So the Lichtman Test so far

[/]

I wouldnt count the Iran Deal, which a majority oppose, or Cuba, which no one cares about or is even aware of, as major foreign policy victories.

Just like your opinions.

[Landslide Lyndon]

I wouldnt count the Iran Deal, which a majority oppose, or Cuba, which no one cares about or is even aware of, as major foreign policy victories.

Just like your opinions.

LOL. I have him on ignore so I can't see his posts unless someone quotes them. Lyndon is still being a douche I see.

And you're still a moron I see.

[Oswald Acted Alone, You Kook]

The Lichtman test is overrated.

It's worked on all but one election in the past 100 years.

Well, sure it does great when you know what result you're aiming for when you retroactively decide the subjective ones.

It was created in the 80's and has been used to predict every election from 1984 to the present. Again, all but one time it was correct. Then it was retroactively applied to elections going all the way back to 1860 and it still worked.

[/]

The Lichtman test is overrated.

It's worked on all but one election in the past 100 years.

Well, sure it does great when you know what result you're aiming for when you retroactively decide the subjective ones.

It was created in the 80's and has been used to predict every election from 1984 to the present. Again, all but one time it was correct. Then it was retroactively applied to elections going all the way back to 1860 and it still worked.

Ever head of the Redskins Rule? It has the same record. As I said, electoral truisms are hilariously inane.

[SteveRogers]

The Lichtman test is overrated.

It's worked on all but one election in the past 100 years.

Well, sure it does great when you know what result you're aiming for when you retroactively decide the subjective ones.

It was created in the 80's and has been used to predict every election from 1984 to the present. Again, all but one time it was correct. Then it was retroactively applied to elections going all the way back to 1860 and it still worked.

Um, nope. Nope nope nope. Lichtman looked at every election from 1860 to 1980 and identified common factors in order to create the model. The model cannot predict the data points that were used to derive the model in the first place. When you plug in an election from 1860 to 1980, the model spits out the right answer because it was designed (overfitted) in the first place specifically to account for each and every one of those elections.

As for whether or not the model got 2000 right, doesn't that just kind of go to show the inherent uselessness of a model for U.S. presidential elections that doesn't even purport to tell you what percentage of the popular vote a candidate will win by?

[Nym90]

This article from Nate Silver in 2011 helps explain some of the flaws of the keys--most significantly that while they may have been correct in each election regarding the popular vote winner, they are poor at predicting the popular vote percentage, and thus their success rate seems based more on luck than anything.

http://fivethirtyeight.com/features/despite-keys-obama-is-no-lock/

Lichtman's response to this is that the keys aren't designed to predict the popular vote percentage, only the winner. OK, but how can he be so confident that the tipping point voter will always vote on the basis of the keys if other voters don't seem to swinging in proportion to them? If the keys are truly accurate predictors, than the more keys a candidate has, the better they should perform in the popular vote, and there hasn't been a very strong correlation.

Here's Lichtman's response: http://fivethirtyeight.com/features/keys-to-the-white-house-historian-responds/

[windjammer]

Here's Lichtman's own analysis from last month.

1. MAN: false (Pubs hold more seats than after the 2010 midterms)
2. CON: undecided (becomes true if Clinton wraps up the race early)
3. INC: false (Obama isn't running)
4. 3RD: undecided (no forecast until next year)
5. STE: true (the economy is not headed for recession in 2016)
6. LTE: true (real per capita growth in 2013-2016 exceeds the average of the two previous terms)
7. POL: false (Obama has no policy change like Obamacare this term)
8. UNR: true (there is no sustained social unrest)
9. SCA: true (there is no scandal that touches the presidency)
10. FMF: true (Obama has avoided any notable foreign or military failure)
11. FMS: undecided (Obama has no major foreign success to date)
12. ICH: false (Clinton is not charismatic nor a military hero)
13. CCH: true (none of the Pubs are charismatic or a military heroes)

That makes 4 false and 3 undecided. Unless two more keys go false the Dems should win.

He says nothing about the Iran Deal?

[Oswald Acted Alone, You Kook]

The Lichtman test is overrated.

It's worked on all but one election in the past 100 years.

Well, sure it does great when you know what result you're aiming for when you retroactively decide the subjective ones.

It was created in the 80's and has been used to predict every election from 1984 to the present. Again, all but one time it was correct. Then it was retroactively applied to elections going all the way back to 1860 and it still worked.

Um, nope. Nope nope nope. Lichtman looked at every election from 1860 to 1980 and identified common factors in order to create the model. The model cannot predict the data points that were used to derive the model in the first place. When you plug in an election from 1860 to 1980, the model spits out the right answer because it was designed (overfitted) in the first place specifically to account for each and every one of those elections.

As for whether or not the model got 2000 right, doesn't that just kind of go to show the inherent uselessness of a model for U.S. presidential elections that doesn't even purport to tell you what percentage of the popular vote a candidate will win by?

The only reason it got 2000 wrong was because 2000 was stolen by Bush.

[SteveRogers]

The Lichtman test is overrated.

It's worked on all but one election in the past 100 years.

Well, sure it does great when you know what result you're aiming for when you retroactively decide the subjective ones.

It was created in the 80's and has been used to predict every election from 1984 to the present. Again, all but one time it was correct. Then it was retroactively applied to elections going all the way back to 1860 and it still worked.

Um, nope. Nope nope nope. Lichtman looked at every election from 1860 to 1980 and identified common factors in order to create the model. The model cannot predict the data points that were used to derive the model in the first place. When you plug in an election from 1860 to 1980, the model spits out the right answer because it was designed (overfitted) in the first place specifically to account for each and every one of those elections.

As for whether or not the model got 2000 right, doesn't that just kind of go to show the inherent uselessness of a model for U.S. presidential elections that doesn't even purport to tell you what percentage of the popular vote a candidate will win by?

The only reason it got 2000 wrong was because 2000 was stolen by Bush.

That's not actually a response to any of my points.

[pbrower2a]

The Lichtman test is overrated.

It's worked on all but one election in the past 100 years.

Well, sure it does great when you know what result you're aiming for when you retroactively decide the subjective ones.

It was created in the 80's and has been used to predict every election from 1984 to the present. Again, all but one time it was correct. Then it was retroactively applied to elections going all the way back to 1860 and it still worked.

Ever head of the Redskins Rule? It has the same record. As I said, electoral truisms are hilariously inane.

...or who wins the World Series.






At least the criteria of the Lichtman test make some sense.  The incumbent (Gerald Ford is the obvious exception) has typically shown the ability to win at the least the VP race. Economy good, no scandals, no social unrest, success in midterms, no military or diplomatic disasters, some achievements in domestic and foreign policy, no rifts within the incumbent's Party...

[○∙◄☻¥tπ[╪AV┼cVê└]

The Lichtman test is overrated.

It's worked on all but one election in the past 100 years.

Well, sure it does great when you know what result you're aiming for when you retroactively decide the subjective ones.

It was created in the 80's and has been used to predict every election from 1984 to the present. Again, all but one time it was correct. Then it was retroactively applied to elections going all the way back to 1860 and it still worked.

Um, nope. Nope nope nope. Lichtman looked at every election from 1860 to 1980 and identified common factors in order to create the model. The model cannot predict the data points that were used to derive the model in the first place. When you plug in an election from 1860 to 1980, the model spits out the right answer because it was designed (overfitted) in the first place specifically to account for each and every one of those elections.

As for whether or not the model got 2000 right, doesn't that just kind of go to show the inherent uselessness of a model for U.S. presidential elections that doesn't even purport to tell you what percentage of the popular vote a candidate will win by?

The only reason it got 2000 wrong was because 2000 was stolen by Bush.

Lichtman says it predicts the popular vote winner. Hilariously the Republicans had only 4 of the 8 keys they need to win in 1876, but since Tilden won the popular vote, Litchtman counts that as an election predicted correctly. You'd think if you were 4 keys short you wouldn't actually end up in the White House.

[bobloblaw]

Here's Lichtman's own analysis from last month.

1. MAN: false (Pubs hold more seats than after the 2010 midterms)
2. CON: undecided (becomes true if Clinton wraps up the race early)
3. INC: false (Obama isn't running)
4. 3RD: undecided (no forecast until next year)
5. STE: true (the economy is not headed for recession in 2016)
6. LTE: true (real per capita growth in 2013-2016 exceeds the average of the two previous terms)
7. POL: false (Obama has no policy change like Obamacare this term)
8. UNR: true (there is no sustained social unrest)
9. SCA: true (there is no scandal that touches the presidency)
10. FMF: true (Obama has avoided any notable foreign or military failure)
11. FMS: undecided (Obama has no major foreign success to date)
12. ICH: false (Clinton is not charismatic nor a military hero)
13. CCH: true (none of the Pubs are charismatic or a military heroes)

That makes 4 false and 3 undecided. Unless two more keys go false the Dems should win.

He doesn't conceder The Iran Deal or Cuba FMS?

The Iran deal is unpopular and Cuba is irrelevant.

[muon2]

Here's Lichtman's own analysis from last month.

1. MAN: false (Pubs hold more seats than after the 2010 midterms)
2. CON: undecided (becomes true if Clinton wraps up the race early)
3. INC: false (Obama isn't running)
4. 3RD: undecided (no forecast until next year)
5. STE: true (the economy is not headed for recession in 2016)
6. LTE: true (real per capita growth in 2013-2016 exceeds the average of the two previous terms)
7. POL: false (Obama has no policy change like Obamacare this term)
8. UNR: true (there is no sustained social unrest)
9. SCA: true (there is no scandal that touches the presidency)
10. FMF: true (Obama has avoided any notable foreign or military failure)
11. FMS: undecided (Obama has no major foreign success to date)
12. ICH: false (Clinton is not charismatic nor a military hero)
13. CCH: true (none of the Pubs are charismatic or a military heroes)

That makes 4 false and 3 undecided. Unless two more keys go false the Dems should win.

He doesn't conceder The Iran Deal or Cuba FMS?

The Iran deal is unpopular and Cuba is irrelevant.

I think it is more that Lichtman does not consider the Iran deal in the same league as Carter's Camp David Accords or the renormalization with Cuba in the same league as Nixon's trip to China. His threshold for most of the keys is quite high.

[BlueSwan]

It is intuitively unclear why Key 4 even exists. It doesn't follow logically that a strong third party candidate would always hurt the incumbent party. Infact, in 1996, Ross Perot probably helped Clinton win (even if he didn't need the help). This year, if we have Clinton vs Bush or Clinton vs Rubio and Trump decides to run as an independent, that definitely helps the incumbent party.

Outside of that, the keys all seem like relevant factors, but the fact that they aren't weighted makes them fairly useless. A 12 to 1 split wouldn't make me confident of an incumbent win, if that one false Key was an absolutely disasterous economy or a major personal scandal occuring close to the election day.

[muon2]

It is intuitively unclear why Key 4 even exists. It doesn't follow logically that a strong third party candidate would always hurt the incumbent party. Infact, in 1996, Ross Perot probably helped Clinton win (even if he didn't need the help). This year, if we have Clinton vs Bush or Clinton vs Rubio and Trump decides to run as an independent, that definitely helps the incumbent party.

Outside of that, the keys all seem like relevant factors, but the fact that they aren't weighted makes them fairly useless. A 12 to 1 split wouldn't make me confident of an incumbent win, if that one false Key was an absolutely disasterous economy or a major personal scandal occuring close to the election day.

Lichtman recognizes that some keys have more individual predictive power than others and that many keys are redundant, in that there are subsets that also predict the outcome. He explains that they started with a set of 30 factors that could be answered yes/no. It's based on a forecasting technique used by geophysicists to predict earthquakes. A series of statistical and pattern recognition filters were applied to arrive at the final set of 13 keys.

[Torie]

It is intuitively unclear why Key 4 even exists. It doesn't follow logically that a strong third party candidate would always hurt the incumbent party. Infact, in 1996, Ross Perot probably helped Clinton win (even if he didn't need the help). This year, if we have Clinton vs Bush or Clinton vs Rubio and Trump decides to run as an independent, that definitely helps the incumbent party.

Outside of that, the keys all seem like relevant factors, but the fact that they aren't weighted makes them fairly useless. A 12 to 1 split wouldn't make me confident of an incumbent win, if that one false Key was an absolutely disasterous economy or a major personal scandal occuring close to the election day.

Lichtman recognizes that some keys have more individual predictive power than others and that many keys are redundant, in that there are subsets that also predict the outcome. He explains that they started with a set of 30 factors that could be answered yes/no. It's based on a forecasting technique used by geophysicists to predict earthquakes. A series of statistical and pattern recognition filters were applied to arrive at the final set of 13 keys.

Are there really enough data points, with enough similarities, to make this exercise statistically meaningful at all, or is just more of a curiosity? Obviously some of the factors have something to do with how popular the incumbent regime might be, all things being equal, but that gets one only so far.

[muon2]

It is intuitively unclear why Key 4 even exists. It doesn't follow logically that a strong third party candidate would always hurt the incumbent party. Infact, in 1996, Ross Perot probably helped Clinton win (even if he didn't need the help). This year, if we have Clinton vs Bush or Clinton vs Rubio and Trump decides to run as an independent, that definitely helps the incumbent party.

Outside of that, the keys all seem like relevant factors, but the fact that they aren't weighted makes them fairly useless. A 12 to 1 split wouldn't make me confident of an incumbent win, if that one false Key was an absolutely disasterous economy or a major personal scandal occuring close to the election day.

Lichtman recognizes that some keys have more individual predictive power than others and that many keys are redundant, in that there are subsets that also predict the outcome. He explains that they started with a set of 30 factors that could be answered yes/no. It's based on a forecasting technique used by geophysicists to predict earthquakes. A series of statistical and pattern recognition filters were applied to arrive at the final set of 13 keys.

Are there really enough data points, with enough similarities, to make this exercise statistically meaningful at all, or is just more of a curiosity? Obviously some of the factors have something to do with how popular the incumbent regime might be, all things being equal, but that gets one only so far.

Statistically meaningful - probably yes. Statistically perfect - highly unlikely. It used 31 elections with 30 factors on each, that's 930 input points. The model has been successful in 8 subsequent elections with one minor change to how short term economy is calculated (ie use the status as understood during the campaign, not the official NBER value that comes later). Getting 8 out of 8 right from predetermined factors, not based on polls, is 1/64 or about 2%. Getting 7 out of 8 right is about 14% likely, which is still pretty good from a list of predetermined factors.

[Torie]

It is intuitively unclear why Key 4 even exists. It doesn't follow logically that a strong third party candidate would always hurt the incumbent party. Infact, in 1996, Ross Perot probably helped Clinton win (even if he didn't need the help). This year, if we have Clinton vs Bush or Clinton vs Rubio and Trump decides to run as an independent, that definitely helps the incumbent party.

Outside of that, the keys all seem like relevant factors, but the fact that they aren't weighted makes them fairly useless. A 12 to 1 split wouldn't make me confident of an incumbent win, if that one false Key was an absolutely disasterous economy or a major personal scandal occuring close to the election day.

Lichtman recognizes that some keys have more individual predictive power than others and that many keys are redundant, in that there are subsets that also predict the outcome. He explains that they started with a set of 30 factors that could be answered yes/no. It's based on a forecasting technique used by geophysicists to predict earthquakes. A series of statistical and pattern recognition filters were applied to arrive at the final set of 13 keys.

Are there really enough data points, with enough similarities, to make this exercise statistically meaningful at all, or is just more of a curiosity? Obviously some of the factors have something to do with how popular the incumbent regime might be, all things being equal, but that gets one only so far.

Statistically meaningful - probably yes. Statistically perfect - highly unlikely. It used 31 elections with 30 factors on each, that's 930 input points. The model has been successful in 8 subsequent elections with one minor change to how short term economy is calculated (ie use the status as understood during the campaign, not the official NBER value that comes later). Getting 8 out of 8 right from predetermined factors, not based on polls, is 1/64 or about 2%. Getting 7 out of 8 right is about 14% likely, which is still pretty good from a list of predetermined factors.

Unless the model was based on looking at the past and getting the best fit. That is data mining. If the model then worked after it was created in subsequent elections, and a fair number of them, that is more impressive. How about the matter that the model falls apart when it comes to margins, as opposed to who got the most votes? How would that theoretically be explained?

[Torie]

OK, and then it is just an accident the popularity thing never actually changed who got the most votes based on the model, or accidentally always coincided with who got the most votes based on the model. In that case, the model just accidentally, even if not data mined based on the past to get the best fit, turns out to have worked.

[windjammer]

Torie, this model is quite reliable:
Lichtman predicted the victory of Bush in 88 when everyone was believing he was going to lose, and he predicted the victory of Clinton when everyone was thinking Bush was going to win.
And these predictions were made much before the election.

[Nym90]

The Lichtman test is overrated.

It's worked on all but one election in the past 100 years.

Well, sure it does great when you know what result you're aiming for when you retroactively decide the subjective ones.

It was created in the 80's and has been used to predict every election from 1984 to the present. Again, all but one time it was correct. Then it was retroactively applied to elections going all the way back to 1860 and it still worked.

Um, nope. Nope nope nope. Lichtman looked at every election from 1860 to 1980 and identified common factors in order to create the model. The model cannot predict the data points that were used to derive the model in the first place. When you plug in an election from 1860 to 1980, the model spits out the right answer because it was designed (overfitted) in the first place specifically to account for each and every one of those elections.

As for whether or not the model got 2000 right, doesn't that just kind of go to show the inherent uselessness of a model for U.S. presidential elections that doesn't even purport to tell you what percentage of the popular vote a candidate will win by?

The only reason it got 2000 wrong was because 2000 was stolen by Bush.

Lichtman says it predicts the popular vote winner. Hilariously the Republicans had only 4 of the 8 keys they need to win in 1876, but since Tilden won the popular vote, Litchtman counts that as an election predicted correctly. You'd think if you were 4 keys short you wouldn't actually end up in the White House.

Another good example is 1960, when Nixon also had only four keys in his favor yet only lost the popular vote by 0.2%, and it can be argued that he actually did win it depending upon how you choose to allocate the Democratic vote in Alabama.

Pretty hard to believe that it would have taken four more keys swinging his way to persuade the 0.1% of the electorate he would have needed to win it.

[OSR stands with Israel]

Lets do 1992

Key 1: Party Mandate: Republicans had 177 seats in 1986 compared to 1967 in 1990(DNC win)
Key 2: No significant party challenge: True as Bush won every primary in 1992(GOP win)
Key 3: Incumbency: GOP win
Key 4: Third Party:  Perot took 19% of the vote( DNC win)
Key 5: Short Term Economy:  economy was in recission(DNC win)
Key 6: Long Term Economy : GDP grew by .95% per year under HW Bush compared to 2.65% a year under Reagan(DNC win)
Key 7: Policy Change: No Major Policy Change So  False(DNC win)
Key 8: Social Unrest: No sustained social unrest during HW term(GOP win)
Key 9: Scandal: Iran- Contra had fizzled out and wasnt considered big by 92 (GOP win)
Key 10: Foreign/ Military Failure: None (GOP win)
Key 11: Foreign / Military Success: Gulf War, cold war ending peacefully (GOP win)
Key 12: Incumbent Charisma: Bush had no charisma(DNC win)
Key 13: Challenge Charisma:: Clinton had lots of Charisma(DNC win)

Democrats had 7 Keys go for them in 92 and 6 keys went for the Republicans.

[Nym90]

The GOP actually had seven in 1992 as Iran-Contra didn't rise to the level of a major enough scandal to turn key 9. But the incumbent party needs 8 keys to win, and thus they still fell one short.

Also I think you got keys 7 and 13 backwards. No major policy change was a loss for the GOP, but Clinton didn't qualify as charismatic.

It is harder to turn some of the keys than many would assume.

[Oswald Acted Alone, You Kook]

Bumped because I was thinking:

Considering that nearly Sanders won Iowa, could he turn out to  qualify as a serious competitor, thus turning a key for the Republicans?

By the way, after looking up info, the only candidates that qualified for #13 (for Lichtman himself) are Roosevelt, Eisenhower, Kennedy, and Reagan. Trump is not as charismatic as either of those guys, no matter what the people at AAD will tell you.

[Oswald Acted Alone, You Kook]

Now that the nominees are pretty much in the bag, what does it look like now?