About this Blog

The purpose of this blog is for my personal use. It serves as my personal diary as I investigate Chinese internet/gaming companies for investment purpose. If you have any comments or disagreement, please give me feedbacks.

Wednesday, May 30, 2007

Sohu – TL – TL’s Peak Concurrent User estimate

This is the final part of 4 part series. The 3nd part is here:

Now we have all the data we can use to provide some estimate for TL’s PCUs. I will be doing this through two completely different approaches. The first approach is by comparing to its peer. The second approach is by extrapolate from its server counts.

Approach A. Estimates from peers.

The idea in this approach is by comparing with it peer’s PCU, we can get an estimate of TL’s PCU. In this case, peer means a MMORPG game that just started open beta.

We can’t compare it with an existing game that had been in operation for years because relationship between the number of downloads and the number of searches to PCU is different between new games and older games. For a new game, there will be more downloads simply because it is a new game.

In this case, NTCY’s SUN provides a reasonable base of comparison because SUN is a new game. From the data in the last two articles, we can formulate the following table:

________iResearch___Baidu____Xunlei Download___ Xunlei Search

For this study, I will ignore the last two columns (Xunlei download and Xunlei search). There is no way TL has 4 times as many users as SUN.

Xunlei’s one day numbers seem to be skewed. A possible reason is that since these are one day numbers, they may be subject to wide fluctuations. (As a matter of fact, Xunlei’s one day numbers always seem to have problem with accuracies.)

Now, if we take the result from the iResearch (which is really Xunlei’s one week statistics) and Baidu column and average them out, we would have gotten a combined TL/SUN = 155%.

Since NCTY proclaim that SUN has a PCU of 400k, we can estimate the PCU for TL to be 1.55 * 400k = 620k.

Approach B. Estimate from the number of servers.

As of now, TL has 159 servers. Each server’s maximum capacity is 5200 players. Sohu had been adding servers like crazy. It means tremendous new users. But after more than one month, I believe it is able to support almost all users now. Therefore, I estimated that a small number of servers still have servers that are almost full (30) while the majority has a very good balance of users in them (90).

But since Sohu is adding servers so fast, it also means that some older servers may neve get the chance to get fill up. I picked an estimate of 39 servers that have numbers that are lower than expected, but not low enough to warrant server consolidation.

Thus, the calculation is as follows:

30 servers @ 4500 players per server
+ 90 servers @ 3000 players per server
+ 39 servers @ 2000 players per server
= 483,000 players.

Or PCU = 483k.

Clearly, this is a highly subjective estimate. By varying the numbers, you could get very different results.

Now, if we combined the two numbers (just average them),

We get a PCU = 552k

This is a truly amazing number from a game nobody saw it coming. Well, I predicted TL will be a successful game. But I was thinking about a mid-level game that can give 5M per quarter surprise to Sohu. But a PCU of 552k for open beta is indicating something completely different.

PCU like this for open beta is un-precedent. It is indicating a major blockbuster in the making.

In summary, in this series, I pointed out 1 possible inconsistency, I dispelled 2 myths and I gave 5 reasons for a strong TL. In the end, the evidence is just too strong in one direction. Hopefully, I answered the question why I think TL is on his way to be a blockbuster.

Tuesday, May 29, 2007

Sohu – TL – Reasons I think TL is popular

This is the 3nd part of 4 part series. The 2nd part is here:

Reason 1: iResearch said so
See the link: http://www.iresearch.com.cn/html/online_game/detail_views_id_44137.html
It is the cumulative number of downloads between 5/14 and 5/20 as compiled by Xunlei. There are 571k downloads of TL. TL is ranked 3rd behind two casual games. Thus, it is number 1 in MMORPG games. Netease’s XYQ is number 4, Zhengtu is number 5, WOW is number 6 and NCTY’s SUN is number 8.

Reason 2: Xunlei’s download ranking said so
See the following for Xunlei game download ranking for the previous day:
TL is number 1 with 36688. SUN is number 4 with 9292. XYQ is number 7 with 6240.

Reason 3: Xunlei’s search ranking said so
See the following for Xunlei game search ranking:
TL is number 3 behind two casual games.XYQ is number 4, Zhengtu is number 6, WOW is number 7, and SUN is number 14.

Reason 4: Baidu Game ranking said so
See the following for Baidu’s top 50 game search ranking:
TL is number 2 and is sandwiched between the two casual games. WOW is number 4 and SUN is number 5.

Note that Xunlei is a download accelerator and a P2P program. Sohu’s biggest focus now is P2P video (with download accelerator function provided for free). Thus, they are direct competitors.

In addition, Xunlei is allied with Google. Google is Sohu’s direct competitor because Sogou is Sohu’s property. For the same reason, Baidu is direct competitor to Sohu also. There is no reason that Xunlei and Baidu will do TL’s bidding by hyping TL’s popularity.

In addition, this is not something you can fake. We are talking about half of millions of downloads or searches. If all these downloads or searches are coming from a few thousands IP addresses from a few cities, alarm bells in Xuelei and Baidu would ring of possible fraud. I am sure Xuelei and Baidu would lick their chops in sticking it to Sohu without mercy if that is the case.

Reason 5: It (faking server numbers) could kill the game
To me, this is the biggest reason why I think TL is popular. There are 159 servers. TL can’t fake it. If there are phantom servers, players would know since they can’t get into these servers, nor can they register with these servers. Assuming Sohu is faking it, Sohu could annouce that these servers are in maintenance or full. But how can they be full or in maintenance all the time. Players will get awfully suspicious if any server becomes full on 4am every day.

Similarly, if there are very few players in the servers, players will notice also. A server can handle 5200 players. If there are only 500 players during peak time, players will notice and complain because it would be no fun to play in a sparsely populated game. It would be awfully easy to spot in various game forums.

But there is a possibility that the sever only have one or two thousand players during peak time. Players might not complain and TL gives people the impression that they are 2 or 3 time more popular than they really are.

But I don’t think it is the case. By doing that, they could kill the game completely.

The reason they potentially want to inflate the server number is to create hype to attract players. But if they start to adds new servers when the old servers (in the same geographic location) only has 1000 or 2000 players, they will start to create ghost servers.

Ghost servers are naturally created event for MMORPG. Netease’s XYQ, XY2, and DaTang consolidate servers to get rid of these ghost servers. Zhengtu consolidates ghost servers almost every two weeks.

There is a threshold on the server that when the number of player goes below that, that server starts to lose players quickly. As more players leave, it becomes less fun for existing players. It then becomes a vicious cycle.

This is not something game companies would do. They would do anything to get rid of ghost servers. They definitely won’t do anything to help create many ghost servers.

Tomorrow I will conclude this series by providing my calculation for TL’s PCU.

Monday, May 28, 2007

Sohu – TL – Myths that TL is not popular

This is the 2nd part of 4 part series. The 1st part is here:

Before I went further, I just want to say that I created this article on 5/25/2007. Therefore, all the data supplied are up to the date of 5/25/2007. As of today, TL has 159 servers.

Myth 1 – Why does the Sogou Game ranking ranks TL so low?

Well, they rank them so low that you can’t even find it on Sogou’s top50 game search ranking:
Actually, the answer to this is very simple.

If you go to the Sogou top50 book ranking:
There is it, number 3 right behind “adult novel” and “Love novel”. There is no way TL the book is so popular. TL is popular not because of the book, but the game.

The lesson from this is not TL is not popular, but Sogou is still immature in classifying phrases. Another lesson is Sohu’s various divisions are very independent. Most times, I got the impression that there are very few coordination between the divisions.

One thing is for sure; there are is no vast Sohu conspiracy to lie about TL’s popularity since both Sogou and TL are Sohu’s property.

Myth 2 – Why does 17173.com rank TL so low?

Again, another Sohu’s property seemingly dissing TL! From the following 17173 game download ranking:
TL is down all the way to number 211!

I will give a possible reason.

If you go to TL’s download site:
You can see the latest version is dated 5/10/2007 and the file has 471MB of data. But if you go to 17173’s download site for TL:
You see the version is dated 4/28/2007 and the file is 456.15MB. 17173’s file is a couple of versions back.

Actually, it doesn’t matter whether you start out with 5/10 or the 4/28 version, during the installation process, it will auto-patch to the latest version anyway. But new players don’t know that. Given the choice, they download the latest one. They either got it from Xunlei or from the official site, or other places that has the latest, rather than from 17173.

The answer is really simple. 17173’s game ranking ranks by the number of downloads. TL has low ranking because gamers don’t download TL from 17173. They don’t download from 17173 not because TL is not popular but because 17173 doesn’t have the latest file for TL.

Again, no vast Sohu conspiracy to lie about TL’s popularity since 17173, Sogou and TL are all Sohu’s property.

Tomorrow, I will provide 5 reasons why I think TL is extremely popular.

Sunday, May 27, 2007

Sohu – TL – Is its server counts for real?

My previous post on TL is here:

The author of the intelligentinvesting blog had the following comments:
“Please explain why you are so certain that Sohu's TL is a blockbuster. I am still not very sure about that right now even though I really wish TL will be a big hit. One concern I do have is whether Sohu's claim of TL operating 145 servers is credible. Yes, TL is ranked top on Baidu's most searched online games. But it is not ranked at all on Sogou's website(http://www.sogou.com/top/class19_1.html). It is also not ranked very high on its own 17173.com webiste or xunlei.com, another popular online game portal.”

As the matter of fact, the same concern bothers me a great deal until I looked more into it. In the next series of articles, I will dig deeper into it. By the end of this series, I will provide my estimate for TL’s PCU (Peak Concurrent User).

In today’s article, I will talk about one possible reason why TL’s server numbers are just marketing hype. In tomorrow’s article, I will provide many reasons why TL’s number is for real. In the third and concluding article, I will provide my calculation for TL’s PCU estimate.

The case for caution:

I can only see one symptom that might arouse caution. Why didn’t Sohu published its PCU number if TL is so hot? It makes no sense. Game companies want to hype its games to create momentum. They hype their games by announcing PCU numbers. Average gamers understand PCU numbers. They don’t really care about the number of servers.

One possible reason could be that they are holding the PCU number back until commercial operation to give maximum PR impact.

Or one could say since Sohu is a public company, they don’t want to give out false PCU numbers.

As of 5/26/2007, they have 159 servers. I believe these are 159 physical servers for TL. But each server might only have 500 players playing rather than 5000 players playing. This way, they are not lying about the number of servers, but they also accomplish the purpose of giving people a false impression that TL is extremely popular.

Well, there are major holes for arguing pro or con of why Sohu didn’t publish its PCU number. But either way, it is very unusual why a company won’t publish PCU number for a popular game.

Tomorrow, I will dispel the myths that TL is not popular. After that, I will provide 5 reasons that that indicates TL is popular. In part 4, I will provide the calculation of TL’s PCU.

Sunday, May 20, 2007

NTES - XYQ - server count

My previous post on Netease's XYQ:

Tomorrow is Netease's earning conference. For those of you who actually check for server counts, just a friendly reminder that XYQ is in the mist of server consolidation.

Therefore, about 40 to 60 (I think it is 49 servers) that are being consolidated. These are the ghost servers. They are usually in yellow or green (little or no players), but are now pink or red (many players or completely full). In addition, they are being double counted.

In a nutshell, there are about 49 extra servers that are listed as full.

These newly created servers were doubled counted to ease the player transition. But they also let the server count looks a lot more impressive than reality.

Wednesday, May 16, 2007

Sohu – TL – Start charging money during open beta

My previous post on TL is here:

TL was a successful story in Sohu. Sohu gives TL’s chief designer (hired away from Kingsoft) plenty of time and resources to develop the game right. It is a game that will make a lot of money for Sohu investors.

But 2 things happened recently that might cast a dark cloud over it. I won’t talk about one because I need to investigate it more. But I will talk about the other now.

I think Sohu is starting to charge players now. First, its final stage of “close beta” is really an open beta. And now they are treating the open beta as commercial operations?

When in beta stages, players are not supposed to pay!

From the following official site:

TL had started to charge money for in game items. There are four places you can purchase the game card:
1. Internet Bar
2. UDPay
3. Convenient Stores, Supermarket, and Newspaper Stand
4. Through Sohu’s Online Store: http://esales.sohu.com/online/tlcost/index.jsp

To purchase in-games items, go to the online stores (where you can buy goodies).
Not a lot of stuff yet, but I am sure they will add a lot more goodies as time goes by. Also, players who killed too many other players can lessen the punishment by paying game currency.

Some players had already started buying the game card already.

At this point, the game is still growing like crazy (nothing I ever seen before). I am pretty certain this game is going to be a blockbuster.

In the just concluded 1Q earning conference, Sohu said that there won’t be any revenue from TL in the 2Q. Now, it looks like there might be some significant revenues in the 2Q from TL after all.

I am sure it will make short term investors very happy. But is it necessary for Sohu to do this? It is awefully short-sighted of Sohu.

Friday, May 11, 2007

Earning Season

It is the earning season the next 2 to 3 weeks. This blog is just a hobby of mine and my time is limited. Most of the free time in the next few weeks will be in listening in on the earning conference calls and reviewing what is being said.

I won't talk about anything being said in the earning conferences since there it won't provide anything extra.

Therefore, I won't update this blog as often in the next three weeks.

A lot of people had asked me for an estimate for the revenue for Sohu's new game TL. I said it might be 5 million per quarter a couple of months ago. Given what had happened in the last two months, it is clear that estimate is way too conservative.

I am in the middle of doing a series of articles on turmoil happening in China's gaming industry, the new Free-to-Play game model vs the existing Paying game model. By the end of that series, I will have a new revenue estimates of TL.

I haven't work out all the numbers yet, but it looks like my TL's revenue estimates could be many times of my original estimates. I also need to recalculate the numbers to make sure I didn't add a 0 somewhere.

Anyway, for the people who is waiting for that article, I am still working on it. But it might be a few weeks later.

Have fun with the earning season!

Thursday, May 10, 2007

SOHU – TL – Open Beta

My previous post on TL is here:


TL just keeps on surprising me. From the “Final Closed Beta” (which is really the open beta) on 4/4/2007, it started out with 16 servers. It continues on growing and went all the way to a total of 103 servers on 5/8/2007. Most games would have started to peak after few weeks. It is doubly surprising considering that NCTY’s SUN went into open beta on 4/18/2007.

But what is truly surprising happened yesterday. Yesterday (5/9/2007) is the date of the official “Open Beta” for TL. Netease planned to add 10 more servers. But the servers were overwhelmed. Netease eventually added 24 more servers in one day. See the following link:


It is not surprising that a game has big jump in the number of game players for an open beta. But it is surprising that a game has big jump for a second open beta.

As of today, TL has 127 servers. With the maximum capacity of each server at 5200 players, I would conservatively estimate a PCU = 550k.

One thing that bugs me is that Sohu is not advertising its PCU number. Usually, game companies would have broadcast the PCU number to build up the hype.

But I am pretty sure this is a very popular game. There are a lot of players everywhere you go and players were still complaining to being too crowded and asking for new servers.

Maybe Sohu did change to a low-key company. The Sohu of a year ago would have the CEO and other senior management in the news pushing the story. Well, the CEO of Sohu is climbing the high mountains for Sohu's Olympics events. Maybe that has something to do with the low profile.

It is interesting to see if TL will just keep on going up. I do expect that they will add a few servers before this weekend. But the key to see if TL is still growing is if they are still adding servers the next few weeks. Another key point is if they will lose a lot of users after commercial operation (free-to-play games don’t usually lose as much players as Paying games right after commercial operation).

From the game players comments and game developers reaction, there are less and less bugs, and TL doesn’t have a lot of bugs to beginning with. TL is starting to test out the pay-for-item features (by giving players free credits).

Unless there is a big hidden design flaw (which is unlikely), I can’t see this open beta last more than a month.

Wednesday, May 9, 2007

Ntes – TX2 – Progress 3 – What shall NTES do?

My previous post in this article is here:


From the following article:


Netease had changed the chief designer of TX2. See also the following article for about the same news:


The chief design has a personal blog, it is on http://blog.163.com/gao_bell_ring/ His name is Gao and he is a 29 years old young man.

I had been reading his blog for a few months to get some insight. It is a pretty good read. His blog gives some hints about Netease the company. It confirmed my impression that Ntes is a technology company that really treasures talents. It gives tremendous amount of responsibility and trust to its legends of lead game designers.

The chief designer of a game has a lot of authorities. Sometimes, I got the impression that they regards Vice-presidents’ and CEO’s directions as suggestions rather than orders.

Ntes may have over-done it in this case. Sometimes the game designer wants to develop the most revolutionary or innovative game rather than the most profitable game. The game designer wants to build the most artistic game rather than the game most gamers want to play. A game other game designers would woo and aah, but also a game that can’t keep players.

After reading his blog for the last six months, I do get the sense of knowing the guy (of course, I don’t). He is very talented and artistic. He has a sensitive soul. He is also supremely confident. He seems to have a fragile ego that he doesn’t take criticism well. He is so confident in his grand vision that game players will be in awe once he finished TX2’s development. In his mind, TX2’s full potential is still many months (or year) away.

He believe that as long as he build it, they (the game players) will come. He is supremely talented, but somewhere along the way, he starts to deviate from what the gamers want. Let me give an example:

Gamers had been complaining about the background music in one of main playing area on the game map. These music come from the true ancient Chinese music. The chief designers spent a lot of resources to get historical and musical experts to duplicate these music. They modify some for modern time. They do capture the essence of a time so ancient that most players are definitely not familiar with.

Most players hated it. But the chief designer believe in it and the music stays.

An analogy would be for Lord of the Real Online uses Shakespearean English as the mean to communicate! Or to make Beethoven or Mozart's music as the background music for Harry Potter's game.

It maybe truly be a master piece, but it doesn’t satisfy the popular taste.

Something drastic must had happened for Netease to do this. Most other companies would have try to get another guy in charge while making the current design lead the deputy. Maybe the egos involved are just too big for them to coexist together. But it will lose whatever continuity for the game development.

TX2’s Chief designer is probably not feeling very well now. His baby for the past 4 years had been taken away from him. But I don’t feel sad for him. I am sure by now, CEOs of NCTY, SNDA, and 10 others game companies had already left messages on his answering machine to beg for appointments.

TX2’s chief designer and his core team were the one who developed XYQ. The rumor is Netease uses its XY2’s core team to take over (it is just a rumor and I can’t confirm it).

Something this drastic has to cause cascading affect to the whole company. We are talking about basically the two best design teams Netease have.

This will not just affect TX2, it will probably have a negative affect on XY3 as some of its designers were shifted away to TX2. I am certain most other new major games will be affected as well.

At this point, Netease is probably trying to figure out what is the next step for TX2. If I were Netease, I would not have a regular open beta for TX2.

A regular open beta usually last for 1 to 2 months. That would be just too soon. I wish TX2 could be revamped, not just patched.

As I mentioned many times, TX2 (even if successful) would just be a niche product. A game with so much emphasis on gang warfares won’t ever be a true blockbuster. What TX2 aught to be a WOW clone.

As I mentioned before, there is a great need for a WOW-clone game with the only difference of that this game builds around ancient Chinese mythology. Make TX2 be that game. Build lots of quests. Put more relationship into the game.

Doesn’t need to down-grade the existing PvP features (they are unique and shall be a selling point), just don’t make them the only selling points.

WOW’s expansion pact is rumored to be out in China at the end of June. It would be suicidal for TX2 to compete head to head with the Burning Crusade.

Give TX2 and its new chief designer 9 to 12 months to revamp the game.

Yes, it would probably cause short term pain on Ntes’s stock price. But I invest for the long term, and I hope Ntes ship herself right again.

Oh, Ntes, how about start marketing for TX2 now. Sohu started marketing for TL more than a year ago. NCTY starts marketing for SUN long time ago also.

Ntes, please forget about these stock buybacks.

Tuesday, May 8, 2007

Ntes – TX2 – Progress 2 – Development Direction

My previous post in this article is here:


There is another TX2 movie that can be downloaded:


Right click and save as to download the file. You would need winrar to unzip.

It is another fan made movie. It has a sappy plot (good guys got killed, bad guys got killed. Good guys become bad guys become good guys become animals become good guys). But all the graphics are coming from in-game graphics. It is not specially produced trailer made by professional computer graphic artists. It is made by ordinary game players. It has to be captured and then compressed. Those two steps degrade the graphics. But even with that, as you can see, the graphics is absolutely gorgeous.

Next, let me talk about some of the major playing features the game designers had added.

In every sense, Netease is adding tremendous amount of new contents every week during the open beta. It allows gangs/tribes to conduct castle sieges. At the beginning of the open beta, there were 2 cities. Now, there are 5 cities. Eventually, each server would have 9 cities.

After the gang/tribe captured a city from enemy gang, it can start to build out the city. The game developers were busying in adding the amount of goodies that can be build in a city.

In addition, Netease adds another Player vs. Player (PvP) feature, gang warfare, a couple of weeks ago. While castle siege is the domain of major gangs (with member greater than 200), gang warfare allows minor gang a chance to fight it out with a major gang. While city siege is limited to a city and its vicinity, gang warfare is conducted spreading out all over the whole map. Therefore, it allows for smaller gangs a chance to prevail by conducting guerrilla warfare.

Couple of weeks ago, Netease came out with a very unique way for individual PK (player kills). It groups all users into three groups, sheriffs, bandits, and regular guys. Players become bandits after they killed a regular. The sheriffs can kill bandits and get the stuff bandits have. On the other hand, bandits can also kill sheriff. Bandit would only be penalized after they killed regular players.

I will have to say that this is truly unique. Unlike where WOW or most other games that allows unlimited PK only on special servers, TX2 gives its game a rich environment much like ancient world where fights breaks out from time to time while most innocent players were left alone.

Some of these contents were really unique. Big gang can fight it out with another big gang in castle siege. Unlike the severe player limits of 40 for WOW (War of Warcraft) and 12 for Guild War, TX2’s castle siege has no upper limit (it has some severe side effects, but that is another story for the future). Every day, there are battles involve 4 or 5 hundred players battling out each other.

In addition, smaller gangs can get into the act by doing gang warfare. For individuals, they can do PK with other players either as a sheriff or as a bandit.

As you can tell, TX2 development team had been busying in adding PvP features.

There were flaws in TX2. Actually, there are a lot of flaws. Most of them are in PvE (Player vs. Environments) area.

For one, there are insufficient help for the new users. There really is no official user guide on its official web page. There are some “unofficial guides” written by game players. But they are written during closed beta and are not up to date.

Frankly, it is not a big deal. Write some users guides and add more dialogs for NPCs (Non Player Characters) to give more guidance to newbie are really not that time consuming.

After the first 10 levels, the players are no longer the newbies. For the first 30 levels, the main storylines are rich, fun and rewarding. But after level 35, it becomes very boring. The main story lines are not finished. All that players can do is to kill monsters.

The main emphasis of the young chief designer in on PvP features. Even something as important as the main storyline are assigned to lower priority. Simple things like more help for new players are left unfinished because it has lower priority.

In addition, the economic systems is incomplete. The relationship system is incomplete.

Rich relationship features are the fortes of all existing Netease’s games. If a player doesn’t want to fight, he may become a doctor who goes out to collect herbs to make into medicine. He may marry another player, etc.

There is none of that in TX2 yet. I am sure these features will eventually be part of TX2. But they are just not the priority to TX2’s design team.

In addition, as I predicted in my previous articles:




Some server becomes dominated by a gang. The gang becomes so dominate, no other gangs can compete on that particular server. As a result, that server starts to lose players.

Finally, as I mention many times, TX2 is a high resource game. Players who have old computer is not going to have a good time in TX2. I thought they improved a lot on this front. But this is going to be a never ending quest to optimize the game code.

So, in a nutshell, that is what happens to TX2:

Netease won’t provide marketing for TX2. There is little new players to start out with. The game is not friendly to new players (in terms of documentation and other things that only require minor changes). For those who is not part of a gang or for those who don’t want to PK (Player Kill) all the time, these are very little reason for them to be there. Even for members of the gangs, there is little reason for them to hang around after their gang loses the war.

Next, I will talk about what I hope Netease will do to fix TX2.

Monday, May 7, 2007

Ntes – TX2 – Progress 1 - Status

My previous articles on TX2 can be found here:







On April 30, 2007, NTES provides the following news: see link:


NetEase said its Tianxia II research and development team will work on making the game easier for beginners and improving group play and hardware functionality. They are going to add another phase of open beta.

It sounds somewhat minor. Adding another phase of open beta usually mean another 1 to 2 months of open beta.

But I think it is going to be a lot worse than that! For one, you probably had heard by now the lead designer had been reassigned. Many of the group leads had changed.

But before I talk about it, I would give some background first. Today, I will talk about the progress of the game to this date.

So far in this stage of open beta, TX2 is still losing players, although at a very slow pace.

The marketing is starting, although at a very slow pace. From 4/1/2007, they put commercials on the bus in six cities (not 1st tier cities, only six second tier cities) , see the following:


From 4/6/2007, they put ads in the subway in a southern city. They planned to expand it to all other cities and in bus stations, subways, and convenient stores. See the following:


During the last week of April, they started to put ads on some internet sites and applications such as Xunlei. But there is still no ads on 17173.com yet.

Of course, given what happens now, the current marketing program probably won’t be expanded.

In some sense, Netease had never given its young chief designer a chance. With so little marketing, TX2 is not going to get too many game players.

Ntes has so much money, they would spend 200m to buy their own stock, why are they pinching pennies in marketing their flagship game?

On the other hand, from the way Netease changes course in DaTang and TX2, it is clear that Netease has certain internal control. Its games probably have to pass certain performance criteria before the game can go to the next stage. Maybe Netease checks the percentage of the number of player (who started the day of open beta) remaining after 1 month over the number of players the day of open beta.

Regarless of what performance criteria Netease uses, it allows Netease to do mid-course correction to salvage the game.

In the end, I think it is fault of both Netease and TX2’s chief designer. Netease didn’t provide its young chief designer enough marketing support. On the other hand, TX2’s chief designer designed an innovative game that game designers (but not game players) like.

The game is so far capped at level 60. Therefore, with virtually no marketing, there is little new users. As the players get to level 60, they don’t have much to do (except those who are in a gang and is involved in gang battles), older players start to leave.

The game developers were adding a lot of contents every week. But it is also clear that the game is far from a finished product. Compare to another recent game, Sohu’s TL, TX2 has much less playable features.

Tomorrow, I will talk about which direction TX2’s development team had gone so far in the open beta.

Thursday, May 3, 2007

Sohu – TL – Game Prediction

My previous post on TL can be found here:


I had always have the idea that Netease is the best game company in China by far. But it is also a company like Apple. Just like a big fat apple, ready to be picked by a company like a Microsoft.

Netease is a great technology company. But it has so many flaws. The company has a distaste on marketing bordering on psychosis. It would try to save a few millions on marketing so it can spend 200 or 300 millions to buy back its own stocks.

It won’t expand its market share, it won’t buy companies or technology and it can’t do more than one thing at a time.

But how to be successful to be an online gaming in China? Easy, just follow what Netease do.

Kingsoft tried to do that. But Kingsoft has an overbearing CEO who think he knows everything.

I had been following TL for more than 6 months now. Sohu gives TL more than 100 developers and 3 years of time and ample of room for the developers to do their things.

It is also clear that Sohu will keeps on devoting resources to continuing updating TL.

In a nutshell, Sohu is following Netease’s formula to a tee. But Sohu definitely don’t have Netease’s distaste in marketing.

There is a difference in how to be a good paying game vs. free-to-play game. To be a good free-to-play game, you want the players to feel somewhat comfortable, but not too comfortable. It shall be somewhat fun to play for the players so they want to play the game. But it can’t be too much fun unless they pay for the in-game items.

TL seems to have the right balance. It allows for very liberal PK (player kills). There is definitely incentives for players to buy powerful swords, etc. so they can hold their own in PK battles. It allows for players castles. For gangs who want to develop their own castles and gangs, there are many things to buy. It will definitely be costly for the gangs to make their castle invincible.

I am certain that their recent experience in converting Blade Online from a paying game to a free-to-play game helps.

I used to think Sohu’s CEO could be somewhat overbearing. But Sohu, as a company, seems to slowly changing in the last 6 months. It is starting basic research in search. It build a lab (Sogou Lab) to do research on the internet. It is slowly mutating from a basic text-based internet company to a multimedia company. It has a game that will give Sohu a great run in the next 4 to 5 years. Dr. Charles Zhang seems to mutate from a CEO who gets into every details of the company into a CEO who gives out grand strategic vision while his vice-presidents do the day-to-day. Maybe the promotion of Dr. Yu Gong to COO had convert Sohu from a company of short-term profit (a company who would hide spyware in its software) into a company of long term vision.

Everybody seem to be so excited about next year, the year of Olympics. For me, I am more excited about 2009 and beyond when Sohu and Sina will mutate to completely different companies.

That mean there will be a lot of expanse on R&D in the next two years to prepare for the future. But for Sohu, there are the TL and Olympic to absorb the cost.

I guess that is the difference between me and almost everybody else. I see the Olympic and TL as merely the stepping stone for companies like Sohu and Sina to become the dominant multimedia companies of the next few decades.

That will also means they might disappoint the stock market as the margins fall. But they are not spending money out of the weakness, but out of strength. They are not spending money so they can keep their market share.

They are spending money so they can be one of the few dominant multimedia companies. It will allow them to be a dominant multimedia internet company. It would give them tremendous leverage in the 3G mobile phone’s content battle. It allows them into making new shows and TV productions. They could follow Sina and try to be the platform of choice for legal music and video contents (in case China does come down on piracy). They will also continue to invest in search.

For those who keeps on asking me of my prediction of what will the stock price be. I don’t give those out for the simple reason that I don’t really know myself.

I am a fundamental investor and I won’t do options. I try to find as much company information as I can. But I don’t really know how to put all these information into stock price prediction.

But I do know that TL will be a surprised hit.

Wednesday, May 2, 2007

Sohu – TL – Game Progress

My previous posts on TL can be found here:




Also, for much more detailed coverage on TL, the following blog site seems to do a better job than I can: http://intelligentlyinvest.blogspot.com/

Anyway, let me get to the point.

I was going to report on the horrible news on TX2 of Netease. It is not a minor course correction. Adding another phase of open beta usually mean another 1 to 2 months of open beta.

But it is going to be a lot worse than that!

I was going to go through the game progress on all of Netease’s games. But with the Sohu’s latest earning conference call, I will just spend two days on Sohu’s TL.

On 4/4/2007, TL went into the final stages of “closed beta”. But since it allows anybody to play and the player stats would not be erased, it is really an “open beta” except in name only.

On 4/4/2007, Wednesday, it starts out with 15 servers, it adds 16 more servers later on. It has a total of 31 servers.

On 4/5/2007, Thursday, it adds 3 new servers, It has a total of 34 servers.

On 4/6/2007, Friday, it adds 8 new servers. It has a total of 42 servers.

On 4/7/2007, Saturday, it adds 6 new servers. It has a total of 48 servers.

On 4/9/2007, Monday, it adds 4 new servers. It has a total of 52 servers.

On 4/10/2007, Tuesday, it adds 6 new servers. It has a total of 58 servers.

On 4/11/2007, Wednesday, it adds 4 new servers. It has a total of 62 servers.

On 4/12/2007, Thursday, it adds 2 new servers. It has a total of 64 servers.

On 4/13/2007, Friday, it adds 5 new servers. It has a total of 69 servers.

On 4/14/2007, Saturday, it adds 5 new servers. It has a total of 74 servers.

On 4/15/2007, Sunday, it adds 2 new servers. It has a total of 76 servers.

On 4/16/2007, Monday, it adds 3 new servers. It has a total of 79 servers.

On 4/17/2007, Tuesday, it adds 2 new servers. It has a total of 81 servers.

On 4/18/2007, Wednesday, it adds 4 new servers. It has a total of 85 servers.

On 4/20/2007, Friday, it adds 1 new server. It has a total of 86 servers.

On 4/21/2007, Saturday, it adds 2 new servers. It has a total of 88 servers.

On 4/24/2007, Tuesday, it adds 2 new servers. It has a total of 90 servers.

On 4/25/2007, Wednesday, it adds 2 new servers. It has a total of 92 servers.

On 4/27/2007, Friday, it adds 5 new servers. It has a total of 97 servers.

On 4/30/2007, Monday, it adds 4 new servers. It has a total of 101 servers.

On 5/1/2007, Tuesday, it adds 2 new servers. It has a total of 103 servers.

As can be seen, it has even more success than what I expected. (and I was expecting very good success). It hits a little rough spot on 4/18/2007, the date that NCTY’s SUN went into open beta. But it seems to start growing again.

The game is developed by the Firefox branch of the company. It takes about 100 developers 3 years and about 100 million RMB to make.

At this point, the server can hold 5200 players. Therefore, I would estimate that TL had already achieved a PCU of 450k.

Also, it is official now, TL will be a free-to-play game. See the link:


As I mentioned, it spells good things about TL. TL can’t compete with the likes of TX2 or WOW. But as a free-to-play game, it is going to be a very successful game.

A new trailer had come out (right click, save as): http://files.sohu.com/tlbb/tlbb_trailer.avi

From the movie, it is clear that the graphic is nothing to brag about. If TL is successful (and I predict it will), it is another confirmation that game-play features are much more important than eye-candy.

On 5/9/2007, TL will go into open beta. See the following link:


I will talk about why I think TL will be a very success game tomorrow.


A test for now