• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

(Sales Age) An introduction to theSimExchange.com (Now with a NeoGAF league-post 29)

This is more of any FYI for those of you interested in our whole "sales-age" bent here on NeoGAF. I've been following them since they started up and found it interesting.

I know some of you are already aware of The simExchange.com , but I thought some of you might not and be interested in it. They were recently included in Pachter's monthly NPD preview and they've raised some eyebrows out in the industry with their forecast application. One of the more difficult things for the industry is to try and be more accurate in SW forecasting. The big franchises can sometimes follow a predictable trend from release to release, but frequently do not and new titles... Forget about it. It's a complete crap shoot much of the time.

They've applied prediction market's methodology to the VG industry. Essentially applying a stock market model where gamers buy and sell contracts based on their expectation of how a given game or system will perform. Predictors who do better are rewarded with more funds to make predictions and those who do worse with less funds, weighing the prediction towards the better participants.

They've got more information on the whole methodology on their site, but in the end, it's a "wisdom of the crowd" application, which studies have shown to outperform individual analysts the majority of the time.

I've talked with them and the interesting thing about it is that they don't make any predictions themselves. It's all based on the user participation and performance. I haven't done a comparison for their forecasts vs ours, but I believe neoGAF has done better the majority of the time when we compare HW results against our average, so it will be interesting to see how these trend against each other over time.

Since I've done it for Pachter, I thought I'd see what you guys think about them, and just clue you in on them as I think you might be hearing more about them in the future. I'm also including their latest monthly preview as fodder for discussion. It's really more of a recap of what the forecasts are, but it does pepper in commentary for what it means for the various parties involved.

I'm interested to see what you guys think and to make sure that NeoGAF raises it's prediction game up a level as I'll make sure we include TSE (the sim exchange) along with our predictions and Pachter's each month.

More of the preview is available @ http://www.thesimexchange.com/research.php

SimExchange said:
October Preview Report
Jesse Divnich
November 12, 2007

Key Points

Wii and Nintendo DS expected to lead their product groups
40GB PS3 Model expected to increase sales
Activision expected to dominate the holidays
First-person shoots are cannibalizing each other's sales
Halo 3 sales not affected by PS3 hardware news or PS3 exclusives
Sony continues to struggle with its PS3 exclusives
Introduction

This report is intended to aid industry professionals and the press in drawing insight and discovering trends occurring in the video game industry through the use of the simExchange video game prediction market (www.thesimexchange.com), a data aggregation system in which gamers, developers, and industry investors trade virtual video game stocks and futures to predict how video game products will sell and how they will be critically received.

Results based on NPD data have consistently shown that The simExchange prediction market can predict results not only more accurately than traditional models and surveys, but also more quickly adjust forecasts based on industry and product news.

The simExchange's Market Predictions

US Hardware in October 2007Title Predicted unit sales
Xbox 360 412,000 units
PLAYSTATION 3 136,000 units
Wii 467,000 units
Nintendo DS 460,000 units
PlayStation Portable 251,000 units

US Software in September 2007Title Predicted unit sales
Enemy Territory: Quake Wars (PC) 124,000 units
Final Fantasy Tactics: The War of The Lions (PSP) 132,000 units
Guitar Hero III: Legends of Rock (Xbox 360) 502,000 units
Guitar Hero III: Legends of Rock (Wii) 213,000 units
Guitar Hero III: Legends of Rock (PS3) 113,000 units
Half-Life 2: The Orange Box (Xbox 360) 419,000 units
Halo 3 (Xbox 360) 738,000 units
Project Gotham Racing 4 (Xbox 360) 89,000 units
Ratchet & Clank Future: Tools of Destruction (PS3) 131,000 units
The Legend of Zelda: Phantom Hourglass (DS) 269,000 units
 

Jesse2040

Member
Originally Posted by sonycowboy:
Donny has already done so. They actually beat the GAF aggregate the majority of the time.

Good point sonycowboy, but dosen't the top 15 GAF estimaters beat us(TSE)?.

I think the important thing here is that gamers do a better job at predicting the market than financial Analysts. Whether its GAF or TSE, both of their predictions are pretty solid.

-
Jesse Divnich
 
Jesse2040 said:
Good point sonycowboy, but dosen't the top 15 GAF estimaters beat us?.

Would they beat your top 15 estimators?

The PGR4 prediction has to be high, FYI. Check the global scoreboards. Even considering that many buyers are offline, the numbers don't work out.
 
Thanks for the introduction. I am Brian Shiau, the founder of the simExchange.

I posted about this on another thread, but it is equally relevant here:

Both the NeoGAF and the simExchange predictions are based on the idea of "the Wisdom of Crowds"--that all of us are smarter than just some of us. In my chats with Michael, he admits that the aggregated intelligence of many should be smarter than individual analysts. After all, he is an analyst of the stock market, a world in which the belief of many trumps all.

Although both a survey and a prediction market aggregates the opinions of many, I do want to share with you some flexibilities of a prediction market.

(1) A prediction market requires participants to back their opinions using some limited resource, in our case "DKP." In a prediction market, participants will only express an opinion if they feel strongly enough about their opinion. Opinions are weighted, so that people who believe their information is better will bet more and affect the prediction more. Additionally, there is a profit incentive to quickly price in any information you learn from the news. Our predictions are running months to years in advance.

(2) A prediction market is self-correcting. There will naturally be outlier participants who are predicting well beyond the norm for a console or a game. The nature of a prediction market allows rational participants to correct for the outlier participants. For example, if a player is very bullish on the Wii and moves the prediction upwards, other players can short sell the prediction back down.

(3) Prediction markets become more accurate over time. It does take users some time to get a hang of how a prediction market works. However, one reason prediction markets get better is that accurate predictors get rewarded with more virtual money to make more predictions. Bad predictors lose their virtual money and make less predictions.

Just to provide more background on the methodology: prediction markets have been implemented by universities and corporations to forecast the future for a long time. The Iowa Electronic Market forecasts which candidate will win the US Presidency. Compared with 596 polls, the prediction market was more accurate 75% of the time. HP used prediction markets to forecast their printer sales, and over 3 years their prediction market outperformed their own analysts 75% of the time. The simExchange applies the same technology to the video game industry.

Due to the way a prediction market works, we can provide forecasts in real-time, months to years in advance. You can check out our forecasts free from our data site.

It is good to see so many people interested in game sales on NeoGAF. At the simExchange, we don't view ourselves as in competition with analysts or communities like this one. Our goal is to provide another data set which may be useful in your analysis. We also welcome you to participate in the prediction market as you continue to make monthly predictions here.

Thanks for your time.

Brian Shiau
 

Evlar

Banned
I've commented before that GAF's "wisdom of the crowd" predictions are really damn good, spookily good, at least when applied to HW sales. If simExchange's selective pruning can shape a userbase that's as knowledgeable about SW as GAF is about HW this can be impressive.
 

RobertM

Member
Jesse2040 said:
Originally Posted by sonycowboy:


Good point sonycowboy, but dosen't the top 15 GAF estimaters beat us(TSE)?.

I think the important thing here is that gamers do a better job at predicting the market than financial Analysts. Whether its GAF or TSE, both of their predictions are pretty solid.

-
Jesse Divnich
Because we see the hype, know what games released, and compare previous sales figures.
 
beermonkey@tehbias said:
Would they beat top 15 estimators?

The PGR4 prediction has to be high, FYI. Check the global scoreboards. Even considering that many buyers are offline, the numbers don't work out.

It is unclear how the top 15 predictors on the simExchange would compare as they do not express discrete predictions--they only buy if they believe a forecast is low and they short sell if they believe a forecast is high. It is interesting that the "elite" predictors on GAF outperform the whole. The argument in James Surowieki's "The Wisdom of Crowds" is that the group as a whole should be better than the individuals, including the smartest and most informed ones. Taking the predictions of a whole company, rather than just the committee of its execs and analysts, should yield better results.

The beauty of a prediction market is you can go and short sell the prediction for PGR4 down if you believe it is too high. Naturally, there are those of the other opinion, and so it's a matter of who feels more strongly about their forecast and will bet more.
 
KultofCows said:
It is good to see so many people interested in game sales on NeoGAF. At the simExchange, we don't view ourselves as in competition with analysts or communities like this one.

This is Neogaf, dude. Everything is a competition or an argument. :lol

But seriously, cool site.
 
RBH said:
Any reason for the exclusion of the PS2 in these predictions?

We had originally intended to forecast the next-gen consoles (the public beta for the prediction market was launched in November 2006) and never ended up picking up PS2 hardware as a metric we forecast. We had several PS2 games listed for lifetime global sales forecasts in our earlier days, but these listings have trickled off.

One question we often get is why X game isn't listed for trading. Games that are listed for trading are submitted by users and we track the interest in trading that particular game. The reason for this is that too many games on a prediction market dilutes the liquidity (trading interest/activity) across the board, so we have to manage the number of predictions that are running at any given time.

Predictions on the simExchange include (1) global lifetime sales, (2) NPD reported monthly sales, and (3) Metacritic Metascores.
 
BenjaminBirdie said:
So it's kinda like HSX for games?

Like HSX, the simExchange is a prediction market. How things are setup on the simExchange are a bit different, but it does share the same idea of trading virtual stock in media properties to predict how the property will do (for those who don't know, HSX is a virtual stock market for movies).

Our trading/market is more based on actual market dynamics. The simExchange uses a double-call auction trading system like the NASDAQ. Our site also uses AJAX for a more fluid trading experience. However, we will let you be the judge. We are always open to suggestions on how we can improve the site.
 
It's great to have both Jesse and Brian here, I do have a question followed by a proposal:

Does the Sim Exchange subscribe to NPD, if not, any plans to do so? Have you explored the idea of subscribing to the NPD monthly service and providing your members with a subscription-based service where they can access the data?

In light of how protective the NPD group with its data, how about developing a plan to solicit the help of your users to collect local data (national surveillance network) from the major retailers? There's a statistical model that can be implemented to provide accurate data based on a certain number of samples. NPD does not provide a margin of error (at least publicly) on the data it collects, so there's no absolute correct value to the number of units sold, it's based on estimations and extrapolation. You could eventually develop a real alternative to NPD...
 

Parl

Member
KultofCows said:
Like HSX, the simExchange is a prediction market. How things are setup on the simExchange are a bit different, but it does share the same idea of trading virtual stock in media properties to predict how the property will do (for those who don't know, HSX is a virtual stock market for movies).

Our trading/market is more based on actual market dynamics. The simExchange uses a double-call auction trading system like the NASDAQ. Our site also uses AJAX for a more fluid trading experience. However, we will let you be the judge. We are always open to suggestions on how we can improve the site.

Less viral marketing?

j/k
 
KultofCows said:
We had originally intended to forecast the next-gen consoles (the public beta for the prediction market was launched in November 2006) and never ended up picking up PS2 hardware as a metric we forecast. We had several PS2 games listed for lifetime global sales forecasts in our earlier days, but these listings have trickled off.

Welcome! I look forward to seeing how the predictions compare.

I do, of course, have to note that the lack of PS2 and GBA predictions will count against TSE when it comes time to tally the results. Would it be possible to start those up so we can have a more complete comparison in the future? I'm sure there'd be adequate interest among your users.

Thanks for joining in!
 

Thraktor

Member
It looks like a nice site, too bad there isn't a European version (I know, you'd need actual sales data for that), but I might sign up if I get the time.

KultofCows said:
Predictions on the simExchange include (1) global lifetime sales, (2) NPD reported monthly sales, and (3) Metacritic Metascores.

How exactly do you work the Metacritic scores? Given that reviews come in gradually, and the aggregate score can vary for up to a month after release (as magazines come out), when do you cut it off, say "That's the Metacritic score" and make the payout?
 

elostyle

Never forget! I'm Dumb!
Can't we apply the same thing, weighting posters predictions by there previous results and calculate an aggregate?
Would be fun to compete.
 
Amused_To_Death said:
It's great to have both Jesse and Brian here, I do have a question followed by a proposal:

Does the Sim Exchange subscribe to NPD, if not, any plans to do so? Have you explored the idea of subscribing to the NPD monthly service and providing your members with a subscription-based service where they can access the data?

An NPD subscription does not work that way. Any client subscribing to NPD data is bound to keep the information confidential. We could not subscribe and then resell subscriptions at a smaller consumer price to cover a single NPD subscription. This would be like buying a CD for $15, then selling 15 burned copies of the CD at $1 each to cover the original $15. To sum it up, this type of service would not be workable for us.

Amused_To_Death said:
In light of how protective the NPD group with its data, how about developing a plan to solicit the help of your users to collect local data (national surveillance network) from the major retailers? There's a statistical model that can be implemented to provide accurate data based on a certain number of samples. NPD does not provide a margin of error (at least publicly) on the data it collects, so there's no absolute correct value to the number of units sold, it's based on estimations and extrapolation. You could eventually develop a real alternative to NPD...

This would be an interesting project; however, not part of the goal/methodology/expertise of the simExchange prediction market. The way a prediction market works does do what you suggest to a certain degree. When people are trading our NPD Futures or the Lifetime Global Sales stocks, people are using all information available to them, include anecdotal data they've seen at their local retailers. They are pricing this information into the prediction. So we are collecting that sort of local sales data in a way; however, rather than modeling the data, the data is aggregated via prediction market.
 
Thraktor said:
It looks like a nice site, too bad there isn't a European version (I know, you'd need actual sales data for that), but I might sign up if I get the time.

We are always expanding our offerings as our users demand them. We do have a large following from Europe and we hear there is a lot of interest for following that market. Currently, our global lifetime sales stocks are most applicable for international users.

Thraktor said:
How exactly do you work the Metacritic scores? Given that reviews come in gradually, and the aggregate score can vary for up to a month after release (as magazines come out), when do you cut it off, say "That's the Metacritic score" and make the payout?

This is a good question and something we wrangled with initially. What we decided to do is freeze trading 14 days after the game's release and use the Metascore from that day. We had originally planned on freezing trading prior to any Metascore assignment, but this proved impossible as some games may have a Metascore a month in advance of the game's release.
 

Uncle

Member
I remember playing this when there was an earlier thread about it here. It was fun for a time. I sucked, though. Made more DKP with contributions (which I mainly found here).
 

Thraktor

Member
KultofCows said:
We are always expanding our offerings as our users demand them. We do have a large following from Europe and we hear there is a lot of interest for following that market. Currently, our global lifetime sales stocks are most applicable for international users.

Well, a European market similar to what you've got for NPD console sales simply wouldn't be possible, as the sales data collected by agencies like ChartTrack isn't made public. I suppose you could, for instance, try to predict the chart position of particular games each week on the UK charts, but it wouldn't be quite as easy, or get anywhere near as many people, as the monthly NPD predictions.

This is a good question and something we wrangled with initially. What we decided to do is freeze trading 14 days after the game's release and use the Metascore from that day. We had originally planned on freezing trading prior to any Metascore assignment, but this proved impossible as some games may have a Metascore a month in advance of the game's release.

You could, perhaps, freeze trading afer 5 reviews come in, and then assign the score after the 15th review? The difficulty is that once scores start coming in, it obviously becomes a lot easier to predict the final Metascore.
 

Pachael

Member
I love the concept (Wisdom in Crowds) and it's a popular idea these days what with simEx for games and CAPS (fool.com) for the stock market with the idea that a large aggregate of the population will generally do better than a small number of candidates however good they are individually.

So good job guys and keep up the good work :)
 
Thraktor said:
You could, perhaps, freeze trading afer 5 reviews come in, and then assign the score after the 15th review? The difficulty is that once scores start coming in, it obviously becomes a lot easier to predict the final Metascore.

I hear that the game companies that trade their own version of Metacritic Futures in their internal prediction markets often freeze trading when the game goes gold or when the game is shipped. The reason for this is they can't improve the game any further so there isn't any point to continuing to trade the Metacritic forecast. The reason companies use prediction markets is to make better business decisions, so they stop trading once decisions have been made.

For us, part of it is just for the entertainment. Having reviews roll in and trading as the Metacritic metascore is updated is an exciting dynamic. Most prediction markets don't freeze trading before the event is running. Part of the fun of Presidential prediction markets is trading on Election night as the results come in. Prediction markets for sports generally allow trading up to the very end of the game. A preliminary Metascore isn't a certain result, but it does close in the range of reviews (just like having 50% of the electoral votes coming in for an election). However, we freeze NPD Futures trading the night before because the release of that actual data is a singular event.
 

GhaleonEB

Member
These markets are quite interesting. We did a case study on the exchange system run by the University of Iowa when I was in school (at the U of I). It will be fun to see how GAF compares, as these markets move steadily, while GAF predictions trickle in over time. I'm not convinced a predictor that moves visibly like this will be more accurate than ours (where we get the aggregate when submissions are closed).

I took a bath on the Iowa market, btw. Seems predicting console and software sales are more my forte. :lol
 
FYI:

I know a good number of you frequent simExchange as it provides a bit more structure and feedback to the sales age predictions we all participate in.

Well, now you not only can represent yourself but also NeoGAF. theSimExchange has recently added a NeoGAF handle to their user profiles, so you can associate yourself with NeoGAF, and we can accumulate league results which can be compared against other NeoGAF members, the whole of the simExchange as well as additional communities which they'll be adding. Currently, they're starting with NeoGAF, EvilAvatar, & cheapassgamer, but will be adding more as time goes on. Once a level of participation is reached, they'll be providing results specific to each league.

And, really, we should crush everyone else out there. :D

For more information, see here:

http://www.thesimexchange.com/blogpost.php?post_id=448


Also, I'd be interested to see how those of you participate perform and what you think of the site. They've gotten thousands of folks actively participating, and from what I here, if you're willing to invest the time to understand how it works, it's pretty interesting stuff.
 

Jesse2040

Member
Thanks sonycowboy,

It looks like we already have some closet neogaf'ers on our site who joined on.

Once we get enough sign-ups we will begin to make visible the community rankings against others.

It will be interesting to see how the communities perform.
 
The NeoGAF League is now online for view. Currently, the communities competing are NeoGAF, Evil Avatar, and Cheap Ass Gamer.

We've noticed many NeoGAF users have signed up on the simExchange, but to participate in the special NeoGAF league, you must enter your NeoGAF username in your Profile. Simply log into your simExchange account, goto the "Profile" page under the "Accounts" tab, and enter your NeoGAF username into the form and click update.

Please let us know if you run into any problems.
 
KultofCows said:
The NeoGAF League is now online for view. Currently, the communities competing are NeoGAF, Evil Avatar, and Cheap Ass Gamer.

We've noticed many NeoGAF users have signed up on the simExchange, but to participate in the special NeoGAF league, you must enter your NeoGAF username in your Profile. Simply log into your simExchange account, goto the "Profile" page under the "Accounts" tab, and enter your NeoGAF username into the form and click update.

Please let us know if you run into any problems.

Thanks for that guys. It's a bit addictive and I placed my first trade yesterday and now I'm out of money until NPD is released I guess. I haven't looked too closely, I just wanted to see how simple it was without researching it too much. However, my trades were based on market prices that I felt were out of whack with what I've been reading, seeing via channel checks, and from friends out there.

For those who are interested, here are the only folks who haver registered so far:

NeoGAF LeagueRank Name NeoGAF Alias Net Worth
At Close %Change MTD
1 rikitikitik rikitikitik 32,750,364.82 -2.29%
2 charron Charron 17,413,924.05 -0.31%
3 dinodoh Dino 2,618,306.43 -0.32%
4 mj1108 mj1108 2,108,635.67 -1.06%
5 jj984jj jj984jj 1,471,276.55 -2.35%
6 sonycowboy sonycowboy 1,034,958.75 +3.50%
7 felipeko felipeko 1,009,228.19 0.00%
8 squicken squicken 1,000,000.00 0.00%
9 apujanata apujanata 999,905.33 -0.01%
 

Meier

Member
I can't be arsed to keep up with it as much as I did before, but I added my name to the league (MiyazakiFan).
 

jj984jj

He's a pretty swell guy in my books anyway.
argon said:
I no longer play actively, but I added my name to the neoGAF league just for the hell of it.
Me too, though this league had me go make a few new bids, they aren't going to well from the looks of my rank. :lol
 

Jesse2040

Member
There is still time to get your bids in before the NPD results.

It is pretty simple, if you think the current estimate of any of the NPD November futures is too high, then sell (yes you can sell what you don’t have), if you think it’s too low, then buy.

Also, please keep in mind that this isn’t competing with NeoGAF’s monthly NPD estimates; The simExchange is just another way to add some fun to being an armchair analyst. It’s comparing apples to orange, one set of figures is an aggregate (NeoGAF); the other is from a prediction market (TSE). Both are great ways of forecasting data, which is why both estimates are relatively close to each other every month.

Looking forward to seeing more NeoGAF’ers participating.
 

GhaleonEB

Member
None of the "Trade This" buttons work for me. I'm signed in and trying to trade, but the pages load with errors and the buttons don't register at all. No audible click, can't open in a new tab, etc. I guessing I need to change some settings on my browser, but I'm not sure what. Any suggestions?
 
GhaleonEB said:
None of the "Trade This" buttons work for me. I'm signed in and trying to trade, but the pages load with errors and the buttons don't register at all. No audible click, can't open in a new tab, etc. I guessing I need to change some settings on my browser, but I'm not sure what. Any suggestions?

Is Javascript enabled on your browser? The simExchange is largely AJAX based and Javascript is required.
 
A Link to the Snitch said:
So I'm dumb right, buying means you think it'll hit the number, selling means you don't?
You buy when you think the "real" price is higher than the current price.
You sell/short when you think the "Real" price is lower than the current price.
 
A Link to the Snitch said:
So I'm dumb right, buying means you think it'll hit the number, selling means you don't?

If you are buying a stock, you are betting the game will sell more than the current forecast. If you are short selling, you are betting the game will sell less than the current forecast.

For example, Mario Kart (Wii) is currently forecast to sell 13.65 million copies globally over the lifetime of the game. This corresponds to a stock price of 1,365.09 DKP. If you think the game will sell 14 million copies, you would be willing to buy the stock under 1,400 DKP. If you think the game will only sell 10 million copies, you would be willing to short sell the stock down to 1,000 DKP.
 
GhaleonEB said:
These markets are quite interesting. We did a case study on the exchange system run by the University of Iowa when I was in school (at the U of I). It will be fun to see how GAF compares, as these markets move steadily, while GAF predictions trickle in over time. I'm not convinced a predictor that moves visibly like this will be more accurate than ours (where we get the aggregate when submissions are closed).

I took a bath on the Iowa market, btw. Seems predicting console and software sales are more my forte. :lol

The University of Iowa's Iowa Electronic Market is actually one of the most famous examples of a prediction market. We actually mention this on our data homepage. What was particularly interesting with their presidential markets is that it has been more accurate than polls, such as the Gallup Poll, at predicting the next President 75% of the time over the past couple of decades. There are several presidential markets running around the web which are interesting to track following debates and other developments.
 
GhaleonEB said:

Java and Javascript are different technologies. The simExchange doesn't use Java at all, but Javascript is required.

Does that popup occur when you click the "Trade this" button or does it only come up when you double click the error loading icon at the bottom of the browser?

Are you able to use the pull-down menus on the left navigation bar?
 
Top Bottom