Book Review: Gaming the Vote (2008)


gaming the vote

by Alan F. Zundel

“Gaming the Vote: Why Elections Aren’t Fair (and What We Can Do About It)” is an entertaining primer on the craziness of our usual method of running elections in this country and the various alternative methods proposed to replace it. Long on anecdotes and short on in-depth analysis, it whets the desire for electoral reform without making a definitive case for “what we can do about it.”

Not that it doesn’t reach a conclusion. The book was recommended to me by an online acquaintance who takes issue with my promotion of ranked choice voting (called instant runoff voting in this book) and argues that range voting is superior. The book does come down on the side of range voting as worth a try, which is a tepid recommendation. It made me want to learn more about range voting, but the contents of the book were insufficient to get me to give up my efforts on behalf of ranked choice.

Our current electoral system is bedeviled by the “spoiler” effect, which can (and often does) result in a candidate winning even though a majority of voters are opposed to that candidate. The 2000 Presidential election is often presented as a spoiler election (I am not so sure, but it makes for a familiar illustration). George W. Bush won with a minority of the votes, even though the combined majority of voters for Ralph Nader and Al Gore were opposed to Bush.

There are plenty of other examples if you look for them, and the author, William Poundstone, is deft at picking the most colorful ones to make his point. The most literally colorful is the case of “the Blue Man,” whose skin turned blue from drinking what he believed is a “natural” antibiotic. His hopeless 2006 campaign for U.S. Senator in Montana (he got 2.55% of the total vote) resulted in the Republican loss of a majority in the Senate.

Poundstone describes the shady maneuvering of politicians and their consultants to exploit this electoral flaw, but probably no reasonably intelligent observer needs to be convinced that something is wrong with our electoral system. The first section of the book describing the problem is thus overkill. It is the second part, on the proposed alternative electoral systems, that is the meat of the book.

He runs through the most prominent alternatives, doing a good job of making it clear how they differ and keeping the reader’s interest with stories of the history of electoral reform debates. He raises problems with each of them, lending weight to Professor Kenneth Arrow’s famous dictum that there is no perfect democratic electoral system. The consensus of those who have looked into the issue is that the system we now have may well be the worst of the lot. But what should we replace it with?

What the alternatives offer is the opportunity for voters to give more information about their desires. One set of proposals allows voters to rank candidates according to each voter’s preferences. This set includes ranked choice voting, as well as the Condorcet and Borda voting methods. They are all subject to a flaw known as cycling, which in particular (hopefully rare) circumstances can produce results that defy our notions of fairness. Professor Arrow’s dictum was based on studying these systems, and resulted in what is known as the “Impossibility Theorem.”

The second set of proposals, however, takes a different tack. This set is based on voters rating each candidate, and includes both approval voting and range voting. Approval voting lets a voter give a straight thumbs up-or-down for each candidate, and has some serious recognized flaws. Range voting lets you rank each candidate on a scale, such as from zero to ten. It may well escape the Impossibility Theorem.

Without going into all the arguments, suffice to say that the debate has narrowed down to a competition between ranked choice voting and range voting. Ranked choice voting is superior to our current system and has the advantage that it already has a lot of political support and is making headway in the U.S. Range voting may have the advantage that it is the best of all the systems generally proposed—if its proponents are correct.

It is hard to judge whether they are right without a more in-depth study of the question. The primary evidence for its superiority is a computer simulation of all of the alternatives under various assumptions that found that voters would have the least reason to regret the outcomes of range voting. I would need two things to be persuaded that this analysis is correct. One is to examine the parameters that the simulation was built on, along with hard thinking about whether it missed anything important. The other is more testing of the method, not only in corroborating computer simulations but in real life.

Lacking that information, I rely for a first impression on imagining how I would respond to a ballot using alternative electoral methods. In instant runoff voting the voter needs to rank the candidates in order of preference: I like this one the best, this one second best, this one third best, and so forth, stopping at any point one chooses to. Range voting requires the voter to give a rating to each of the candidates (or as many as the voter wants to rate): this one is a six, this one a ten, this one a zero, and so forth.

To me, the ranked choice ballot is doable. I can pretty easily order my preferences for most of the candidates. The range voting ballot makes me have to think harder. The candidate I like best is easy: I give the highest possible rating. And the candidate I like least I give the lowest rating. But what about candidates in the middle? A six? A four? A seven or five? Not only do I need to think about how much I like them, I have to think about how my rating might affect the race overall.

Poundstone claims the rating system is easy for people because we do it all the time: rating books on Amazon for example, or movies on IMDb. But that’s because the implications of your rating in the latter cases are inconsequential. People feel how they vote is more important than how they rate a book, movie, or hotel, and thus the decision takes more thought. Maybe too much thought for most of us.

Whatever the merits of this reservation, I am intellectually intrigued enough to look into range voting further, but on a practical level I still find ranked choice voting the more attractive alternative. Ranked choice voting is better known, it has a political support infrastructure, and it has been tested a lot in the real world. (It has been used in other parts of the world for many decades.)

“Gaming the Vote” makes a good case for changing our electoral system, but it also inadvertently makes the case that theoretical debates about the best system may go on for a long, long time. To make headway the proponents of range voting will need to build political and intellectual support for the method, which is also going to take time (not to mention money and political skill).

We don’t have to wait that long to make things better. There is another dictum that applies here. It’s the old adage, “don’t make the perfect the enemy of the good.”

Buy on

(Note: I plan to do a fuller comparison of ranked choice and range voting in a week or two on my other blog,



  1. Warren D Smith (PhD) September 27, 2015 at 9:34 am

    your own writeup here also was long on anecdotes and short on analysis. I’d recommend the website which has over 1000 subpages if you want analysis. E.g. two subpages about instant runoff are
    and the current Trump-dominated race is considered here

  2. Matthias Bendewald September 28, 2015 at 12:40 am

    You write that instead of trusting computer simulations you would “rely for a first impression on imagining how I would … ”
    Well, I don’t believe anyone (me included) to be able to get a meaningfull idea of any voting system and it’s flaws by just “thinking about it” or even by making tons of mathematical criteria (like condorcet criterium or the like). I am convinced that only a good combination of something like those computer simulations, real-live experiences and an unbiased discussion of the results of the real-live votings leads to a good conclusion.
    Your own mental model will always differ from others mental models, like mine. Maybe you’re smarter than me – but does that mean that you can’t be wrong in this case?

  3. Steve Cobb September 28, 2015 at 5:50 am

    The scale in Range Voting can be sized to the problem at hand. If there are many candidates, the scale need not be 0-10, but maybe 0-5, 0-2, or even 0-1 (Approval Voting). It’s analogous to a computer monitor’s color depth: for some purposes you need full color, but in others monochrome is not only sufficient but superior.

    Calling IRV “Ranked Choice Voting” creates confusion with the other ranked voting methods (e.g. Condorcet Voting and Borda Count) which are arguably better and simpler.

  4. Alan October 1, 2015 at 7:12 pm

    Warren: This is a book review; not meant to be an in-depth analysis of any particular system of voting.

    Matthias: As you quoted me, my “first impression” would be based on thinking about it. For a considered judgment I would have to do more of course.

    Steve: It seems likely any range will revert to approval voting with a “bullet” strategy for helping the one candidate you like best. Second and third choices, etc., just get a zero for strategic voting purposes.

    Glad you are all so passionate about the problem of electoral reform, but remember this is a book review not a electoral methods review.


Leave a Reply

Your email address will not be published. Required fields are marked *