Double Blind

A frequent argument made by anti-road activists is that more roads will lead to more driving which will lead to more pollution. The environmentalist lobby typically makes the same arguments. They refuse to believe that the addition of more roads is the solution to traffic congestion. While they correctly point out that the extra roads will proceed to fill to capacity, this is because the addition of extra roads makes the area more accessible and thus livable – more people flock to areas with better road access. The trick is the causality – I claim that population density is the cause of extra traffic, not the addition of new roads. The trick is to spread the population density so it cannot create concentrated choke-points on the traffic grid. Geographically in an area like Greater Vancouver, this is difficult because most of the significant regions are separated by water. When there’s water, you need bridges.

As I live and work in Richmond, I thankfully do not have to encounter any of the bridges that most commuters have to face each day. Each bridge is a choke point that inevitably results in backed up traffic. Other than the Port Mann Bridge, which I covered in a previous article, the second worst choke point in the road network is the Lion’s Gate Bridge. The upgrading of the Lion’s Gate Bridge was a political mess and I will not talk about how such a decision stalled a real road network in Vancouver for at least another 30 years.

The George Massey Tunnel, however, is an interesting choke point on what little of a freeway system our area has. The tunnel connects Richmond to Delta, and this freeway is also used to connect to Tsawwassen and the ferry terminal to Victoria and also White Rock. Originally before the highway 91 freeway was connected, the 99 freeway was the optimal way to reach south Surrey from Richmond. Suffice to say, the tunnel services a lot of traffic.

The tunnel is also four lanes and can be adjusted to move traffic in one direction with three lanes and the other direction with a single lane. Typically the commute north through the tunnel is very heavy during the morning and is very heavy southbound in the evenings. The traffic has become bad enough that highway 99 saturates and the overflow begins to spill into the freeway exits (which are the entries into the freeway). In the reverse direction, even though the tunnel only provides one lane of traffic flow, one can typically get through the tunnel after about 10 minutes of queuing up as the lanes converge.

The map to the left illustrates with the red spray paint what congestion one typically sees on a Friday afternoon in Richmond. First of all, highway 99 itself, southbound, is completely saturated with stop-and-go traffic. There is one issue involved here – the freeway has two regular traffic lanes and a HOV (3 passengers or higher) lane which is poor design because the HOV lane is on the right of the traffic. As a result, anybody entering into the freeway has to merge onto the HOV lanes first, slowing down traffic when they have to get into the left two lanes since they’re the only drivers in their car. This is exceptionally bad at the Steveston Exit/Onramp (Exit 32 on the map), where you have a steady stream of traffic blocking the entrance into the tunnel. This forces traffic to slow to a crawl as they align to get into the tunnel, which of course backs up until you’ve reached another choke point, the Oak Street Bridge (connecting Richmond to Vancouver).

The second reason, indirectly, is the tunnel. It can only service three lanes of traffic in one direction. Although three lanes is enough to service the existing Highway 99 as-is, the inability to supply more lanes to the tunnel prevents the construction of what Highway 99 really needs – a fourth lane to handle all the merging traffic. If the fourth lane is built without upgrading the capacity of the tunnel, the new choke point will be the tunnel itself, opposed to the Onramp near Exit 32.

The result of saturated traffic is that cars eventually slip beyond the freeway entrance points at the Onramp – they begin to slip out onto the main streets. When driving home from work (westbound on Westminster Highway), I noticed that the traffic queue to get into Highway 99 (between exits 36/37) was stretched all the way between Garden City and No. 4 road. Tuning into the radio, I discovered that Steveston Highway was backed up to nearly No. 3 road (not surprising), and the other entry from Richmond (via Bridgeport and Garden City) would probably be backed up a little. All of these cars waiting to just get onto a saturated freeway cannot be good for the mileage.

It is very bad for freeway congestion to spill into the city streets. It ruins the ability of people to get around in the city legitimately (for example, the corner of Steveston and No. 5 is an area with grocery stores and restaurants) and it also wastes the time and resources of those commuting around the city. The worst example of this was in California – State Highway 91 services much of the traffic travelling between Los Angeles and its eastern neighbour, Riverside County. The freeway got congested and they sold the rights to the toll lanes to a company. The only problem is that the contract included a clause that said that the California Transportation Department could not compete with the toll roads in 25 years! What happened is that as a result of this, congestion did not improve and the City of Corona sued unsuccessfully because of all of the traffic that jammed up the city streets.

Unlike the Lion’s Gate Bridge, there is something that can actually be done with Highway 99’s problem – expand the capacity of the tunnel and expand the capacity of Highway 99 by putting in an extra lane southbound. Put a toll on the tunnel southbound from 3pm to 7pm to pay for the project. The southbound traffic capacity can handle the subsequent flow and this will be the quickest way of getting cars off the side streets as they race to get home for the Friday.

I received a letter in the mail from the Insurance Corporation of British Columbia (ICBC) which politely informed me that my automobile insurance was due for renewal in a month. I was rather shocked when I opened the letter and discovered that I will be paying less for insurance this year than I did in 2003. I paid $1196 in 2003, while I will be paying $1080 in 2004 for the same basic and optional coverage. Although I am in the 40% discount category this year (compared to 35% the previous year) even when you eliminate that difference, my insurance this year is cheaper. Since a generous estimation of the book value of my ancient vehicle is $2500, I am basically covered for people that decide to bump into me at sufficiently low velocities to cause two thousand dollars of damage or less. If my car is totaled, the insurance firm will pay me book value on the vehicle and I will have to find a new one.

What caught my attention, however, was the option that they gave people to pay in monthly installments, called “Autoplan12″. Instead of paying $1080 in one lump sum, I could pay 12 monthly installments of approximately $95. 12 times $95 is $1140 so thus the interest charge of using Autoplan12 would be $60 over the sum of a year. $60 over $1080 is approximately 5.6%, which is not a bad rate of interest, right?

This is an incorrect calculation! The problem here is that the payments are performed monthly, so you will not have the utilization of the full $1080 during the entire year. Every month, starting with the initial payment, $95 will be subtracted from your account until 12 payments have been made. So if you are starting with $1080 of capital, you deduct $95 for your first payment and then start earning interest on the $995 of capital you have left. You repeat this process for 11 more iterations and then all payments have been made. How much of a rate of interest must you achieve with your remaining capital in order to break-even with Autoplan12?

The answer is 12%. This is significantly higher than the paper napkin calculation above that somebody uneducated in finance would make. 5.6% is the typical interest rate you would receive on a 5-year secured loan (such as a mortgage), while 12% is extremely expensive even for an unsecured loan to a credit-worthy person. Despite the fact that I have more confidence in my investing abilities to pull a risky return greater than 12%, my decision to make the annual payment of $1080 was obvious since the 12% was risk-free.

I find it very good marketing on the behalf of ICBC to conveniently omit the fact that the interest implied in the payment plan is so ridiculously high. There is nothing illegal about it, nor should there be any laws that would make it illegal. People simply have to be better educated in matters of finance and I do not feel any sympathy for the suckers that do decide to take the Autoplan12 package.

I was having a discussion with a friend one day and he claimed that he liked sitting between two players at a Caribbean Stud table because he could see their cards and thus use that information to make “winning decisions” at a casino game. Despite the fact that you are not supposed to share information with other players around the table, peeking is frequently done and is tolerated by the casinos because there is still not enough information available to make winning decisions. I explained this to my friend and also explained that the only way that he could win at Caribbean Stud was to collude with all seven people around the table. My friend still adamantly claimed that he was a ‘winning player’ at Caribbean Stud and over several sessions that he was ahead. I called “bullshit” on this and left it at that – there was no amount of arguing or presentation of proof that could be done in that situation to show how full of it he was.

Here I will present evidence to back my claims. Or rather, most of the work has been done for me by the Wizard of Odds, who is a mathematical consultant for casino games and has other credentials in the gaming industry. Specifically, his article on the Analysis of Caribbean Stud provides ample evidence for me to back my claims up. Since I don’t want the reader to wade through pages of information, I will summarize it.

You have to put down an initial bet. Let’s say this is $10. You are dealt five cards and the dealer is dealt five cards. The dealer has one card that is face up. You have two choices. Either you can fold and the dealer will take your $10. Or you can raise, which means you put another $20 in front of you and then you compare hands. You cannot raise any more or any less money.

If the dealer does not have at least an ace and king in his hand (or something better like a pair, two pair, etc.) then you get the $20 that you raised plus you get 1:1 on the original $10 that you placed. In other words, you will end up with a profit of $10 for that hand. If the dealer does have an ace and king in his hand (or better) then the player with the higher poker hand wins. If the dealer has the higher poker hand, he gets the $30 you placed on the table (i.e. you lose your original $10 and $20 raise). If you have the better hand, you get 1:1 on your original $10 bet, and your $20 raise back, plus an increasing amount of money depending on what hand you had. It amounts to a profit of $30 if you had a pair, $50 if you had two pair, $70 if you had a three of a kind, $90 for a straight, $110 for a flush, $150 for a full house and more for hands that you and I will never see in our lifetimes.

The table underneath the “Strategy” section tells you how much money you will lose in the game.

If you bet $10 in a hand, you will lose 52.24 cents per hand on average assuming you play perfectly. Now as I explained above, you have the option of putting more money on the table (i.e. raising in a game) and if you play perfectly, you will lose 76.65 cents per hand in a game where you originally bet $10. This is what the Wizard of Odds calls the “element of risk” in that you are actually risking $30, but your expected loss in a percentage form is less.

So let’s assume that you play 100 games, at $10 a piece. You will typically walk out of the casino with $52.24 less in your pocket at the end of the day.

Now, it was mentioned by my friend that he said that he can increase his chances of winning by looking at the cards of the player to his left and the right. If you look at the very bottom of the article under “Player Collusion”, you see the following:

---

According to the paper ‘An Analysis of Caribbean Stud Poker’ by Peter Griffin and John M. Gwynn Jr., which appears in the book Finding the Edge in the perfect situation of having 7 colluding players it would be possible to narrow down the dealer’s unseen cards to just 16. Using a computer to analyze all 1820 possible 4-card sets out of 16 the player would have an advantage of 2.3%. In a six player game the house would still have an edge of 0.4%.

---

In other words, if you were able to look at all 7 hands around the table and use perfect analysis, you could walk away with 23 cents per $10 bet. However, if you could only take a look at six players around the table, you will lose an average of 4 cents per $10 bet. The conclusion here is that if you look at less than seven players’ hands around the table (which is illegal, but assuming you can pull it off anyway) you will lose money, on average, playing Caribbean Stud. Thus, my friend’s claim of being able to beat the game by just looking at the hands of the players to his left and right is complete garbage.