The Texas sharpshooter is a fabled marksman who fires his gun randomly at the side of a barn, then paints a bullseye around the spot where the most bullet holes cluster. The story of this Lone Star state shooter has given its name to a fallacy apparently first described in the field of epidemiology, which studies how cases of disease cluster in a population.
http://www.fallacyfiles.org/texsharp.html[/url]

It’s been a while since we had a puzzle and this seems like a twofer to me in the sense that it is both a puzzle and a discussion.

My question is, What is wrong with the Texas Sharpshooter Fallacy?

You should answer your own question. Before you do, consider the logical implications of the following:

"Argument from fallacy: assumes that if an argument for some conclusion is fallacious, then the conclusion is false."

This wouldn't mean that your conclusion is alternatively true, it would only mean that since you committed a fallacy in drawing a parallel, you must find some other [non-fallacious] means to support your claim.

Interbane wrote:You should answer your own question. Before you do, consider the logical implications of the following:

"Argument from fallacy: assumes that if an argument for some conclusion is fallacious, then the conclusion is false."

This wouldn't mean that your conclusion is alternatively true, it would only mean that since you committed a fallacy in drawing a parallel, you must find some other [non-fallacious] means to support your claim.

Just because one's argument is fallacious doesn't mean that it is necessarily wrong does it. Nor does it mean my conclusion is automatically wrong only that it most likely is wrong.

As for TSF I am curious to see if anyone else sees an inherent fault, or faults in it. I believe that there are three.

Terms like "most likely" or "automatically" are signifiers of some sort of application of probability to your claim. There are methods for calculating such probability, Bayesian for example, but in this case there is no assignment of probability. Which means, based on your arguments alone, we can't determine if it's most likely wrong or most likely correct. All that we can say is that the argument you're using to support your claim is wrong. You could potentially use a different argument, or even modify your argument. For example, if there were details that matched up uncannily precise to my claim that the end of religion is inevitable, the specificity would be a case against my claim that you commit the TSF.

However, even a non-fallacious argument is not automatically correct. It merely means it is a sound argument. Consideration must be dispassionately and objectively given to sound arguments from the opposing side. Even then, the stronger argument is not necessarily correct. However, we could formalize the arguments into hypotheses and apply Bayesian inference to determine which is more parsimonious.

Interbane wrote:Terms like "most likely" or "automatically" are signifiers of some sort of application of probability to your claim. There are methods for calculating such probability, Bayesian for example, but in this case there is no assignment of probability. Which means, based on your arguments alone, we can't determine if it's most likely wrong or most likely correct. All that we can say is that the argument you're using to support your claim is wrong. You could potentially use a different argument, or even modify your argument. For example, if there were details that matched up uncannily precise to my claim that the end of religion is inevitable, the specificity would be a case against my claim that you commit the TSF.

However, even a non-fallacious argument is not automatically correct. It merely means it is a sound argument. Consideration must be dispassionately and objectively given to sound arguments from the opposing side. Even then, the stronger argument is not necessarily correct. However, we could formalize the arguments into hypotheses and apply Bayesian inference to determine which is more parsimonious.

I'm sorry I posted the defense because it distracted from the topic. What are the problems with TSF?

You're the one claiming there are problems. What are they?

See below pictures.
1) These are scatter plots of 50 shots and 200 shots at a 20 x 20 foot area.
The problem is that since they are random, you might get a grouping or two, but the other non-member strikes are there calling attention to your lack of marksmanship ability.
[In my case I called attention to a specific verse about 'The Inevitable' in that forum. The objection was that it was an example of the TSF. Was that because there are 1 million other verses in the Bible which say nothing about the end of Christianity or because there are verses which contradict what I said?]

Of course the problem with the TSF is that there is not way to avoid the non-memebers as the stipulation is that the firing is random. Perhaps the shooter is blindfolded and spun around as once he starts aiming the shots are no longer random are they?

2) The application I saw of the TSF was for epidemiology. Certainly one would be more interested and accepting of groupings in those cases as the consequences of rejecting a potential pattern are greater than mistaking a pattern.

1) These are scatter plots of 50 shots and 200 shots at a 20 x 20 foot area.
The problem is that since they are random, you might get a grouping or two, but the other non-member strikes are there calling attention to your lack of marksmanship ability.

Correct. Thus when you call attention to a particular grouping, you are potentially committing the TSF. You are pointing out why it is a valid fallacy. Do you realize that? All you have to do to see through the fallacious thinking is consider all the non-member dots. However, the example you chose here is more related to clustering illusion than to the way in which you committed the TSF.

2) The application I saw of the TSF was for epidemiology. Certainly one would be more interested and accepting of groupings in those cases as the consequences of rejecting a potential pattern are greater than mistaking a pattern.

Correct. The nuances of any given example would help us determine if the researchers were committing a fallacy or not. It would depend on the experiment the epidemiologists are running. As long as they have good specificity and sensitivity, they can accurately examine the data without committing the TSF. Distinguishing between Type I and Type II errors would be critical for them.

The error you make in drawing a parallel between 'the inevitable' and your passage is that you assigned meaning to the inevitable based on the passage. You made the claim that the sequence of events which has lead up to the inevitable happening thus supports your passage. This is illogical and commits the TSF. For almost any event that happens, you could pick an ambiguous passage at will and use that as the 'bullseye' around current events. It is the same illogic used by people attempting to show that Nostradamus was a prophet. Although most times the fallacy is committed it has more to do with statistics, such as winning the lottery or playing cards or creationist arguments for abiogenesis, the fallacy applies conceptually as well. Another related fallacy is cum hoc ergo propter hoc, false cause. Although a prophecy can't exactly be seen as a cause but rather a prediction, so the TSF is more fitting.

Interbane wrote:1) These are scatter plots of 50 shots and 200 shots at a 20 x 20 foot area.
The problem is that since they are random, you might get a grouping or two, but the other non-member strikes are there calling attention to your lack of marksmanship ability.

Correct. Thus when you call attention to a particular grouping, you are potentially committing the TSF. You are pointing out why it is a valid fallacy. Do you realize that? All you have to do to see through the fallacious thinking is consider all the non-member dots. However, the example you chose here is more related to clustering illusion than to the way in which you committed the TSF.

Is there a difference? I used Excel to generate two sets of random numbers between 0 and 20. One set was for the x axis and one set for the y axis. I then did charts for 50 sets and 200 sets and plotted them. I believe that models the TSF exactly. Perhaps even better than the internet posting as it doesn't say that the 'sharpshooter' is blindfolded when he shoots. He should be to preserve the randomness of the shots shouldn't he. Otherwise he is going to start aiming at hit points and the process will no longer be random.

Anyway, I disagree with your criticism. I understand that was the point the creator was trying to get at but his/her logic fails. Go back to the title of the Fallacy. The 'Texas Sharpshooter'. Say you are driving down the road and see a sign which says: Stop and see the Texas Sharpshooter. You do and he shows you the barn, one of the scatter plots I included. The scatter plot is not going to fool anyone, least of all the sharpshooter/researcher. I think it is a poorly named and conceived fallacy.

Stahrwe wrote:quote2) The application I saw of the TSF was for epidemiology. Certainly one would be more interested and accepting of groupings in those cases as the consequences of rejecting a potential pattern are greater than mistaking a pattern.

interbane wrote:Correct. The nuances of any given example would help us determine if the researchers were committing a fallacy or not. It would depend on the experiment the epidemiologists are running. As long as they have good specificity and sensitivity, they can accurately examine the data without committing the TSF. Distinguishing between Type I and Type II errors would be critical for them.

I steered clear of the terminology because I didn't want to have to explain what they mean but for anyone who doesn't know generally researchers try to frame their hypothesis so that a mistake does the least damage. For example, you would rather tell someone they had cancer when they didn't than tell them they didn't when they did.

interbane wrote:The error you make in drawing a parallel between 'the inevitable' and your passage is that you assigned meaning to the inevitable based on the passage. You made the claim that the sequence of events which has lead up to the inevitable happening thus supports your passage. This is illogical and commits the TSF. For almost any event that happens, you could pick an ambiguous passage at will and use that as the 'bullseye' around current events. It is the same illogic used by people attempting to show that Nostradamus was a prophet. Although most times the fallacy is committed it has more to do with statistics, such as winning the lottery or playing cards or creationist arguments for abiogenesis, the fallacy applies conceptually as well. Another related fallacy is cum hoc ergo propter hoc, false cause. Although a prophecy can't exactly be seen as a cause but rather a prediction, so the TSF is more fitting.

I must disagree again. The discussion is that it is inevitable that humans will abandon Christianity. Throughout the history of the church it has been anticipated that prior to return of Jesus, there would be a turning away, a great apostacy, in fact, a universal religion would replace Christianity. The fact that it is prophecied is the context in which I agreed that it is inevitable. If my comment was an example of TSF or any other fallacy, how would one introduce the association of a prophecy with either current or anticipate events. I agree that one must not see fulfilment under every 'rock' but I was discussing it as a concept and in that context I don't see it as any different than a scientist saying that one day the Sun will explode.

Anyway, I disagree with your criticism. I understand that was the point the creator was trying to get at but his/her logic fails. Go back to the title of the Fallacy. The 'Texas Sharpshooter'. Say you are driving down the road and see a sign which says: Stop and see the Texas Sharpshooter. You do and he shows you the barn, one of the scatter plots I included. The scatter plot is not going to fool anyone, least of all the sharpshooter/researcher. I think it is a poorly named and conceived fallacy.

You're not educated enough on the logic of this fallacy to comment. For one, the story behind the namesake has only a single shot being fired, which the sharpshooter then paints a bullseye around. As I've said before, you're jumping back and forth between the clustering illusion and the TSF. Stop ignoring these details when I mention them, they make all the difference in the world.

For example, you would rather tell someone they had cancer when they didn't than tell them they didn't when they did.

That's not an example of Type I versus Type II errors.

Throughout the history of the church it has been anticipated that prior to return of Jesus, there would be a turning away, a great apostacy, in fact, a universal religion would replace Christianity.

So your problem with the TSF, when we get right down to it, is that it makes all prophecies fallacious? You would rather have faith that prophecies are true than to accept logic?

I agree that one must not see fulfilment under every 'rock' but I was discussing it as a concept and in that context I don't see it as any different than a scientist saying that one day the Sun will explode.

Observation>Induction>Extrapolation on the one hand [magical visions][guesstimates][blind chance] on the other. You don't see it as any different because you don't understand the sound epistemology behind science.