One Incontrovertible, Unequivocal, Undeniable Fact Which Refutes the Diary

[Iconoclast]

I have, a few times, completely agreed that the probability of everyday events are actually extremely low, and have indicated that it would not be a shocker to find that yet again.

RJ, are you paying attention here? Even everyday events coinciding have probabilities attached to them and - as Jeff notes above - some of these probabilities are "extremely low". What Jeff misses out is what I guess he assumes we all know, namely that these events have to be mutually-exclusive (that is, non-causal - neither one has caused the other to occur). If they are not mutually-exclusive, then they could potentially have caused the other to occur and that breaks the rules of simple probability theory which assumes that events which occur are entirely independent of each other.

I've suggested if all one wants to do is examine the improbability of things that happen, then I would argue his flat distribution model needs revising, etc.

RJ, are you paying attention here? Jeff is acknowledging that the improbability of mutually-exclusive events occurring 'simultaneously' can be examined. They would be examined by probability theory which - I think - I may have said once or twice during this torturous debate? What Jeff has doggedly done is refuse to apply basic probability principles because he wants to be uber-literal and attempt to factor-in known and unknown variables which may or may not have influenced the events we observed. I wish he'd just answered the question - even equivocally - with a number and said "Yes, if you push me, the chances of those two events happening for the first time on the record on March 9, 1992 purely by chance alone given that Maybrick expired on May 11, 1889 are 1-in-37,557 or there or thereabouts, given that we need to make allowances for known and unknown variables which are fundamentally impossible for us to quantify but I'm going to obfuscate the case anyway in case I ever get sued over this or accused of being a Maybrickian". Jeff would then add "And you can't read anything into this however low the probability is because improbable events coincide all the time" to which I would have replied "In fairness, that was not the point I was asking you to comment on”.

So if that is all he is interested in, we agree, life is improbable.

RJ, are you paying attention here? Jeff is acknowledging that life is full of improbable events happening. Improbability is not a different concept to probability – it’s just the reverse expression of it. The key thing for you, RJ, is noting that Jeff is talking about improbable events. This means that he understands that probability can be brought to bear on events.

The only thing I've been saying is that those probabilities are irrelevant with regards to the diary provenance, and no matter how rare the above is it cannot be misconstrued as support for anything diary related.

If only that had been what we were debating, though, Jeff! I guess you can’t have been briefed properly otherwise you’d have just stuck to dodging the specific probability whilst acknowledging that events are subject to simple probability theory.

Some of Ike's statements suggest he sees the low prob.

I think you must be joking there, Jeff? That’s exactly what my statements have been saying. The ‘low prob’ is an eye-watering 1-in-37,557.

All events, when we calculate them this way, end up with low probabilities, etc..

I’m not clear what you’re arguing here Jeff. Not all events have low probabilities. Really not sure what you were referring to here. Two known events occurring for the first time on the same day by pure chance alone three days after they first could have occurred – in the absence of any other information – have a 1-in-3 chance of having occurred on the third day. They had a 1-in-3 chance of having occurred on the second day, and a 1-in-3 chance of having occurred on the first day. Probability theory requires that the sum of all possibilities must equal 1 (100%, certainty), like the 6 sides of a dice which each have a 1-in-6 chance of coming-up on any given throw, and therefore sum to 6-in-6 which is 1. If two events are known to have occurred, etc., and they end up occurring on the third day, then by the third day they had a 1-in-3 chance of occurring on the first, second, or third day. On the second day, if two events are known to have occurred then the chances of them occurring on the same day are 1-in-2 for the second day and 1-in-2 for the first day. On the first day, if two events are known to have occurred then the chances of them occurring on the same day are 1-in-1 for the first day which is certainty, and that would make sense, wouldn’t it, if two events are known to have occurred and only one day had passed? It must have been that first day.

Generally speaking, the two events won’t happen on the first day they could have occurred (although they could). If they occurred on the 18^th day they could have occurred, then the probability that they occurred on the same day is 1-in-18, which are by then the same odds as it happening on the 17^th, 16^th, 15^th,14^th, etc., day. RJ, probability theory does not permit you to say that we know the two events happened on March 9, 1992 therefore we’ll calculate it all from there and exclude anything which had passed before it. Probability theory would ask when the first day was they could have happened and then worked out how many days had passed until they did occur, which – as we all know – was 37,557 for March 9, 1992, therefore odds of 1-in-37,557, as indeed it was for March 8, 1992 by March 9, 1992, March 7, 1992 by March 9, 1992, March 6, 1992 by March 9, 1992, March 5, 1992 by March 9, 1992, etc.. Jeff said that statistics could be counter-intuitive and perhaps this is an example of what he meant?

RJ naively thinks that “The intervening years are entirely irrelevant. They play no part in the supposedly astonishing coincidence”. Well, on the contrary, once we know that two or more events have coincided, to understand the likelihood of those events occurring on the same day by chance alone are 1-in-days-passed-since-they-could-occur. Jeff is definitely not arguing that all probabilities are the same – we know this from his ‘fair’ coin tossing thought experiment – so he knows that The Miraculous Day occurring on March 9, 1992 has a low probability, and he also knows that The Miraculous Day occurring on March 9, 21992 would be a miracle of miracles. In truth, it would be a 1-in-7,337,557 chance. Just because it would be the same outcome, RJ, does not mean it therefore has the same probability of having happened on that day. Indeed, in this latter scenario, the odds of The Miraculous Day happening on March 9, 1992 would have therefore risen from 1-in-37,557 to 1-in-7,337,557. Remember, dear readers, these are not probabilities that The Miraculous Day would happen, but simply the probabilities that they would happen on the same day once we knew they had happened. There is a whole world of difference between these two perspectives.

If, however, he recognises that no matter how improbable his calculation is, that it means squat all with respect to diary coming from the floorboards, then we've agreed all along.

RJ, are you paying attention here? Jeff is acknowledging that – at very least – the methodology applied here is correct (though I’m sure that he would want to nuance the value to account for the possible impact of certain possible variables), but he is qualifying the use of this valid use of probability theory to argue that however improbable our odds were, we couldn’t use that alone as evidence that the scrapbook came from Maybrick’s house on March 9, 1992. This is an error of logic I may very well have fallen victim to, and I suspect it's human nature to do so, though I politely remind everyone again that that is a different debate altogether and not the one I was seeking the insight of a statistics lecturer for. Once our Stage 1 analysis (not a statistical term, by the way) has thrown up a probability value so low that our alarm bells are ringing, we can then go back to the scene and get stuck into Stage 2 by starting to look for clues which might help us to understand if in fact chance alone had nothing to do with what was observed.

Ike

[rjpalmer]

RJ, are you paying attention here?

Yes, I am paying attention, Ike (but I won't be for long), and it clear that you have a seemingly limitless capacity for self-deception.

RJ naively thinks that “The intervening years are entirely irrelevant. They play no part in the supposedly astonishing coincidence”. Well, on the contrary, once we know that two or more events have coincided, to understand the likelihood of those events occurring on the same day by chance alone are 1-in-days-passed-since-they-could-occur. Jeff is definitely not arguing that all probabilities are the same – we know this from his ‘fair’ coin tossing thought experiment – so he knows that The Miraculous Day occurring on March 9, 1992 has a low probability, and he also knows that The Miraculous Day occurring on March 9, 21992 would be a miracle of miracles. In truth, it would be a 1-in-7,337,557 chance. Just because it would be the same outcome, RJ, does not mean it therefore has the same probability of having happened on that day. Indeed, in this latter scenario, the odds of The Miraculous Day happening on March 9, 1992 would have therefore risen from 1-in-37,557 to 1-in-7,337,557. Remember, dear readers, these are not probabilities that The Miraculous Day would happen, but simply the probabilities that they would happen on the same day once we knew they had happened. There is a whole world of difference between these two perspectives.

The above statement is bonkers, Ike, as already explained in Post #6749.

The fact that the earth revolves around the sun once every 365.25 days is irrelevant. The fact that the historic James Maybrick died on 11 May 1889 is also irrelevant.

They play no role in the matter.

If we were to take your methods seriously, it would mean that if Keith had a time machine, and was at Scotland Yard in 1889, and noticed that Nurse Yapp had come forward with the confession of James Maybrick on 15 May 1889, and later learned that the floorboards of Battlecrease were lifted on the same day, it would only be a 4 to 1 coincidence. Barely worthy of note. No more strange or unlikely than someone guessing correctly whether a random playing card was a heart, spade, club, or diamond. Oh, perhaps slightly interesting, but not important enough for Inspector Keith to take a train to Liverpool to investigate. Even though the floorboards coincided with the appearance of James Maybrick's confession, it would be a matter of little consequence (are you okay with that, Keith?)

But if the exact circumstance happened instead on 9 March 1992, it would now be a 37,577 to 1 coincidence, on the strength that the earth revolves around the sun, and 37,577 days have passed. The circumstances are the same--it's earth's voyage through the solar system and the date of Maybrick's death that matters (!)

(Thank goodness the earth revolves around the sun)

And, if Ghost of Future Keith notices this coincidence in the year 21,992, it is will now be a 7,337,577 to 1 coincidence, even though he is noticing EXACTLY the same circumstances, and the exact same suspicions are aroused in his mind.

Wow, Ike. Just wow.

As Jeff stated earlier, statistics can be shockingly counter-intuitive, but I didn't realize that fact would be this hard to understand.

Good luck, Ike. I can offer no more insight.

R P ~ aka the naïve drunken baker

[JeffHamm]

RJ, are you paying attention here? Even everyday events coinciding have probabilities attached to them and - as Jeff notes above - some of these probabilities are "extremely low". What Jeff misses out is what I guess he assumes we all know, namely that these events have to be mutually-exclusive (that is, non-causal - neither one has caused the other to occur). If they are not mutually-exclusive, then they could potentially have caused the other to occur and that breaks the rules of simple probability theory which assumes that events which occur are entirely independent of each other.



RJ, are you paying attention here? Jeff is acknowledging that the improbability of mutually-exclusive events occurring 'simultaneously' can be examined. They would be examined by probability theory which - I think - I may have said once or twice during this torturous debate? What Jeff has doggedly done is refuse to apply basic probability principles because he wants to be uber-literal and attempt to factor-in known and unknown variables which may or may not have influenced the events we observed. I wish he'd just answered the question - even equivocally - with a number and said "Yes, if you push me, the chances of those two events happening for the first time on the record on March 9, 1992 purely by chance alone given that Maybrick expired on May 11, 1889 are 1-in-37,557 or there or thereabouts, given that we need to make allowances for known and unknown variables which are fundamentally impossible for us to quantify but I'm going to obfuscate the case anyway in case I ever get sued over this or accused of being a Maybrickian". Jeff would then add "And you can't read anything into this however low the probability is because improbable events coincide all the time" to which I would have replied "In fairness, that was not the point I was asking you to comment on”.



RJ, are you paying attention here? Jeff is acknowledging that life is full of improbable events happening. Improbability is not a different concept to probability – it’s just the reverse expression of it. The key thing for you, RJ, is noting that Jeff is talking about improbable events. This means that he understands that probability can be brought to bear on events.

If only that had been what we were debating, though, Jeff! I guess you can’t have been briefed properly otherwise you’d have just stuck to dodging the specific probability whilst acknowledging that events are subject to simple probability theory.



I think you must be joking there, Jeff? That’s exactly what my statements have been saying. The ‘low prob’ is an eye-watering 1-in-37,557.



I’m not clear what you’re arguing here Jeff. Not all events have low probabilities. Really not sure what you were referring to here. Two known events occurring for the first time on the same day by pure chance alone three days after they first could have occurred – in the absence of any other information – have a 1-in-3 chance of having occurred on the third day. They had a 1-in-3 chance of having occurred on the second day, and a 1-in-3 chance of having occurred on the first day. Probability theory requires that the sum of all possibilities must equal 1 (100%, certainty), like the 6 sides of a dice which each have a 1-in-6 chance of coming-up on any given throw, and therefore sum to 6-in-6 which is 1. If two events are known to have occurred, etc., and they end up occurring on the third day, then by the third day they had a 1-in-3 chance of occurring on the first, second, or third day. On the second day, if two events are known to have occurred then the chances of them occurring on the same day are 1-in-2 for the second day and 1-in-2 for the first day. On the first day, if two events are known to have occurred then the chances of them occurring on the same day are 1-in-1 for the first day which is certainty, and that would make sense, wouldn’t it, if two events are known to have occurred and only one day had passed? It must have been that first day.

Generally speaking, the two events won’t happen on the first day they could have occurred (although they could). If they occurred on the 18^th day they could have occurred, then the probability that they occurred on the same day is 1-in-18, which are by then the same odds as it happening on the 17^th, 16^th, 15^th,14^th, etc., day. RJ, probability theory does not permit you to say that we know the two events happened on March 9, 1992 therefore we’ll calculate it all from there and exclude anything which had passed before it. Probability theory would ask when the first day was they could have happened and then worked out how many days had passed until they did occur, which – as we all know – was 37,557 for March 9, 1992, therefore odds of 1-in-37,557, as indeed it was for March 8, 1992 by March 9, 1992, March 7, 1992 by March 9, 1992, March 6, 1992 by March 9, 1992, March 5, 1992 by March 9, 1992, etc.. Jeff said that statistics could be counter-intuitive and perhaps this is an example of what he meant?

RJ naively thinks that “The intervening years are entirely irrelevant. They play no part in the supposedly astonishing coincidence”. Well, on the contrary, once we know that two or more events have coincided, to understand the likelihood of those events occurring on the same day by chance alone are 1-in-days-passed-since-they-could-occur. Jeff is definitely not arguing that all probabilities are the same – we know this from his ‘fair’ coin tossing thought experiment – so he knows that The Miraculous Day occurring on March 9, 1992 has a low probability, and he also knows that The Miraculous Day occurring on March 9, 21992 would be a miracle of miracles. In truth, it would be a 1-in-7,337,557 chance. Just because it would be the same outcome, RJ, does not mean it therefore has the same probability of having happened on that day. Indeed, in this latter scenario, the odds of The Miraculous Day happening on March 9, 1992 would have therefore risen from 1-in-37,557 to 1-in-7,337,557. Remember, dear readers, these are not probabilities that The Miraculous Day would happen, but simply the probabilities that they would happen on the same day once we knew they had happened. There is a whole world of difference between these two perspectives.



RJ, are you paying attention here? Jeff is acknowledging that – at very least – the methodology applied here is correct (though I’m sure that he would want to nuance the value to account for the possible impact of certain possible variables), but he is qualifying the use of this valid use of probability theory to argue that however improbable our odds were, we couldn’t use that alone as evidence that the scrapbook came from Maybrick’s house on March 9, 1992. This is an error of logic I may very well have fallen victim to, and I suspect it's human nature to do so, though I politely remind everyone again that that is a different debate altogether and not the one I was seeking the insight of a statistics lecturer for. Once our Stage 1 analysis (not a statistical term, by the way) has thrown up a probability value so low that our alarm bells are ringing, we can then go back to the scene and get stuck into Stage 2 by starting to look for clues which might help us to understand if in fact chance alone had nothing to do with what was observed.

Ike

Ike, what you seen to think are admissions are what I've been saying all along. Glad to see you finally comprehended.

However, you seem to think the simplistic examples used to teach probability theory can just be transplanted into the real world they can't because the real world is far more complicated. House work isn't a flat distribution, for example. So if all we're debating is the probability of Battlecrease having floor work donrme on that day then use of a flat distribution will be wrong. Since we don't know how that increase over time function works, we need an estimate for the probability of work around that time.
That's where the 14 days of work in the past year come in ( I think I've seen that said, perhaps I'm misremembering, but again, it's the underlying principle that matters here). So 14/365.25 is about 3.8%. now, that's a bit better, but there are seasonal changes in work rates, some repairs are more common in spring and fall, for example. We would need to do some research, work out how common heater installation is over the year, etc and we need to factor in the age of the house still for that (is heater installation more probable if the house is old? Etc)

We need to know if it has changed ownership recently (increases probability of work being done), and so forth.

See, this is because, as you recognise above and I've been saying all along, there are things that cause the probabilities to change. We need to factor them in when dealing with Battlecrease. Your flat distribution pretends house repairs are entirely random, but that's only how things are presented in class (and why we use coin flips, because people already think of them as random so even if they misunderstood that part of the lesson it is hoped it won't matter. But clearly it does).

So, to work out the probability for heaters being installed on that day, we need a lot more information about the rate of work being done on the house around that time (that gives us a base rate relavant to our time of interest to start from). Then, we need a lot of information about heater installations over the course of the year because heater installation in spring may turn out to be more probable than, say, heater installation over Christmas/new year, even after equating for different time spans.

Basically, to get anything like an accurate probability we need far more information from data than we have. It certainly isn't 1/37000+ though. That's based upon simplistic teaching examples, and this is a real world problem, with a lot of complexities.

And given the rate of work being done at around the time of interest seems to be 14 days out if a year, then the actual probability is more likely to be in the vicinity of 3.8% further adjusted by the seasonal rate of heater installation. Basically, if it is the busy season for heater installation, the the probability will be more than 3.8%, but if it is a slow time of year, the probability will go down. Other factors will also come into play, change of ownership for example
now we could argue over what factors to include, and so forth (we've already agreed that prob of work inreases with property age, so we both agree your flat distribution is wrong) but that seems a strange topic for a forum about JtR?

Why would we care what the probability for heater installation is? Such a probability has no bearing on anything at all related to JtR so I'm confused as to why anyone is concerned about house renovation probability?

See, I'm pretty sure this is "of interest" because it is thought to indicate that the diary cane from under the floorboards if this probability is low.

First, it probably is lowish, but more like 3-4% low. It is definately not 1/37000+, and 1/18 I've seen too, which I have no idea what that is based on but comes to about 5.5%, which seems a bit high but is at least closer to the mark.

But who cares! Whether this probability is vanishingly small or all but guaranteed, it is completely useless information about whether or not the diary was found under the floorboards. It has nothing to do with JtR. And that has been my major focus all along.

So, if my major focus is not your interest in house renovations, the we agree this whole topic has no bearing on JtR and diary discussions.

If you think the above house renovation probability value does, if very small, indicates the probability that there is a connection with Mike making his phone call, then despite my best efforts, you are still wrong on that point.

I'm not interested in home renovation probabilities, so if that's all we're talking about here, then I wish you all a fun time.

If anyone thinks the value being discussed has any place to play in evaluating the diary, please stop. It doesn't..


.- Jeff

[Iconoclast]

First, it probably is lowish, but more like 3-4% low. It is definately not 1/37000+, and 1/18 I've seen too, which I have no idea what that is based on but comes to about 5.5%, which seems a bit high but is at least closer to the mark.

.- Jeff

Jeff,

Two questions:

1) Why do you keep linking the two events we've been talking about to the discovery of the diary? I've now said a few times that this is NOT the issue being discussed here and NOT the reason Herlock asked you to comment but you just keep coming back to it; and

2) What was the rationale by which you calculated your apparently very arbitrary 3-4% This cuts right to the heart of the matter so you really owe it to us to explain where you got it from, please.

Ike

[JeffHamm]

Jeff,

Two questions:

1) Why do you keep linking the two events we've been talking about to the discovery of the diary? I've now said a few times that this is NOT the issue being discussed here and NOT the reason Herlock asked you to comment but you just keep coming back to it; and

2) What was the rationale by which you calculated your apparently very arbitrary 3-4% This cuts right to the heart of the matter so you really owe it to us to explain where you got it from, please.

Ike

1). I've mention in past post's comments you've made that seem to imply you connect house renovation rates like this to evaluations of JtR connections. I can't be bothered to go back and repost them. If you didn't read them then, you won't now.

2) I fully explained it in my post. It's clear you are not reading but skimming. If you want to know my reasoning, read my original post.

Anyway, you acknowledge the entire discussion has nothing to do with evaluating the diary in any way, and discussing house renovation probability distributions is not very appealing, particularly as we wish need to gather an awful lot of I formation about the industry before we could really get into it. It also is so far off topic for a JtR forum that it probably better belongs under pub talk as a general interest topic for home owners. That would help avoid people confusing these probabilities you are discussing as being In any way relevant to any discussions I volving JtR.

- Jeff

[rjpalmer]

It is definitely not 1/37000+, and 1/18 I've seen too, which I have no idea what that is based on but comes to about 5.5%, which seems a bit high but is at least closer to the mark.

Hi Jeff,

Not that it matters, and I agree that this conversation has run its course, but the 1/18 probability was given by a former poster named David Orsam when this discussion first erupted four years ago.

All he was trying to express is that it would be somewhat unremarkable for the owner of Battlecrease to have work done on his house on any random day—the same day when a person might call a literary agency.

By way of example, he used the only available data—receipts & timesheets from the year 1992. During that year, there were 14 days when workmen were at Battlecrease. (There may have been more, but those are all we know about).

A person can only call a literary agency on weekdays and non-holidays, and there were 254 such days in 1992.

14 into 254 leaves the odds at 18 to 1.

I have noted before (and I suspect that Orsam would agree) this is open to the objection that he used a limited sample—the year 1992--but he didn’t disguise this fact, and it was the only data he had. Maybe 1992 was not an average year for workmen at Battlecrease. A ten-year average would produce a different number, and as you note, as a house ages, more upgrades, etc. are likely to occur, so it become hopelessly complicated.

Then again, he was also being fairly generous as far as I am concerned. What might qualify as an opportunity for an unproven provenance story to emerge? Exterminators? Visitors to the house? Plumbers? A house inspector? A rummage sale? A break-in during the night? The diarist doesn’t even state that he planned to leave the diary at Battlecrease. There were other buildings associated with James Maybrick.

Lots of random events would qualify.

Unless it can be proven that something genuinely occurred on 9 March 1992 that links to a phone call to a literary agent—all we are really attempting to calculate is the opportunity for human suspicions to develop when someone notices two events coincided during the same general timeframe.

How does one calculate such a thing?

It can’t be done, but I suspect that 18 to 1 is closer to the mark than 37,577 to 1.

Thanks.

[Iconoclast]

1). I've mention in past post's comments you've made that seem to imply you connect house renovation rates like this to evaluations of JtR connections. I can't be bothered to go back and repost them. If you didn't read them then, you won't now.

I think you weren't well briefed, Jeff, and you've launched in here addressing the wrong issue.

2) I fully explained it in my post. It's clear you are not reading but skimming. If you want to know my reasoning, read my original post.

So you've done the 14 days in 1992 thing? What was so magical about 1992 that you would only address this one year out of the 103 which had passed since Maybrick died? Was 1992 the only year Maybrick's floorboards could have come up? Was 1992 the only year someone could have contacted a literary agent with a diary of Jack the Ripper purportedly written by James Maybrick? Your 3.8% probability confirms that you believe this to be so I know that we're done here. It is as risible as Lord Orsam's terrible 1-in-18.

When I heard we were getting an experienced university lecturer in statistics, I honestly though "Result". Even if I was wrong, we'd at least get to the bottom of the simple probability question which was in dispute. Other posters are welcome to absorb everything you have posted and conclude that you have fully explained how probability theory works.

For those posters who are lost and confused by this exchange, all you need to know to answer the simple question that was asked is that Maybrick died on May 11, 1889, and his floorboards could have come up on the record on any of the 37,557 days which followed his death. As it happens, they came up on the 37,557th - March 9, 1992.

Similarly, someone could have contacted a literary agent with a diary of Jack the Ripper purportedly written by James Maybrick on the record on any of the 37,557 days which followed his death. As it happens, they came up on the 37,557th - March 9, 1992.

The odds those two events happening by chance alone on the same day are therefore 1-in-37,557 in simple probability theory.

For those who are lost and confused by this, that's all I'm pointing-out at this stage and we're done here on this point.

Ike

[JeffHamm]

Hi Ike,

I was briefed by your posts. You phase things like you think the prob. we are talking about has a connection to something meaningful about JtR, etc. So if I appear to have missed your point, then perhaps you need to monitor your phasing. Despite my telling you where and how I got my information, you try to phrase it as if I were briefed by others, etc it's a technique of sophistry I see frequently, and which reflects an interest creating a particular impression regardless of the fact it is a false one.
one.And yes, the repair rate around the time of the discovery is what matters. The fact it was not worked in much in 1910 has little meaning to us, who are interested in repair rates of a house 80 years later. As you pointed out, the calculation requires we know if events are independent or not. Work and revocations tend to cluster, so they are not independent events. If I know Battlecrease has been worked in 14 times in the past year, I probably should expect a higher probability of a further repair/renovation.
.your simple model of equally spreading renovation rates equally over time is unrealistic so tmit is the local temporal rate/Freq if repair that will be most accurate to use as our base rate.

Anyway, as I say, I am not particularly interested in architectural issues like renovation rates in old buildings. Also, I'm familiar enough with the boards to recognise the difference between the unconvinced and the unconvincable. If you want to divide one by a bigger number, go ahead, and you can even use that result in sentences where you imply, or even outright state, that this low number means there is likely to be a causal link between the floorboards and Mike's calling a publisher (because we both know that is the goal here, to create that impression).

None of those are correct though. The flat distribution you want to use is overly simplistic. Repairs are not independent events, they tend to cluster, so knowing there were 14 repairs within the last year gives us a better estimate of the rate of work at the time in question. And finally, none of this relates in any way to anything involving Mike's phonecall.

So, if you want to know what I think and why, try reading my posts. If you just want to keep posting 1/37000+, then go ahead but I can't see the point in answering questions only to find the question asker isn't going to read my reply.

- Jeff

[Al Bundy's Eyes]

I think you weren't well briefed, Jeff, and you've launched in here addressing the wrong issue.

The odds of Jeff doing his homework are pretty high. If he's got a point to make, you can bet your ass he's well versed.



So you've done the 14 days in 1992 thing? What was so magical about 1992 that you would only address this one year out of the 103 which had passed since Maybrick died? Was 1992 the only year Maybrick's floorboards could have come up? Was 1992 the only year someone could have contacted a literary agent with a diary of Jack the Ripper purportedly written by James Maybrick? Your 3.8% probability confirms that you believe this to be so I know that we're done here. It is as risible as Lord Orsam's terrible 1-in-18.

Risable? Come on Ike, play nicely. Jeff's input may not align with your stance on odds, but he's been open and honest with his view, a view from a position of authority on the matter. Jeff's posts on any thread can be called long winded, perhaps boring (perhaps!), but risable? No.

When I heard we were getting an experienced university lecturer in statistics, I honestly though "Result".

Result indeed. And he's made fair observations on a subject that isn't black and white, or quite as simple as it may at first seem. He doesn't agree with your odds, or taking a simplistic view of events, but that's how he works, that's fine. Look at Jeff's contributions elsewhere, and you'll see the most unbiased and rational poster out there. He's not out to shoot down your 1 in 37,000 odds for the hell of it, he's proffering his view as a guy who works with statistics. As you yourself like to see things from all angles, take on board what he's saying, because he's right. That doesn't make him the defacto authority on the diary, he's definitely not, but his opinion (well, facts) on odds and probability carry weight. Take it onboard and work with it.

For those who are lost and confused by this, that's all I'm pointing-out at this stage and we're done here on this point.

Ike

Well thank Christ for that. So we're leaving coin tosses and apple trees alone now?

[Iconoclast]

Well thank Christ for that. So we're leaving coin tosses and apple trees alone now?

Hi Abe,

Good to hear from you old friend! Yes, we are leaving coin tossing and fruit alone for now - along with fireworks - until such time, of course, that we feel the need to correct more wrongs, whatever they are, and whenever they appear. All those dreadful, turgid, boring, drawn-out posts are gone - hopefully never to be repeated. Maybe we should even instigate a Twitteresque limited-character rule to save our dear readers from this sort of unnecessary obfuscation, bluff, and dreary bluster? I make this pledge to you all, no more lengthy posts from Ike Iconoclast, New Man of (Few) Letters!

We finally agreed (by dint of examples being given) that we had established that the principle of numerator over denominator was indeed the appropriate route to answering the simple probability question we had asked, just as I have argued all along. It then became a reasonably simple issue of what to include in the numerator (I'm still open to accepting more than one if reasonable grounds can be given for it, but it doesn't matter because you have to define the terms of your question and hope that others don't get distracted, introduce questionable variables, or even reframe the terms of the question entirely) and how to define the denominator. If you want to calculate the denominator based upon the whole of 1992 then, of course, you get a different probability to the one we were asking ourselves (because you are asking a different question of known events). You should really clarify while you're at it why you would choose to start your denominator on a completely arbitrary day (January 1, 1992), excluding every day which passed before it; and why you would end your denominator on December 31, 1992 and not go beyond it, nor indeed stop on March 9, 1992 when the event in question had occurred (the 13 subsequent occasions work was done in Maybrick's house in 1992 immediately becoming irrelevant because the other event - the 'phonecall - had been made on the same day the first occasion of work was done and therefore the question we were asking ourselves by implication stopped at March 9, 1992). So we have agreed that the principle is numerator over denominator and two posters have claimed that must be no more than 14/252 or 14/366 (1992 was a leap year) which implies that we should only consider days in the magic year of 1992. As it happens, 1992 was a very magic year for Ol' Ike himself (our beloved Queen did not concur, of course), but I assume we haven't factored that in so I am at a loss as to what else made it such an exclusively magical annum.

I think, in retrospect, we might have made more progress had we turned the question around and asked what was the probability that 1) someone would contact a literary agent with Jack's diary written by James Maybrick on the same day 2) that Maybrick's floorboards were known to be raised for the first time on the record? The former is a hugely unlikely event (surely by anyone's definition) so the chances of extraneous variables affecting its otherwise flat distribution would be slender at best and perhaps therefore something we could ignore for the purposes of our present question. If only four days had passed since Maybrick's death and those two events had definitely occurred then the probability that they happened on the same day was indeed 1-in-4. It is true that it doesn't mean that we have identified evidence of a link between the two events - 1-in-4 is no biggee - so no-one would raise their eyebrows and say "Wow, I wouldn't have expected that outcome" and therefore contact the cwoffee and donut gang to investigate further. Nor does it mean that there was a 1-in-4 chance that either or both would indeed happen at all but simply that once we knew that they had happened, then it's really easy to calculate how likely it is that they occurred on the same day. Really, really easy. And that easiness didn't magically stop on Day 4 but carried on potentially indefinitely getting less and less likely until eventually it happened (if it ever did).

As I suggested above, I think someone contacting a literary agent with Jack's diary written by James Maybrick is about as flat a distribution as one could hope to get - I don't think it increases in any meaningful way (if indeed at all) with time passing with the possible caveat - which a good detective would bear in mind - that the 1988 and 1989 centenaries of the crimes and Maybrick's death may very well have temporarily increased the likelihood of someone contacting a literary agent with Jack's diary written by James Maybrick and if we wanted we could have calculated only from that point and missed out the first century altogether but that, of course, would pose an entirely different question because our timeframes would have changed and most definitely would not justify us arbitrarily skipping the whole of the rest of 1989, and the whole of 1990 and 1991 so that we only start counting on the perfectly arbitrary January 1, 1992. In that event, however, we would still stop calculating on the day it happened and certainly not have continued onwards until we reached the equally arbitrary December 31, 1992. So we definitely would not have got 1/37,557, but equally we would definitely not have got 1/18 or 1/26.13. I'm resolutely happy that we should start when the possibilities first started, and for that we can argue the toss but I'll happily plump for May 12, 1889. Others may potentially disagree, but that does not mean that they have necessarily made a convincing argument, nor indeed that those who disagree with their calculations are thereby 'unconvincable', regardless of how unimpeachable the former’s credentials may appear to be.

But the principle has rather obviously now been agreed by all parties - it's now just about what you choose to include in your numerator and how you choose to define your denominator which is really where we were at weeks ago. And if you can make a properly convincing argument, you might be justified in expecting even the 'unconvincable' to yield a little or indeed a lot. I think that cuts both ways though and is not simply the preserve of what appears to be the majority view?

But yes (I should have added) from my perspective this debate is done (for now).

Cheers,

Ike

[Herlock Sholmes]

I haven’t a clue on this issue but I’m betting that the odds are favourable that there’ll be some serious work going on at Orsam Towers at the moment.

[Iconoclast]

I haven’t a clue on this issue but I’m betting that the odds are favourable that there’ll be some serious work going on at Orsam Towers at the moment.

OMG Herlock - can you imagine what’s coming my way?????????????????

Trust you’re well, old boy.

[Herlock Sholmes]

OMG Herlock - can you imagine what’s coming my way?????????????????

Trust you’re well, old boy.

I am Ike. Hope you’re well too.

I imagine that David’s priming the heavy artillery as we speak.

[Iconoclast]

I am Ike. Hope you’re well too.

I imagine that David’s priming the heavy artillery as we speak.

I suspect the people who run the power grid near Chigwell Manors (it's a semi but he likes to think it's grander) are currently bemused by the sudden surge in demand for the last two weeks solid, day and night.

[rjpalmer]

Hi Ike,

I think it is time to bring out the traditional black knight.

https://giphy.com/gifs/monty-python-fab-5DfGL75M9spG0