Instant Runoff Elections

So I was wondering why in Atlanta, and many other regions of the US we have runoff elections if there is no majority winner in the primary election rather than simply doing an instant runoff.

Instead of voting for your most preferred candidate in the primary election, you would rank order the candidates. In this case, if there is not a majority winner in the primary election, you simply eliminate the candidate with the least votes and the votes are recounted. This process is repeated until a majority winner is found.

This process of voting only once instead of multiple runoffs saves the taxpayers significant expense of holding such elections. So I wondered what's the downside? Well there is plenty of debate. On the one hand, proponents of the instant runoff say that it not only is more efficient, but voters evaluate the primary candidates more thoroughly and increases voter turnout. Opponents on the other hand, say that although the primary candidates may be more thoroughly evaluated, the winner does not get as much scrutiny as they would if they had a separate runoff election.

http://www.sacbee.com/325/story/2371451.html

The following site talks about the benefits of the instant runoff in Australia:

http://www.fairvote.org/blog/2009/06/instant-runoff-voting-in-australia-guest-blog-from-ben-raue/

Auctions: GREAT Expirements on Revenue Equivalence

I'm starting to see the theory of revenue equivalence the same way I see the theory of efficient markets ... They both involve fundamental assumptions that make analysis and forecasting simpler, but they simple don't hold true in most cases.

This started in my analysis of all-pay auctions in the context of crowdsourcing websites (shameless plug: check out our presentation and final paper, it WILL change the way you think about the future of the labor market). The following link continues to break down the theory of revenue equivalence among auction types/formats.

http://www.econport.org/econport/request?page=man_auctions_experiments

The structure of the auction is immensely important to the bids tendered and to the revenues reaped by the auctioneer.

Follow up: Ravens 4th down decision

Interesting follow up. I was watching the Ravens/Steelers game last night and they had an eerily similar 4th down decision to the one I previously described in the Pats/Colts game. The Ravens had the ball, behind 14-17 with ~4 minutes left, 4th down and 5 to go at their own 45. Do you punt and hope for a defensive stop? or go for it? The mixed strategy equilibrium says on average you are better off going for it. Well, they went for it ... and this time ... they made it. They went on to win the game in overtime.

So if we can channel Parzen for a second: "point estimates are for suckers". In similar situations, one can make the same decision and have drastically different results. On average though, if you have a "conveyor belt" of these decisions and are relatively risk neutral, you are better off on average going for it in these scenarios. However, are they really risk neutral? Probably not. Although they will make many of these decisions throughout their careers as a coach, you will only make a handful in each game.

After all the lambasting that Belichek received after going for it in a similar situation, there is not a word in today's press questioning Harbaugh's decision to go for it. As far as the media is concerned, it is the right decision if you make it, but it an utterly unacceptable and ridiculous decision if you miss it.

Pats/Colts: Belichek was CORRECT to go for it on 4th down

Alright, so after staying up late night to watch the amazing finish to the Pats/Colts game, I figured I should get something useful out of it. So I did some research on what game theory says he should have done on that crucial 4th down play.

For those of you who aren't football fans, there was a call made by the Patriot's head coach to go for it on fourth down and 2 yards to go, deep in their own territory. If they made it, they would have an opportunity to run out the clock and win the game. If they didn't make it, the Colts would have fantastic field position and a golden opportunity to get back in the game.

The result: the Pats did not get the first down, the Colts subsequently scored and went on to a most improbable win.

Well, as you probably guessed, the equilibrium lies in mixed strategies, but the real question is: how often should you go for it on fourth down and 2 deep in your own territory? After some searching, I found a study published that statistically models the mixed strategy equilibria in various scenarios, and it has quite surprising implications.

http://elsa.berkeley.edu/users/dromer/papers/nber9024.pdf

In general, the results of this model show that coaches are not aggressive enough ... by a long shot. In fact, in this particular situation, the study finds that "Even on its own 10 yard line -- 90 yards from a score -- a team within 3 yards of a 1st down is better off on average going for it". In actuality, the patriots were out past the 20 and only had 2 yards to go.

Belichek is vindicated!!! All those doubters of his strategic master mind take note!!!!!

Disclaimer: I grew up in Boston ... there's some egocentric bias going on here ... Go PATS!

Location for Group Meeting: Efficient, Equitable and Envy Free

Last night my group was negotiating a meeting place to work on our SDA final project, which was mostly dependant on how long everybody's drive times would be. I will be driving from my in-laws in Woodstock, Ryan from Smyrna, Doug from Marietta, and Dave from all the way out in John's creek. Our options for meeting location were infinite, but we whittled them down to the two locations that had the facilities we needed: my work in Kennesaw, and Emory. The payouts for each player are shown below, measured in minutes of driving distance.

Kennesaw results the shortest drive for Ryan, Doug, and myself. However, this makes Dave's drive longer than his drive to Decatur. For Dave in this case, Kennesaw is an inequitable solution and he is full of envy of our short drive times. However, Kennesaw results in the least overall driving time for the overall group (95 minutes vs 190 minutes to Decatur). At this point, I realized, this is an opportunity for cash (or other) side payments. A potential efficient, equitable, and envy free solution: we meet in Kennesaw and the group buys Dave a 12-pack of beer (depending on how Dave values his time ... and beer).

China's monetary policy: Peg Yuan to US Dollar

As Obama makes his way through Asia, there is a lot of talk about chinese economic policy, and rightly so. For the last year and a half, China has resumed their policy of pegging the yuan to the US dollar (see chart) despite the massive trade surpluses, thereby thwarting the natural tendency of the dollar to weaken against the yuan. Allowing the dollar to weaken would cause chinese product pricing to increase in USD, reducing US demand for chinese goods .... at the same time, it would devalue the US debt that China currently holds.

This LINK shows a good basic analysis of the Chinese strategy to peg the yuan to the US Dollar (keeping the exchange rate constant). It also includes an ill-constructed gametable showing the payoffs of the players, but what do you expect from NYU guys.

Condo Association Prisoner's Dilemma and Agincourt

So I was drafted into service for my neighborhood condo association, and have blissfully served as president for the last few years, so the subject of homeowners paying/not paying their association fees is always on my mind. My first aha moment about the structure of this fee paying game occured in a class discussing a multiplayer, simultaneous game. This is exactly what this game is, and unfortunately, it's actually a multiplayer prisoner's dilemma.

The structure of a 3 player version of this multiplayer game can be seen below, click on the picture to view full scale image.

This analysis assumes association fees are $150. Further, if all owners pay, the result is a well kept community worth $150 to an owner. If all but one owner pays, the result is a slightly less kept community, worth $100. If only one owner pays, the upkeep of the community is only worth $50. If no owners pay, the community gets no upkeep, worth $0.

As you can see, I solved for the equilibrium of this game by best responses. As you can see, the dominant strategy for all three players is NOT TO PAY. This results in an equilibrium that when owners are left to their own selfish devices, they would have the community in shambles. So, how can we avoid this? Well, we need to change the owners payouts ala Agincourt. We need to instill a sense of community. I tried this just a few weeks ago with a "beautification" day where everyone comes out to clean the property, paint fences, install new pavers, etc. These activities foster a a sense of ownership and pride in the community. It creates relationships and friendships. Who would want to screw their own friends? After focusing on these kinds of activities, there is a notable difference in the attitude of the community. Owners are tending more towards paying down their debts and delinquencies are dropping. Thanks SDA!

Welcome to my Strategic Decision Analysis Blog!

Here I will be posting several of my keen insights on strategy as it relates to the Strategic Decision Analysis coursework. Please comment (and challenge me) as much as you dare!