereafter.

Professionally, I don't like people playing loose and fast with numbers. Let me very specific about what I did. I made no assumption about people's motives and actions. I took the assumptions as stated about travel time, crowd capacity, start and finish times, and I made a clearly described set of calculations. I am not able to reconcile my number with the claimed total of 530,000. I explained what I did and I made no judgment. Period. My professional reputation is that I am infuriatingly logical, precise and unemotional. So please don't come after me with a battle-axe ...

As protection against revisionism, here is the original article in Chinese for the record:

民陣：53萬人上街勝去年
7月 1日星期四 21:20 更新

七一遊行龍尾在八時前抵達政府總部，歷時五個半小時，民陣指有五十三萬人參加遊行，警方估計有二十萬人。

民陣宣布，今日參加七一遊行的人數有五十三萬人，比去年的五十萬人的數字為高，民陣發言人孔令瑜表示，遊行隊頭在下午兩時半出發，用了九十分鐘時間抵達政府總部，而根據政府所給的資料，由維園至政府總部可容納約十七萬人，他們以此推算數字，另外民陣亦派義工在不同地點計算人流。而警方估計，今日只有二十萬人參加遊行。

立法會議員司徒華表示，對於有五十三萬上街，感到驕傲，「你不是比一粒糖佢食，就可以呃到佢地。」

由於參與遊行的市民在二時三十分前已差不多聚集滿維園六個足球場，主辦團體民間人權陣線與警方協議後，決定提前半小時，於下午二時半出發。

大會安排潮水式放行，讓足球場內的市民分批加入遊行，以免維園外聚集得太多人。龍頭隊伍在四時前抵達政府總部，維園內仍有許多人等待出發。而龍尾亦在五時半前出發，並於八時前抵達。

在酷熱天氣下，無減參加遊行市民的熱誠。市民高喊各式各樣的口號，包括「還政於民」、「爭取07/08普選」、「重建希望」。此外，亦有市民喊出「董建華下台」的口號，亦有要求保護海港、保飯碗及反對官商勾結。

有傷殘人士坐輪椅亦參加遊行。由於天氣酷熱，不少市民攜帶雨傘出席，形成一片「傘海」，天文台錄得今年最高溫度34.6度，並發出酷熱天氣警告，有逾五百人在維園或遊行途中不適，要到醫療站接受治療，四十二人被送院。今日是有紀錄以來氣溫最高的七月一日。

在去年七一遊行，天文台亦曾發出酷熱天氣警告，中午氣溫高達攝氏31度，當時民陣估計有五十萬參加遊行，警方則估計至少有三十五萬人參加遊行。

(The Sun) Updated at 3:00 am on July 2, 2004.

[translation] As for the vast difference between the 530,000 and 200,000 estimates, the police said that they used several methods to estimate the crowd size, including counting the number of people who departed from Victoria Park, the number of people in the middle at designated points and times and then they used the average. They also used the number of people who exited the subway stations along the route as reference. Civil Human Rights Front spokesperson Jackie Hung Ling-yu said that they also used the method suggested by the police to estimate the crowd size. If the route from Causeway Bay to Central is filled, there should be 170,000 persons. A total of three groups of people passed through, so that the total is 510,000. Another 20,000 did not go up to the Central Government Office. Therefore, the total number of marchers was 530,000.

Lecturer Sung Lap-kung at the City University School of Professional Studies thinks that the Civil Human Rights Front over-estimated and the police under-estimated. Research statistician Wong Kar-Ying from the Chinese University of Hong Kong Pan-Asian Research Center questioned that the density of the march this year was less than last year, so that a figure between 360,000 and 390,000 was more reasonable.

對於統計遊行人數出現五十三萬人與二十萬人的大差距，警方消息說，他們主要透過幾個方法計算人數，包括從維園每個球場面積估計出發時的人數、遊行至中段又以特定地點及時間作統計，再取出兩者平均數值，並以沿途地鐵站出閘人數作參考。民陣發言人孔令瑜卻說，他們也是用警方提供的方式統計遊行人數，即由銅鑼灣至中環遊行路線上擠滿人時的人數有十七萬，共有三批遊行市民經過，以此推算有五十一萬人遊行，另加兩萬名沒有行上政府總部的市民，總數五十三萬人。

城大專業進修學院首席講師宋立功認為民陣高估遊行人數，警方則低估。中大亞太研究所研究統籌員王家英亦質疑，今年遊行隊伍較去年鬆動，總人數應為三十六至三十九萬人較為合理。

Here is the scenario according to Jackie Hung Ling-yu. There were three groups of people, each group of size 170,000. The first left at 2:30pm and finished at 4:00pm. The second left at 4:00pm and finished 5:30pm. The third left at 5:30pm and finished at 7:00pm. And there was an additional 20,000 somewhere; this 20,000 is a 'fudged' number because the estimate of the number of people who did not finish is extremely difficult. I agree that these people exist, but why 20,000? Why not 15,000? Or 30,000?

Overall, this is a seemingly plausible scenario, except that it is not physically possible. The groups actually have to occupy space on the road. They cannot all leave from the same spot at the same time and arrive at the same spot at the same time. For example, the leader of the first group left at 2:30pm and finished at 4:00pm. That does not mean that the entire group has finished. At 4:00pm, the end of the first group had just left the starting point and they would not finish until 5:30pm. This is clear from the chart above that I made. It takes a grand total of 3 hours for a group of 170,000 to start and finish completely. So when the third group left at 5:30pm, the leaders would indeed finish at 7:00pm. However, the end of that third 170,000 group did not leave until 7:00pm and so they wouldn't finish until 8:30pm. The march was reported to have been over by 8:00pm, and the Transportation Department re-opened the roads at 8:05pm. So this is not a physically feasible scenario.

The following letter was sent to the South China Morning Post. It disclosed an unintended piece of information that challenges one of the assumptions used by the Civil Human Rights Front.

(Sunday Morning Post) Why the count was low. Letter to the editors from Dee Hartland-Swann of Repulse Bay. July 4, 2004.

Perhaps the police, besides march organisers could explain to your readers how they arrived in their tally of 200,000 as against 530,000 marchers ("Police taught us how to count, say organisers", July 3).

Where along the route did they stand to assess the numbers? If people were being counted at the end point of the march, the Central Government Offices (CGO), then estimates will wildly undershoot the actual total. Not all marchers went on to the bitter end. Let me explain, for I was there.

After several hours of good-natured slow walking in exhausting heat, our section of the demonstration finally arrived at the Legislative Council building. Marchers then had to be funnelled on to the narrow paths leading up to the CGO. Many parents felt their children had had enough by then and opted to leave. Others left the march because they were worried that the sheer mass of people was forming a bottleneck as the wide roads of Wan Chai and Central gave way to narrow paths. They also felt that their essential point had been made. If the police were simply counting those who arrived in front of the CGO, then they will have missed well over one-third of the demonstrators, the people who were sensible enough to avoid creating chaos by mindlessly surging forward into the bottleneck.

I note that the Post's aerial photo of Central in the July 2 issue showed the march at precisely the point where the orderly, tightly packed crowds began to disperse. The crowd I was in for more than two hours on Thursday afternoon was a lot more dense than the photo suggests.

Central to the Civil Human Rights Front's assumptions is that it took the marchers 90 minutes to complete the march. That may be true for the march leaders. But Ms. Hartland-Swann said that it took her more than two hours and she did not even reach the true finishing point. If the average time was in fact 2 hours, then the crowd size would in fact be smaller by a factor of 1.5 / 2 = 75%.

(Sing Tao Daily via ChineseNewsNet)

[translation] After the Civil Human Rights Front announced that 530,000 attended the 7/1 march, the figure was questioned by many people. On July 3, Sing Tao daily received a letter from a reader who is a graduate of Hong Kong University to point out that the Civil Human Rights Front's calculation was in error. They had turned two waves into three, and therefore inflated the march crowd by one-third. When the reader's explanation was shown to Lee Cheuk-yan and Jackie Hung at the Civil Human Rights Front, they did not rebut and therefore tacitly accepted that the calculation was in error. But they did not change their estimate of 530,000.

According to Sing Tao Daily, Jackie Hung waved her hands and said, "What do you want me to do?" She emphasized that she had no need to 'exaggerate' the number. Apart from the formula, they also counted with humans and the number was similar. She also pointed out that the number was not important, and that the really important thing was that the government has to make a resonse to the demonstration.

The Civil Human Rights Front has been form on the 530,000 estimate which is based upon a set of calculations. The method is based upon police data which say that the route from Victoria Park to the Main Government Building can contain 170,000 persons who will take 90 minutes to complete. Since the march took 5.5 hours to complete, the Civil Human Rights Front divided this by 1.5 to obtain three waves of marchers. 170,000 x 3 = 510,000. If 20,000 more are added as those joined midway, the total became 530,000. But this calculation involves an elementary school error present in the "tree planting" problem by double counting the last wave.

On July 3, Sing Tao Daily received a letter from reader Yuen Chi-keung, who had just graduated in biology from Hong Kong University. He pointed out that there was a serious error in calculation, and he said that the crowd estimate was a hot topic at Internet forums.

Yuen Chi-keung pointed out that based upon the facts of the march and the assumptions made by the Civil Human Rights Front, the route can contain 170,000 persons at a time and the trip took 90 minutes. Between 230pm and 400pm, the first group of 170,000 departed. The second wave departed between 400pm and 530pm. The third wave departed at 545pm. In other words, the third wave did not contain 170,000. The whole march only contains 2.17 waves, which gives 2.17 x 170,000 + 20,000 = 390,000.

The newspaper also interviewed Lingnan University Public Administration Research Department head Lee Peng-kwong and he agreed with the calculations of reader Yuen Chi-Keung while declaring the Civil Human Rights Front's number to be in error. When Civil Human Rights Front's Lee Cheuk-yan was interviewed, he also agreed that Yuen Chi-keung's calculation was correct but he believes that there were many people who joined in the middle. When asked if their calculation was wrong, he said that he needed to go back and discuss with his people.

Jackie Hung said that apart from this calculation, they also used human observers to make counts. When asked about why Human Rights Monitor's count was only 160,000, Hung said that Human Rights Monitor Chief Executive Law Yuk-kai explained to her that if the error in the count was factored it, the number would be more than 200,000 and if more than half the people joined the march in the middle, they can still arrive at a 400,000 to 500,000 total.

香港民陣公布七一大遊行有五十三萬人，連日受到多番質疑，星島日報七月三日收到一位港大畢業的讀者來函，指出民陣計算方法出現錯誤，將本來只有兩輪的遊行錯計為三輪，以至遊行人數高估了三分之一。當天將讀者的論據向民陣的李卓人和孔令瑜求證，他們都未能提出反駁，默認計算方法可能出錯，但就未有改變五十三萬人的說法。

星島日報四日報導說，孔令瑜聽完讀者的計算方法後無奈說，「你想我點答呢﹖」她強調，自己沒有「作大」的必要，而他們除了依據公式推算外，還進行人手點算，亦得出相若的數字。她又指最重要不是數字，而是政府要回應訴求。

民陣連日來堅持五十三萬的遊行人數，是建基於一套計算方法。這套方法引用警方的數據，指由維園到政府總部這條遊行路，可一次過容納十七萬人﹔而由維園行至終點，需時個半小時。民陣於是以遊行由出發到終結的五個半小時，除以九十分鐘計算，計得可以有三輪遊行。以十七萬乘以三輪，即是五十一萬，再計算兩萬中間插入人士，即五十三萬。不過，民陣這一計算方式，卻犯了一個類似小學算術中的「植樹問題」錯誤，重複計算了最後一輪的人數。

星島日報三日收到一名剛在港大生物化學系畢業的讀者袁智強的來函，指證民陣計算方法出現嚴重錯誤，他並指上街遊行數目連日來已成網上討論區的熱門話題。 (chinesenewsnet.com)

袁智強指出，根據遊行資料和民陣提出的基本假設，整段遊行路面每次可容納十七萬人，而遊行路程為個半小時，由兩點半至四點鐘這個半小時內出發的，即為第一輪十七萬人﹔第二輪十七萬人則是在下午四時至五時半出發的人。由於最後一批出發的人士是在五時四十五分，換言之，第三輪根本沒有十七萬人。整個遊行只有二點一七輪人次，再將二點一七人次乘十七萬人再加中途插入的兩萬人，即應有三十九萬人。

報導說，嶺南大學公共管治研究部主任李彭廣三日接受查詢時，認同讀者袁智強的推算方式，同意民陣的計算方法出錯。而民陣李卓人接受查詢時，亦同意袁智強的計算無錯，但他認為可能有很多人是在中間插入。問他民陣是否計算錯誤，他表示要回去再討論。

孔令瑜說，民陣除了這個計算外，亦有進行人手點算。對於民陣成員人權監察進行的人手點算只有十六萬人，孔令瑜說，人權監察總幹事羅沃啟曾向她解釋，若將人手點算的誤差計算在內，可以有二十多萬，若有一半人是從中間插入，亦可以推算出四、五十萬的數目。

The deduction by Sing Tao Daily reader Yuen Cheuk-yan is similar to my physicality argument. There is an additional piece of information that Yuen Cheuk-yan used: the third wave departed between 530pm and 545pm, which meant that the size of the third wave was only 1/6-th of 170,000. That was how he got to 2.17 x 170,000 + 20,000 = 389,000. The 545pm departure time for the last significant number of marchers is not being disputed by the Civil Human Rights Front.

As the Sing Tao Daily report said, this is elementary school mathematics and is logically irrefutable. However, the Civil Human Rights Front is assuming an unethical position in which they have made up their minds that they needed to reach 500,000+ and they are rationalising in other ways to get there.

After the central theoretical foundation of their calculation had just collapsed due to an irrefutable error, they claimed that there could have been more people who joined in the middle. In fact, there will be conveniently enough people to reach 400,000 to 500,000.

Then they said that their calculation was supplemented by counts from human observers, for which they did not cite the results. How low was that number that they were unwilling to disclose initially? The Sing Tao Daily article said that Human Rights Monitor counted 160,000 marchers, a number which the Civil Human Rights Front spokesperson does not dispute. So we are now left with the need to come up with 200,000+ phantom marchers who all joined in the middle in order to hit the production quota. This is strongly reminescent of production statistics in the Great Leap Forward era. This is 'fudgy' statistics, and the statistician in me finds this appalling.

Look, this estimation business with the Civil Human Rights Front is getting silly. Here is my recommendation for a public position:

"We thank people from different walks of life for pointing out that we made a technical error in calculating the crowd size. We agree that we were wrong. At this point, we cannot definitively provide an accurate estimate based upon the data that we have collected. All those of you who have discussed this issue would surely recognize that there are a number of theoretical ways of estimating crowd sizes, but they all seemed to have some kind of flaw. We will consult with the community to find the best method of counting for future projects, and this method is transparent and will be published beforehand in order to solicit feedback.

Meanwhile, the precise number of marchers is less important than the undeniable fact that the people of Hong Kong have once again demonstrated that they wanted ... [insert list of demands] ... to which the Hong Kong SAR and Chinese central governments are now obliged to address."

The more they try to fudge, the greater the loss in credibilty. Ask yourself: would you want the people in the Civil Human Rights Front to manage the government budget? Frightening thought, isn't it? Move on already ...

The Civil Human Rights Front posted a press release on their website dated July 4, 2004.

[translation] Concerning the problem about the number of marchers on July 1, the Civil Human Rights Front emphasized again: on that day, from the first arrival at the Central Government Office to the last arrival took 4 hours; the time to go from Victoria Park to the Central Government Office is 90 minutes. Therefore, there were about 3 trips. According to the estimate provided by the police to the Front of 170,000 per trip, adding people who joined the march in the middle, we estimated that there were about 530,000 persons.

The Front's method of calculation may contain imperfections. Therefore, the Front welcomes other organizations to disclose their estimates and methods in order to let the citizens decide for themselves and also for the Front to consult in the future.

就七月一日大遊行參與人數問題，民間人權線重申，當日以第一個到達政府總部至最後一個到達的人士共需時四小時；而由維園到政府總部共需約一小時三十分鐘，故共接近三轉。而每轉按警方對本陣線表示共十七萬人推算，再加上中途插隊人數，故推算出約五十三萬人。

本陣線推算方法定有不盡完善的地方，故本陣線歡迎其他機構公佈有關點算結果及方法，讓市民自行判斷，並作本陣線日後參考。

The Civil Human Rights Front seemed to be in disarray regarding a response. On July 5, a day before the Front members convene to go over the issues, the following conflicting remarks were reported by Sing Tao Daily:

Civil Human Rights Front spokesperson Hung Ling-yu admits that the announced figure is definitely not authoritative. She said that they will discuss the method of calculation in tomorrow's meeting. She emphasized that the counting done by the volunteers is reliable. She said that it is possible that they will get more volunteers to work in the future. She admitted frankly that the work of these volunteers was stressful, which was why there were only a few volunteers this time. [Comment: That is missing the point! It is not the quality or quantity of work done by the volunteers that is the problem. Nobody is blaming the volunteers, whose work was ignored in the 530,000 number anyway. To arrive at the 530,000 number, they required no resources whatsoever. They met with the police and obtained a density estimate of 170,000; they saw that the march started at 230pm and ended at 800pm; they plugged the information in a faulty formula and added an arbitrary 20,000 more without any basis. The whole job can be done by one person without ever getting out of the chair. There is no point in getting an army of volunteers if you are going to use a bad formula that ignores the fruits of their labor at the end.]

She said that that the Front will discuss the problem of the march number tomorrow. But she did not directly respond as to whether they will make a public announcement that there was a problem with the number. She only said that if there really was a problem with the number that needs to be publicly disclosed, then they will discuss it. [Comment: Very circular and confusing; lexicographic hair-splitting in the tradition of White House spokesmen Ari Fleischer and Scott McClelland.]

The assistant spokesperson Tam Chun-yin said that if the error was too large, then there should be a public announcement. [Comment: This is dodgy but absolutely correct for public relations. Large error=bad. Everybody agrees with that. But he is not saying what constitutes a large error.] Tam said that the Front's method is the only thing that they can do with the limited resources. He admits that they did not consider the problem of density. He also said that they will examine the methods suggested by the scholars and attempt to recalculate using different methods. If the numbers are too different, they will make an announcement. [Comment: Dodgy again and quite correct for public relations. This man has a future in this business.]

Another spokesperson Rose Wu Lo-sai said that they will not make another statement on the number again. She believes that there is no absolutely correct method of calculation right now, so they are not going to re-calcuate the march number. The recommendations from the scholars can be used by citizens for reference to make their own determinations. [Comment: Bad! Really bad! This problem is not that there are five different ways of estimation, each with their own minor flaws, so that it is impossible to choose. The problem is that the Front's number is now known to be very wrong due to an elementary school arithmetic error and therefore completely out of line with the other numbers. The Front does not have to be concerned about anyone else's problems with methodology or calculations, but it is responsible for the Front's arithmetic errors.]

As for any future marches, the smart thing for the Front to do is to sub-contract the job to a reputatable independent professional group, such as the Hong Kong University Department of Statistics and Actuarial Science. Let the other guys take the heat. This is not the core competence of the Civil Human Rights Front and has proven to be a disastrous distraction with respect to delivering the core messages.

On July 5, 2004, in Chai Shiu-fung's column in Sing Tao Daily, he retold the story of the reader's letter and Jackie Hung's response. Then he added the following comments:

[translation] Actually, the question to Jackie Hung Ling-yu was easy to answer. All she has to do is to admit the error, and that is the best solution.

But Hung Ling-yu did not want to admit any errors based upon several factors. First, the whole world has accepted their 530,000 figure. To say that this number was wrong and that it was a calculation error for an elementary school problem would be too weird. They will become an international laughing stock and that would be unbearable. Second, they don't know how to explain it to the citizens. A couple of days ago, a scholar said that he estimated that there were only 200,000 marchers and Hung Ling-yu accused him of being irresponsible for betraying the marchers. If they now admit to having been wrong, they are the ones who betrayed the marchers. Third, to decrease the number of marchers by such a large amount would make people think that there were only a few marchers and therefore dilute the 'power' of the protest.

But Chai Shiu-fung feels that even though this is an enormous joke, the damage would be less than if they insist on continuing to be wrong. If I am someone who is pro-China/Hong Kong, I would not want Hung Ling-yu and and the Civil Human Rights Front to admit being wrong and I would want them to continue to be wrong. Conversely, the best strategy for the Civil Human Rights Front is to admit being wrong. 200,000 is a very big number already, and there is no need to make it ever larger.

Chai Siu-fung will also mention that the number of marchers was less than last year. Yet in the newspapers next day, most of them put down 530,000 marchers underneath the headlines. Last year, the organizers claimed only 500,000. This year, there appeared to be fewer marchers, and yet there was a 530,000 figure. The media simply parroted the number without any critical evaluation.

Chai Shiu-fung has chatted with his media friends about this problem, and realized that they have a certain type of attitude. A lot of people doubted the 530,000 figure. "But since a lot of people were marching, I might as well as follow the people's will. I'll take whatever the Civil Human Rights Front says and I won't use the police estimate. This way, I won't get attacked by the pan-democratic camp or receive complaints from readers." This is an error of partisanship.

In the late 1950's, Mao Zedong initiated the Great Leap Forward in China. There were scientific farmfields that yielded ten thousand jin (a jin equals half a kilogram) of rice per hectare. This came about under the same system of thinking. The organizers were over their heads, the observers on the sidelines banged gongs and drums to publicize the people's will, and a tragedy results.

Please note carefully the discussion of self-censorship in a democratic environment. This is every bit a tyranny as under a dictatorship.

(Tai Kung Po) Was it an olive branch or an unsheathed sword? By Mei Ying. July 6, 2004.

[translation] The 7/1 march represented many demands from the citizens, not just for the double direct elections. Does the Civil Human Rights Front and the Democratic Party represent all of the citizens of Hong Kong? They don't respect the China Liaison Office in Hong Kong, they don't respect the Special Administrative Region government and they want "direct communication" instead. Is this a three-footed stool or a four-footed stool? They want to win votes to paralyze the SAR government and they want to use "the will of the people" to impose on the central government. Does this represent sincerity in communications? Was it an olive branch or an unsheathed sword?

How many people participated in the 7/1 march? Spokesperson Hung Ling-yu originally said 200,000. Then it was gradually raised to 250,000, then to 350,000, then to 400,000 and it was finally announced to be 530,000.

The police said that there was 200,000 persons.

On July 2, other than the patriotic newspapers, the other newspapers showed headlines that said 530,000 persons attended the 7/1 march. I admired Tai Kung Po for insisting on the truth with this unique headline: "Exaggerated numbers at the marcher again, 160,000 became 530,000."

This author was not among the marchers and believes that the mainstream Hong Kong public opinion is represented by those who did nothing and did not march. But I was concerned about how the Civil Human Rights Front spokesperson could arbitrarily increase the numbers and subjectively make up "statistics." So I contacted various friends until late at night, and I asked authoritative persons to provide reliable numbers. At midnight of July 2, I obtained the most reliable and most accurate estimate: the number of 7/1 marchers could not be more than 200,000.

The authoritative source told me: the police estimate could only be more, not less. The reason is very simple: the police is responsible for ensuring the safety of the citizens, and they cannot underestimate the number of marchers; on top of the observed data, they needed to add another 10% to 20% in order to made sure that the police force is adequate.

The authoritative source also told me: a foreign news agency used advanced scientific techniques to estimate the crowd size. At 3pm, that news agency estimated 100,000; at 4pm, it was 150,000; at 7pm, it was 200,000. The conclusion was that is was no more than 200,000.

The authoritative person made these calculations: the distance from the Central Government Office to Victoria Park was 5.2 kilometers. The marchers formed eight people per row, or 16 people for two rows; people are about 30 to 40 centimeters apart; each square meter of roadway can accommodate 32 people; if the roadway was full, there would be 160,000 persons. But the weather was hot and people were further apart than usual. About one half or one third of the march was less dense. "Long Hair" stood under the Goose Neck Bridge to put on a show that caused congestion in order for the media to take films to show the dense crowd. But for the whole trip, it would be lucky if there were 120,000 to 130,000 marchers. When the march leaders reached the Central Government Office, two of the six soccer fields were already empty. Because the weather was hot, many people used umbrellas and fanned themselves. So each soccer field was likely to have only about 6,000 to 7,000 standees. So the number of people at the rear of the march was at most 30,000. Thus, the police estimated that the total number of marchers from the head of the march to the tail of the march was 160,000. They then overestimated by adding another 20% in order to make sure that they have enough police resources to maintain public order and safety. Thus, they made the public estimate of 200,000.

The Civil Human Rights Front reported an inflated number and most of the media used their reported number to deceive their readers. But of the 200,000 from the police and the 160,000 from Tai Kung Pao, which is more accurate?
On July 3, Hong Kong University Department of Statistics and Actuarial Science senior lecturer Yip Siu-fei was interviewed by newspaper reporters and said that the department had sent people to count the number of marchers at two locations during the march and estimated that about 200,000 people were on the street. This is about the same as the police estimate. [Note: The final number by the Yip Siu-fei team would in fact be 165,000].
Thomas Lee is an expert in satellite photo analysis. He is a visiting lecturer teaching the Hong Kong University Department of Geography Remote Sensing course and the Polytechnic University Geographical Positioning Systems course. He used satellite photo analysis computer software to analyze aerial photos taken by a newspaper, after considering light diffraction from a typical human body. He found that there were 100 people in every one hundred square meters in the Victoria Park soccer fields, although it could be as low as 85 people sometimes. Based upon the assumption that the total road surface was 72,000 square meters, there should be 72,000 persons on the roadway at 430pm on July 1. Based upon the assumptions that each trip took 90 minutes and the march lasted 5.5 hours, excluding any people who joined in the middle, the total number of people was 264,000 with an error rate of 15%.
Facing the academic scholars who challenge the 530,000 number of 7/1 marchers, Hung Ling-yu said threateningly: "These scholars ought to behave responsibly. If they say that there were only 200,000 persons, how can they face the citizens who marched?"
Did Hung Lin-yu think about this?: If 530,000 and 200,000 differ by so much, then should the scholars make a scientific analysis? Should the police ignore the objective conditions in terms of assigning resources? All other persons and organizations do not need to conduct any independent investigation, because they can listen to politically motivated Civil Human Rights Front to arbitrarily and intentionally make announcements. Whey happened to freedom of the right to information? On July 1, the march also included many other organizations and people who were celebrating the return of Hong Kong and there were others who joined in the carnival atmosphere. But after the scandal of the exaggerated number from the organizers, the word is that the number does not matter and the important thing is that the central government and the SAR government must respond to the demands of the citizen. Whether it was the exaggerated number, or the scandalous attempt to cover up the exaggerated number, one thing is clear: the opposition is waving an olive branch on one hand b saying "to communicate not for the sake of communicate," to "converse directly without any pre-conditions" and to "let Hong Kong and Beijing cooperate fully"; on the other hand, they are organizing large marches, they are exaggerating the attendance, they are using "the will of the people" to intimate the SAR government and the central government to make the National People's Congress to change their decision about direct elections in 07/08, by "trying to make possible what is impossible."
The 7/1 march represented many demands from the citizens, not just for the double direct elections. Does the Civil Human Rights Front and the Democratic Party represent all of the citizens of Hong Kong? They don't respect the China Liaison Office in Hong Kong, they don't respect the Special Administrative Region government and they want "direct communication" instead. Is this a three-footed stool or a four-footed stool? They want to win votes to paralyze the SAR government and they want to use "the will of the people" to impose on the central government. Does this represent sincerity in communications? Was it an olive branch or an unsheathed sword?

是橄欖枝還是劍出鞘/梅英

「七．一」遊行市民的訴求是多方面的，不是單一的所謂「雙普選」訴求。「民陣」、民主黨能代表香港全體市民麼？不尊重中央駐香港聯絡機構，不尊重特區政府，要取而代之「直接對話」，這是「三腳」還是「四腳」？爭取選票以癱瘓特區政府，挾持「民意」以令中央，這有溝通、合作的誠意麼？這是橄欖枝還是劍出鞘？
七月一日有多少人參加遊行？召集人孔令瑜最初宣布是二十萬人，後來逐步升級為二十五、三十五萬、四十萬，最後晚間宣布為五十三萬。
警方宣布為二十萬人。
七月二日除愛國報刊外，本港各報均以有五十三萬人參加「七．一」遊行列為頭條新聞。我最佩服、最欣賞的是《大公報》。敢於堅持事實，作如下標題的獨家報道：「遊行又報大數，十六萬變五十三萬」。
筆者是不參加遊行的人士之一，確信不參加遊行者「無為政治」所表達的民意，才是香港的主流民意。所以既關注「民陣」召集人所宣布的那種可笑到隨意加碼，憑主觀意願宣布的「統計數字」，更在深夜與各方友好交換意見，並向權威人士索取可靠數字，於七月二日凌晨零時零點獲得最可靠、最準確數據：今年參加「七．一」遊行人數不會超過二十萬。
權威人士告訴我：警方報出的二十萬人，是有多無少。道理非常淺白，警方負責維護遊行市民安全，負有極嚴肅的保安責任，絕不能把參加遊行的人數低估；相反，要在客觀統計評估的基礎上，增加百分之十到百分之二十，以便安排足夠的警力。
權威人士還告訴我：某外國大通訊社，對香港每次大型遊行示威隊伍例必以先進科技進行測算。該通訊社下午三時統計數字為十萬，下午四時為十五萬，晚七時為二十萬，結論不超過二十萬。
權威人士還作了如下計算：從政府總部到維多利亞公園全長約五點二公里。遊行隊伍一排八人，兩排十六人；人距約三十至四十公分，一平方公尺路段內最多只能容下三十二人行進：全程人滿約十六萬人行進。鑑於天氣熱，人距大，約有一半到三分一路段人流疏鬆，儘管「長毛」鵝頸橋下站台演說，造成一段人流擁擠，似意在給傳媒拍攝出遊行隊伍擁擠之效果。但全程觀之，全程實際人數約有十二、三萬就算不錯。當龍頭抵政府總部，龍尾的維多利亞六個球場有兩個大部分空置。天氣炎熱，打傘者眾，普遍搖扇，每個球場只能站六、七千人，龍尾的四個多球場約有三萬之眾。估計警方從龍頭到龍尾以總數十六萬人計，再加上以百分之二十的比例高估，以便布置較充分警力維持治安，便作出遊行人數約二十萬之公開評估。
「民陣」報大數，大部分傳媒以其所公布的五十三萬數字誤導讀者。但警方報二十萬，《大公報》報十六萬，誰的準確？
香港大學統計及精算學系高級講師葉兆輝三日接受報業查詢時表示，該學系在前天「七．一」遊行期間，在遊行沿途路線兩個目標地點派員進行人數精算統計，確定約二十萬人上街，與警方估計相若。
從事衛星圖片科技研究，身兼港大地理系遙感課程及理工大學衛星定位應用課程的客座講師李偉鵬，利用「太空科技衛星圖片分析軟件」，用多光譜分析原理，配合人體的大小及形狀，將報章提供的高空拍攝相片作出分析。顯示每個維園球場中，每一百平方米有一百人，而最疏的人群則為八十五人，加以一百平方米有一百人計，全路段面積約為七點二萬平方米推算，七月一日四時半遊行路面約有七點二萬人，按九十分鐘一程人群計，不計中途插隊人數，遊行五個半小時內，共有二十六萬四千人，而軟件偏差率為百分之十五。
面對學者質疑「七．一」遊行五十三萬人上街，孔令瑜語帶威脅說：「那些學者說話應該負責任，他們說有二十萬人，怎對得起參加遊行的市民！」
孔令瑜是否想到：五十三萬與二十萬差距如此之大，是否學者不應作科學分析，警方不應基於較客觀現場評估而分配警力以策安全？而各方人士、機構不應獨立調查研究，只聽基於特定政治目的「民陣」的隨意性及有目的性的宣布，則自由權利，民眾知情權安在？有人七月一日前鼓吹包容盡可能多的團體和市民，包括表示慶回歸、表示以參加嘉年華心情投入的市民一起來，以壯大遊行隊伍，但在遊行組織者浮誇報大數的醜聞被揭後，又說數字不重要，重要的是中央及特區政府如何回應市民訴求。報大數也好，掩蓋報大數者醜聞也好，都暴露了一點：反對派一方面搖橄欖枝，要「不為溝通而溝通」的溝通，要「不設任何前提」的「直接對話」，「讓香港和北京全面合作」；另一方面，組織大遊行，誇大遊行數字，挾「民意」屈特區政府兼令中央，要求人大常委會改變決定，○七、○八年實行雙普選，「明知不可而為之」。
「七．一」遊行市民的訴求是多方面的，不是單一的所謂「雙普選」訴求。「民陣」、民主黨能代表香港全體市民麼？不尊重中央駐香港聯絡機構，不尊重特區政府，要取而代之「直接對話」，這是「三腳」還是「四腳」？不確認全國人大及其常委會的領導地位，不按憲法、《基本法》及人大常委會的憲制性法律決定辦事，利用多種訴求的遊行作單一的政治籌碼，爭取選票以癱瘓特區政府，挾持「民意」以令中央，這有溝通、合作的誠意麼？這是橄欖枝還是劍出鞘？

(Sing Tao Daily) The Front Will Meet Today to Discuss the March Number. July 6, 2004.

[translation] The Civil Human Rights Front was challenged about its method of estimating the 7/1 march participants and will hold a meeting today to decide whether they need to make a statement. Yip Kwok-him of the DAB believes that if an error was made, then they must state the correct figure. Some scholars believe that the Front's method was problematic, but the Front can be excused for making errors due to their limited resouces. Similarly, the police's estimate is also imprecise, so that both sides must offer explanations.

The Civil Human Rights will hold a meeting to evaluate the 7/1 march. They will discuss the disputed march number during the meeting. The Front's assistant spokesperson Tam Chun-yin was just the estimated number of people on the day of the march, and they will discuss today if it is necessary to make a clarification. Front member Chuang Yiu-kwong suggests that they can conduct a survey as remedy. If 20% of the people said that they participated, then based upon the population of Hong Kong , it means that about 1 million people marched.

Chinese University of Hong Koong Sociology Department Associate Professor Chan Kin-man, who is also a member of the Democratic Development Network, believes that there was actually fewer than 530,000 people, and he believes that the method of calculation has problems even though he understands that the Front has limited resources. He says that the Front should not "deceive themselves even if they are not to be blamed." He believes that it takes a lot of resources to obtain the march number, so that an error is not a serious matter.

The Front's Lee Chuk-yan pointed out that "number is in the people's heart" and they will not revisit the number. But they may examine various methods to see what they should do in the future. The Front's Choi Yiu-cheung was responsible last year for announcing the march number and he said that they counted the marchers at three different locations, and they ended up using the 500,000 found at Wanchai. There may have been more people who joined the march later on, so that they claimed that there were more than 500,000 people. He would not criticize the method used this time, but he respects and supports the number announced by the Front.

Democratic Party vice-chairman Ho Chun-yan still insists that the method of calculation was reasonable, and said that they understated last year's number. Democratic Party's Cheung Man-kwong said that the number is unimportant, because the important thing is to listen the voice of the citizen. People should be more open-minded, and the Front and the scholars can discuss the methods of calculation.

DAB former chairman Tsang Yuk-sing, also a legco member, said that the organizers wanted to get a bigger number than last year, so they "exaggerated" the number. But he says that it is up to the organizers to decide whether they need to clarify. Another legco member Cheng Yiu-tong said that it does not matter how many people were there; there were a lot of people out there and the SAR government needs to consider their sentiments.

民間人權陣  在七一遊行的統計方法，被質疑計錯數，今日會開會討論是否作出澄清。民建聯葉國謙認為，有錯便應公開正確數字。有學者認同民陣的計算方法有問題，但民陣礙於資源有限難免出錯，相反警方的統計也並非準確，認為雙方都應作出交代。

　　民間人權陣  將於今天開會，檢討七一遊行，會上會討論備受爭議的遊行人數問題。民陣副召集人譚駿賢昨表示，五十三萬只是遊行時統計出來的人數，會在今天研究數字真確性，再決定是否澄清。民陣成員莊耀洸建議，進行民調作補救，若結果顯示有兩成人參加，以香港人口計算，便可證明約有一百萬人參加。

　　本身屬民主發展網絡的中大社會學系副教授陳健民認為，實際人數並無五十三萬，亦相信民陣的計算方法有問題，但明白民陣資源有限。他說 ﹕ 「（民陣）唔好自己呃自己，但都唔怪得佢  。」他認為，遊行人數要花很多資源計算，計錯數也不是嚴重事件。

　　民陣的李卓人指，「數字自在人心」，不會重新評估，但會參考不同的評估方法，日後再定如何點算。去年負責公布遊行數字的民陣蔡耀昌表示，去年分三個點點算，以灣仔最多五十萬人計算，當中也有人加入或離去，故公布人數超過五十萬人，他不願評論是次遊行點算方法，但尊重和支持大會公布數字。

　　民主黨副主席何俊仁，依然堅持計算方法合理，又指只是去年「報少數」。民主黨張文光認為，數字並不重要，重要是聽取市民訴求，大家要放開胸懷，民陣與學者可討論點算的方法。

　　民建聯前主席兼行政會議成員曾鈺成表示，主辦者急於想人數較去年多，故「拋大  」，但認為應由主辦者自行判斷需否澄清。另一行會成員鄭耀棠指，不論人數多少，反正都是有這麼多人上街，這是特區政府應考慮的民情。

Some people just don't get it yet, because they are confusing various issues.

First, let me distinguish between the calculation error and the defective method. This year's method was based upon a density provided by the police (170,000 persons on the roadway), the observed time to go from start to finish (90 minutes) and the start and end times (230pm and 800pm). Then the calculation proceeded to make an elementary school-type error by assuming that there was time for 3 trips and thus they obtained 530,000. This is wrong. This is so wrong that an elementary school student would know. They cannot let the calculation error stand because it is scandalous. I would suggest that they re-do the calculation and the reasoning by the Sing Tao reader is fair and will lead to a revised number of 389,000.

The method is defective because it depends on the density provided by the police (170,000), and the scholars have evidence that the density may be lower. On this matter, the Front does not have to answer for it. They will simply state their methodology and the number is whatever comes out of this set of assumptions, which had seemed fair and reasonable at the time that they designed it and was done in good faith. The Front got the density from an authoritative source -- the police -- and everything else (start and finish times, travel times) is not in dispute. If scholars don't agree with the density assumption, they can make their own adjustments and different scholars seem to have different opinions. The Front's number can stay with their set of assumptions. However, that calculation error above is a different story and cannot stay.

Second, the complaint about limited resources is lame. How many people did the Hong Kong University team have? They got 11 students and faculty members to work half a day without pay. The data entry and tabulations are also negligible. They may have done 600+ interviews, but the survey response is only yes/no and all the responses could have been written down on one piece of paper as Y's and N's and then counted up. If the Front doesn't feel comfortable, they could have picked up the phone and asked someone at a university to organize a volunteer student team for an interesting and meaningful project.

Third, forget about doing that survey afterwards. If you thought that getting eleven people to work half a day was beyond the means of your limited resources, you should see the bill for conducting several hundred interviews properly (note: you can get it done a lot cheaper by doing it improperly, of course). Another big problem is that the population has been contaminated by the extended coverage of the dispute over the number, and there is no guarantee that you will get truthful answers.

Of course, everybody agrees that the number doesn't really matter because we all saw a mass of humanity out there. But this is not the point for now. At issue here is the credibility of the Civil Human Rights Front. If they decline to withdraw the 530,000 number, what is a person supposed to think? Here are the two choices:

On one hand, these people are so dense that they can't understand the simple calculation error. If an elementary school student makes this mistake, he/she gets zero points; if the Civil Human Rights Front makes this same mistake, it is alright. What will we tell our kids?

On the other hand, these people knew that they had made a calculation error but they prefer to let the big (and erroneous) number stand, either for political advantage or to avoid public embarrassment.

In either case, do you want this kind of people to be directly elected to handle the government budget and other affairs?

The next article is a report on the outcome of the Civil Human Rights Front meeting on Tuesday.

(Sing Tao) The Civil Human Rights Front insists on 530,000 figure. July 7, 2004.

[translation] The 7/1 march organizer Civil Rights Human Rights indirectly admitted that there was an error in the calculation of the number of marchers that day, but they insisted that the number of 530,000 will not change. They welcome people to propose new methods of calculation. The Civil Rights Human Rights admitted that its method of calcuatlion "definitely has imperfections" and therefore welcome other organizations to publish estimates and methods in order to let the citizens decide on their own.

On the day of July 1, the Civil Human Rights Front calcuated 530,000 persons, based upon people departing for 5.5 hours allowing for 3 waves. According to Hong Kong University graduate Yuen Chi-keung and Hong Kong University Department of Statistics and Actuarial Science Senior Lecturer Yip Siu-fei, this formula double-counted the last wave and mistakenly turned two waves into three waves. At yesterday's meeting of the Civil Human Rights Front, they revised the original formula. They used as their basis the fact that there were 5.5 hours between the time that the first marcher left and the time that the last marcher finished. But Civil Humans Right Front refused to formally admit that the original method of calculation was wrong.

According to the Civil Human Rights Front, the difference between the first and last start times was 4 hours and trip lasts 90 minutes, therefore there was almost 3 waves. Based upon a density of 170,000 per wave, plus the additional people who joined in midway, they therefore estimate the total to be about 530,000.

The Civil Human Rights Front also issued a statement to explain the new calculation. They re-emphasized that the Hong Kong SAR government must respond to the marchers. "Under the hot sun of July 1, many citizens still got out on the street. This should be treated seriously by the government and society, instead of fighting over the number of particpants that day."

Concerning the new method of calculation by the Civil Human Rights Front, Lo Chi-kwong who studied economics and statistics, says that the method of calculation already includes those people who joined in the middle. To be more precisely, based upon the logic of the Civil Human Rights Front, the number of marchers should be 450,000 and not 530,000.

Another mathematician who does not wish to disclose his name said that the Front's new calculation did not double-count. But if this formula is used, then the route should have been packed full with people at 630pm and the last one will reach the Central Government Office at 800pm. Actually, the last group of marchers left Victoria Park at 545pm, and at 630p, much of the road from Victoria Park didn't have people. Therefore, whether this hypothesis is accepted or another hypothesis that the trip time is 2 hours instead of 90 minutes, the number of waves is two instead of being near three.

Civil Human Rights Front organizer Rose Wu Lo-sai was asked after the meeting about whether they are firm on the 530,000 figure and their method of calculation. She said, "There is no possiblity of overturning the 530,000. We have calculated the number from a sound basis. The public can decide." She admits that the Front's method has imperfections, included not determining the density carefully so that they may not achieved the maximum density of 170,000 given by the police. But because the density varied at different times in different places, it is impossible to accurately calculate the number of people on that day.

She emphasized that the Civil Human Rights Front has disclosed their method of calculation and the assumpitons. They have discussed the matter and they believed that the 530,000 is an accurate figure. This figure was based upon the calculation by the Front and can be used by everyone as reference. Different organizations have their own methods, and the citizens have a right to choose to believe or not. The Civil Human Rights Front will consult other methods.

When asked whether this number will be believed by the public, she said, "Whether or not to believe the 530,000 number will be decided by the public. Why do you feel that the public may not believe it? Why don't you go ask the public? Why don't you go ask the marchers? Are the 200,000 claimed by some scholars just as accurate? Every scholar can have a point of view, but no one can offer the most accurate number. The march number is subject to variation, and there will be differences."

Human Rights Monitor made a count on the day of the march and obtained 160,000. Yesterday, Law Yuk-kai said that he did not want to be tangled in the problem of the number of marchers.

　　七一遊行的主辦團體民間人權陣  ，昨天間接承認它在七一當天對遊行人數的計算方法出錯，但就堅持五十三萬遊行人士的數字不變，並重新提出新的推算程式。民陣承認，它的推算方法「肯定有不盡完善的地方」，故歡迎其他機構公布有關點算結果及方法，讓巿民自行判斷。

　　民陣在七一當天，計算出五十三萬人，是以遊行由出發到終結的五個半小時為基礎，推算出可以遊行三輪。但被一位港大畢業生袁智強及港大學者統計精算系高級講師葉兆輝指出，這計算方法重複計算最後一程人數，把兩輪遊行誤計為三輪。民陣昨晚開會檢討後修正了本來的計算方法，不以出發到終結的五個半小時作為基礎，而由第一個到達政府總部到最後一個到達的時間為計算基礎。不過，民陣發言人孔令瑜卻拒絕正式承認當初的計算方法出錯。

　　根據民陣的最新推算，上述時間相差四小時，以每輪個半小時計，故有接近三輪。每輪按十七萬人推算，再加上中途插隊人數，故推算出約五十三萬人。

　　民陣又發表聲明交代最新的推算結果，聲明並重申，強烈要求特區政府正視遊行市民的民主訴求，以及作具體回應，「在七一當日烈日當空，氣溫酷熱，但仍有大批市民上街，應得到政府及社會人士重視，而不是糾纏於當日的參與人數」。

　　對於民陣新的推算方法，讀經濟及統計系出身的羅致光亦表認同，但他指這推算方法本身，已將插隊人數計算在內。若精確計算，並以羅致光及民陣的邏輯推算，遊行人數亦只是四十五萬而非五十三萬。

　　一位不願透露姓名的數學愛好者則指出，民陣最新的計算方式在數學概念上雖沒有重複計算的問題，但若採用這推算方式，晚上六時半的時候，理論上整個遊行道路上，都會鋪滿人，因為最後抵達政府總部的是八時正。可是，最後一批遊行人士在五時四十五分已離開維園，因此六時半的時候，接近維園一段路已沒有人。因此，若採用這方法，較合理的假設，是每個行程兩小時，而非一個半小時，若採這假設，則同樣是兩輪而非接近三輪。

　　民陣召集人胡露茜會後被問及是否堅持五十三萬人和計算方法時說 ﹕ 「推翻（五十三萬）是無可能，我們有基礎計算出來的，由公眾來判斷。」她承認，民陣推算方法的不完善地方，包括遊行人潮的密度未能仔細地統計，有可能未達警方所指的最高密度的十七萬，但由於不同時段密度不同，故無方法準確地統計當日人數。

　　她強調，民陣是如實地報告計算方法，具計算基礎，他們經過討論後認為五十三萬是個值得可信的數字，是民陣的推算及評估，大家可作參考，但不同組織有自己的方法，市民有權選擇是否相信，民陣會參考其他方法。

　　被問及今次數字會否令公眾相信時，她則說 ﹕ 「信不信五十三萬由公眾決定，為何你覺得公眾一定不信，為何你們不去問公眾 ﹖ 不去問遊行人士 ﹖ 某些學者說廿萬是否等同準確 ﹖ 任何學者都有立論點，不能提供最準確人數，遊行可變數高，中間會有些出入。」

　　人權監察於遊行當日以人手點算出遊行人數為十六萬人，羅沃啟昨晚表示不想糾纏於人數問題。

Let me summarize this. The premise of the decision is summarized by the statement, "There is no possibility of overturning the 530,000." So the only question was how to get there. The problem that they faced is that the actual time frame for departing the march is only 4 hours, not the 5.5 hours that they originally assumed. They had assumed that there would be time enough for three waves of people taking 90 minutes each to finish, and they threw in an extra 20,000 for those who joined midway. Thus 170,000 x 3 + 20,000 = 530,000. Within the framework, there were two ways to revise the formula in order to achieve 530,000.

Plan A. There are only 4 / 1.5 = 2.67 waves of people. In order to get to 510,000, the density of the march must be increased from 170,000 to 190,000. Then 2.67 x 190,000 + 20,000 = 507,000 + 20,000 = 527,000. Done deal, except there is no reason to assume that the density is 190,000. If anything, the scholars are suggesting that the density was lower. It was time for Plan B.

Plan B. There were only 4 / 1.5 = 2.67 waves of people. Therefore, the number of marchers is 2.67 x 170,000 = 454,000. However, instead of 20,000 marchers who joined midway, there were actually 76,000 who joined in the middle. The total is now 454,000 + 76,000 = 530,000, same as before.

This new and improved 530,000 is said to be reliable, accurate, precise and all that. Dear reader, do you buy it?

There has been a reference to a "tree planting" problem given in elementary schools in Hong Kong. Here is the statement of the problem: suppose that you plant trees one meter apart in a row. If you planted ten trees, what is the distance between the first and tenth tree? The correct answer is 9 meters. The Civil Human Rights Front gave 10 meters as the answer originally, which would receive zero marks in an elementary school arithmetic class. But the Civil Human Rights Front is determined to get an answer of 10 meters. So they invented an eleventh tree and now the answer is 10 meters.

What will we tell our children? As for me, I am between laughing and crying. Most of the rest of Hong Kong, I suspect, are embracing the new and improved 530,000.

P.S. The unnamed mathematician pointed out that the the number of departures had effectively dropped to zero by 545pm, so that there was only a 3.25 hour window. The old calculation would have said that the march number is 3.25 / 1.5 x 170,000 + 20,000 = 369,000 + 20,000 = 389,000. This calculation reflects the real conditions on the ground. But by now you know what to do. There must have been 161,000 people who joined the march in the middle so that the total number of marchers is 369,000 + 161,000 = 530,000. This number is reliable, accurate, precise and all that, just as before. Blah, blah, blah. We must not let the truth change the magic number of 530,000.

Why did the Civil Human Rights Front feel that it cannot change the 530,000 number? There are two reasons.

First, the logically consistent number would habe been 2.67 x 170,000 + 20,000 = 474,000 because it corrects an arithmetic error while preserving all other previous assumptions. However, it is less than last year's 500,000 number and may reflect an erosion of support. This is off message.

Second, if they issue a press release that acknowledges the well-known arithmetic error but keeping the 530,000 number intact, the media stenographers may deem this not newsworthy and continue to use the 530,000 number from there on.

Actually, they were better than that. The decision had been made before the meeting anyway. On July 4, 2004, the Front posted this message on their website:

[translation] Concerning the problem about the number of marchers on July 1, the Civil Human Rights Front emphasized again: on that day, from the first arrival at the Central Government Office to the last arrival took 4 hours; the time to go from Victoria Park to the Central Government Office is 90 minutes. Therefore, there were about 3 trips. According to the estimate provided by the police to the Front of 170,000 per trip, adding people who joined the march in the middle, we estimated that there were about 530,000 persons.

就七月一日大遊行參與人數問題，民間人權線重申，當日以第一個到達政府總部至最後一個到達的人士共需時四小時；而由維園到政府總部共需約一小時三十分鐘，故共接近三轉。而每轉按警方對本陣線表示共十七萬人推算，再加上中途插隊人數，故推算出約五十三萬人。

This press release is phrased such that it can be interpreted to describe both the old and new calcuations. It says "approximately" three trips, which could mean either the 3.0 or 2.67 and it says "a number of people who joined the march in the middle" which could mean either the 20,000 or 76,000. In fact, there was no acknowledgement of any error, so this should really make the media stenographers happy.

On the same day, an editorial appeared in Sing Tao Daily.

(Sing Tao) The Civil Human Rights Front makes up the number; credibility completely destroyed. July 7, 2004.

[translation] When the Civil Human Rights Front announced a figure of 530,000 for the 7/1 march, it was criticized by various experts and intellectuals. Yesterday, the Front insisted that they were not wrong and they changed their calculation formula in order to reach the pre-determined number of 530,000. The Front did not have the moral courage to admit the mistake, and their announced number has lost crediblity totally. The only thing left to do is to set up a different estimation system for future activities.

When criticized by outsiders, the Civil Human Fights Front refused to even acknowledge that they made an error in elementary school arithmetic. Worse yet, they distorted the data. They were determined to reach the 530,000 figure and they customized a new formula to get that number. The result is that they will be subject to more challenges.

Think about it: the Civil Human Rights Front originally used the method of "planting trees" to estimate the number of marchers. The problem is stated as follows: suppose that you plant trees one meter apart in a row. If you planted ten trees, what is the distance between the first and tenth tree? The correct answer is 9 meters. The Civil Human Rights Front forgot to subtract the first tree and committed an error often made by elementary school students. This calculation error was pointed out many times by the electronic media since the first day and is well known by now. The Front did not come out and say anything. The error is known to everyone, from university teachers to elementary school teachers. But the new calculation yields a number that is identical to the original number. How can this be possible? How can this be justified to the citizens who went out on the street under the hot sun?

To refuse to correct the error is not the same as changing reality. Now that the Civil Human Rights Front is insisting on the 530,000 number, there are at least several results. One is that the related number has become a reference. Any other opinion based upon this number will be distorted. This is like cutting off part of your foot to fit the shoe that you bought.

Another problem is whether the media should continue to use this dubious and flawed number. On the day after the march, most foreign media such as The Asian Wall Street Journal did not use the 530,000 figure. The reports stated that the police estimated 200,000 and the Civil Human Rights Front claimed 530,000. They were very careful. Now that the number from the Civil Human Rights Font is obviously wrong and improper, the media are faced with the difficult choice of whether to continue to use the number.

By insisting on being wrong, the Civil Human Rights Front highlighted the problem of whether a number provided by organizers should be accepted. This creates the need to set up a trustworthy system of estimation in the future. In the past, every single event that involved more than 10,000 persons was subject to dispute with significant differences between police and organizer estimates. Some media have joked that a more reliable number is obtained by obtain the average of the two numbers. This is unscientific, but it reflects the difficulty that the media faced in reporting the crowd size.

The dispute over the 530,000 number does not involve the demands being made the marchers. No matter whether the number is the 200,000 estimated by the scholars or the 530,000 cited by the Front, the opinion of the people will be respected. In a quality society, any person's opinion --- as long as it is fair and just --- should be respected. Confucius says that when it comes to a matter of principle, one would go regardless of the number. Whether the march had 200,000 or 300,000 people is immaterial. The reason that the intellectuals are firmly pursuing the issue of the estimation of the number of marchers is that they insist on and they respect the facts. Some other pro-democratic persons also offer other estimates and methods for the same reasons. We hope that the Civil Human Rights Front will recognize their error.

　　民間人權陣線公布七一遊行人數有五十三萬人，受到各方專家和知識界質疑後，昨日仍堅持錯誤不改，不惜改變計算方式來遷就五十三萬人這個預設數字。民陣沒有認錯的道德勇氣，令其公布的遊行數字公信力全失，惟今之計只有改轅易轍，為日後的活動另訂評估機制。

　　民陣在備受外界質疑下，連犯了小學算術錯誤也拒不認錯，最壞的是為了扭曲事實，在先有五十三萬的數字後，才度身訂造數據的做法，結果只會惹來更多質疑。

　　試想想，民陣本來用計算植樹數目的方程式去評估遊行人數，問題一如是每隔一米種一棵樹，種十棵樹的長度是多少米 ﹖ 由於頭一米要種兩棵樹，答案應該是九米，民陣一時大意沒有扣除最先一次，犯了小學生經常出現的錯誤。這個計算方法在當日電子傳媒已經多番發表，眾所周知之後，連日來也沒有澄清，有關錯誤無論是大學講師或小學教師都一清二楚，但現在新的計法，得出來的答案竟然和原先沿用錯誤計法的答案一樣，這種邏輯上的矛盾猶如偷看答案的小學生，用錯誤算式計出正確答案一樣，試問又怎能自圓其說 ﹖ 這種做法又對得起在酷熱下上街的市民嗎 ﹖

　　拒絕矯正錯誤，不等如可改變事實，現在民陣堅持五十三萬的數字，至少產生幾個後果，一是有關數字成為一個基準，今後作出評估時，雖有繼續扭曲，令人削足適履去遷就。

　　另一個問題，就是傳媒應否繼續使用這個可信性存在極有缺憾的數據。從遊行翌日，大部分外國傳媒如《亞洲華爾街日報》，都沒有採用五十三萬的數字，報道內容亦申明數字由警方的二十萬到民陣的五十三萬，取態十分審慎。現在民陣的數據明顯犯錯而不更正，更令傳媒陷入應否繼續採用的兩難局面。

　　民陣堅持錯誤，突出了遊行組織者點算人數的認受性問題，產生另訂有公信力機制評估人數的需要。過去每年都因各種議題有逾萬人的遊行和集會，有關參加人數統計不易，警方和主辦者每次估計都大有出入。有傳媒中人曾經笑談較可靠的數字可能是把二者加起來取個平均數 ﹔ 這個「平均計算法」並不科學化，但也反映了傳媒在報道遊行人數時面對的困境。

　　五十三萬的數字爭議，完全不涉及上街市民的訴求，無論是學者估計的二十萬還是民陣的五十三萬人，他們的意見都值得尊重。在一個有素質的社會裏，就算是一個人的意見，只要是符合公平、公義，都值得尊重，孔子說遇到值得堅持的事，「雖千萬人吾往矣」，所以多了少了二、三十萬人，對遊行不是最重要的因素。知識界近日對評估遊行人數的執  ，只是對事實的堅持和尊重，部分民主派人士提出其他評估數字和評估方法，也是基於相同原因。迷途知返，我們仍希望民陣能再三反省，還歷史一個真相。

The following opinion column by Chai Shiu-fung appeared in Sing Tao Daily on July 8, 2004.

[translation] The Civil Human Rights Front held a meeting last night and decided not to change the 530,000 participant figure for the 7/1 march. They changed the formula, skirted around the arithmetic error and calculated "anew" the 530,000 figure. Thus, they denied the 170,000 to 260,000 figures obained by scholars at various universities. The Front gave an astonishing display of how politics overpowered scholarship.

Chai Shiu-fung has spoken to various people about this hot subject. We all believe that they ought to count whatever was there and not to make it as large as they want. We all shake our heads and sign at what the Front did. We felt that they turned a good thing into a bad thing. A well-place person in politics also offered another viewpoint that I want to share with everyone.

He believes that the Front magnified this year's number to make it larger than last year's. If it were really true, then that is what it is. If they exaggerated, then they have not thought that this Great Leap Forward may bring negative consequences.

There are many provinces and cities in mainland China which disagree with the gentle way in which the central government has handled the Hong Kong problem. Hong Kong does not contribute taxes, they have no military expend