Valuing a Narrowboat

[toppy007]

Dear CANAL WORLD

 

I am currently writing an open source program to predict the value of a Canal Boat.

 

The program will be trained on collected data. I am currently sorting this data into the features that influence the price

 

My main question is about what features I should Include. 

 

For example I currently:

Boat Age,

Boat Make,

Steel Thickness, 

Size,

Material, 

Engine Type,

Heating System,

Electrical System,

 

What else do you think I should consider?

 

Thanks for any info

 

 

 

[Alan de Enfield]

16 minutes ago, toppy007 said:

Dear CANAL WORLD

 

I am currently writing an open source program to predict the value of a Canal Boat.

 

The program will be trained on collected data. I am currently sorting this data into the features that influence the price

 

My main question is about what features I should Include. 

 

For example I currently:

Boat Age,

Boat Make,

Steel Thickness, 

Size,

Material, 

Engine Type,

Heating System,

Electrical System,

 

What else do you think I should consider?

 

Thanks for any info

 

 

 

 

It is 'subjective' but condition is by far the most variable that affects the price.

 

The same as with cars - You can buy a cheaper, badly maintained Rolls Royce, for less than a well maintained Ford

 

It can be £10,000 to have a boat professionally painted, It can be £7,000+ if a new engine is needed, £1000 to fit a new gearbox, £2,000 to buy and have fitted a new solid fuel fire (if there is not one already in the correct place)

 

You mention hull thickness - but (certainly for me) an overplated boat which may have a thicker steel is not worth more than an unplated, thin steel. boat. Badly applied overplating can actually devalue a boat.

 

You also need to take ito account the state of the economy and peoples disposable income, people are currently leaving / being evicted from rental properties because they cannot afford to pay £1000+ rent per month so they buy a boat creating a demand and increasing the prices.

 

Mortgages are goung up by 'hundreds of pounds per month', interest rates at the highest for many, many years. Folks will be defaulting on their mortgages and have no where to live - buy a boat and live (they think) cheaply'

 

There are a large number of older folks who can now take a lump-sum from their pension pot so they buy a boat, increasing demand and increasing the prices.

 

Since C19 more people are happy to 'stay at home' and invest the £10k they would have spent on a holiday on a boat.

 

If you can encompass all of these intagibles into a spreadsheet you should be somewhere earning £1 million per annum.

Edited November 4, 2022 by Alan de Enfield

[Mike Todd]

6 minutes ago, toppy007 said:

Dear CANAL WORLD

 

I am currently writing an open source program to predict the value of a Canal Boat.

 

The program will be trained on collected data. I am currently sorting this data into the features that influence the price

 

My main question is about what features I should Include. 

 

For example I currently:

Boat Age,

Boat Make,

Steel Thickness, 

Size,

Material, 

Engine Type,

Heating System,

Electrical System,

 

What else do you think I should consider?

 

Thanks for any info

 

 

 

Clearly the most important factor which you have omitted is the toilet system! (If you are plugged into canal social media you will know just how important that is . . .)

 

Stern Type (trad, semi trad, cruiser)

Layout (reverse, trad)

No of beds/bedrooms

No of toilets/.bathrooms - any en suite - applies only to larger boats

heating system - may be more than one eg diesel (webasto/eber) and solid fuel stove

No of batteries - 'electrical system' does not make much sense unless confined to 12v/24v

Inverter/Land line connection

Solar

 

I suggest that you look at boat brochures from some of the better used-boat brokers (Rugby boats, ABNB etc) and trawl for the listed and featured characteristics - which are numerous. My fear is that when you include all or at least most of the factors that folk take into consideration when considering a specific boat then you will have insufficient data, relative to the number of variables, to make your 'training' adequate. You may also need to take into account the difference between asking price and vale (ie what it sells for) As with houses, there can be quite a margin. Difficulty is that the price paid does not have to be registered anywhere.

 

 

[Tracy D'arth]

I would be amazed if your program could produce sensible boat valuations because there is no previous yard stick and too many variables.

 

Boats are not like cars when it comes to valuation.

 

Best of luck.

[nicknorman]

I think the nature and quality of the fit out is important. A boat from a well-reputed manufacturer sold as a sail away and fitted out badly by the owner with cut down domestic units and a lot of freestanding furniture, will be worth a lot less than the same hull fitted out with integral and bespoke furniture installed at build by the manufacturer’s professional workforce.

[David Mack]

Where are you going to get the data from? How are you going to keep it up to date as the market changes? How are you going to deal with the difference between advertised prices and actual sale prices - the relationship between them varies depending on the state of the market?

And at the end of the day, what is the point of your program? If it says a particular boat has a certain value, and the willing seller and willing buyer agree on a different figure who is right?

[Tacet]

The challenge you face in establishing an expert valuation system is the inter-action between the various factors.  As the absence of a fixed is liable to have a smaller negative effect on a short boat than the same issue on a longer boat; same with condition or steel thickness and age.

 

There have been various attempts at creating valuation models for real property - which is much more location dependent, of course.  It can work tolerably well if beacons are established- that is examples, hypothetical or otherwise, of particular types.  Right to buy council house valuations were reasonably well suited. where a standard three bed-mid terrace etc has a (subjectively established) beacon value - and you add a little or take off a bit for houses on the same estate according to a formula for differences in central heating, garaging, new windows etc.  But even then, the resulting figures were inevitably subject to some human oversight.

 

A report has been published this week on property valuations - partly concerned with regulated purpose valuations where the property figures are made known to the stock market etc.  Typically these concern large investment portfolios which lend themselves to a heavily mathematical approach.  But the report considers that human valuers will be around for a while yet.

[MtB]

2 hours ago, toppy007 said:

What else do you think I should consider?

 

1) Location. 

 

A boat on say, the Lancaster might sell for perhaps half or two thirds of what it would make if sold in central London.  

 

2) The mooring.

 

Again, a boat with a really nice residential mooring will sell for a lot more than the same boat sold with no mooring. 

 

 

[tree monkey]

What is impossible to take into account is the intangible bits, it's often mentioned when looking for a boat that people just find "the one for them" 

[harrybsmith]

I can see this being useful, rather than an absolute price you'd probably start with a "price per foot" then skew it lower or higher depending on various factors e.g. 

 

50ft boat, start at £50k (round numbers) 

Rubbish shell builder, -8k 

Nice ish engine, +2k

On the K&A, +2k 

Untidy, needing paint, -£5k 

 

And you end up with a value of £41k, which is somewhere close ish i'd imagine. A useful guide 

[Tam & Di]

It is maybe a fun exercise to occupy some dull moments, but I can't see you coming up with any useful program. There are dozens of unquantifiable factors - how about: The number of similar vessels currently available; do people prefer red boats to green ones?; does the owner have to sell in a hurry?

 

Tam

[toppy007]

Thanks for input everyone, Some interesting points.

 

2 hours ago, David Mack said:

Where are you going to get the data from

 I was going to ask this to the community! ebay auction sales value? approach Individuals? Or acknowledge price predictions based on advertised prices and not final sales? Any thoughts, what do you think.

 

 

[Alan de Enfield]

1 minute ago, toppy007 said:

Thanks for input everyone, Some interesting points.

 

 I was going to ask this to the community! ebay auction sales value? approach Individuals? Or acknowledge price predictions based on advertised prices and not final sales? Any thoughts, what do you think.

 

 

 

 

To be honest I think to are trying to find a solution to a non-problem.

 

It is a cottage industry where every narrowboat is 'hand made', made to customers specification from a choice of options, so every boat is different - they are not cars where 100,000 are all made the same.

[toppy007]

3 hours ago, Mike Todd said:

specific boat then you will have insufficient data

 

Yeah, I think with the 37 features id like to find about 400 training samples!

 

Thoughts on how to go about this would be appreciated. 

 

 

Narrowboat Instant Quote Program

 

Model Output = Price (Prediction)

Model Inputs = Boat Specifications

Training Features 

Current Market Infomation

1. Date Valued

2. Location

3. Mooring Included

Boat Overview

4. Construction Year

5. Boat Builder

6. BSC Date

7. Current Condition

   1. Sail-Away, Project

8. Boat Style

   1. Butty, Tug, Traditional Stern, Cruiser Stern, Semi-Traditional Ster, Other

9. Length

10. Hull Material

    1. Steel, Aluminum, GRP, Wood

11. Hull Condition 

    1. Pitting Evident, Overplated, Last Survey Date.

12. Hull Blacked Date

    1. Year and Type of Blacking (epoxy or bitumen)

Engine Overview

13. Engine Make

14. Engine (hp)

15. Engine Hours (Estimate)

16. Engine Cooling System

    1. Keel/Skin Tank, Raw Water, Other

Fitout Overview 

17. Most Resent Fitout Year

18. Proffectional or Owner Fitted Out

19. Layout Style

    1. Traditional, Reverse Layout 

Fresh Water System

20. Fresh Water Tank Size

21. Year Installed (Including first fit piping)

22. Hot Water System

    1. Calorifier (Shore Electric, Engine Connected, Diesel Heater Connected, Solar Dump), Instant hot water (Electric 12v, Gas), 

Galley Overview

23. Galley Amenities

    1. Sink, Fridge, Freezer

Electrical System Overview 

24. First Fit WIring Age

25. Shore Power Connector

26. Inverter/Charger

    1. Size, Make, Age

27. Solar 

    1. Size, Type, Age, Controller make 

28. House Battery Chemisty 

29. House Battery Size (Ah)

30. House Battery Age

Heads Overview

31. Heads Fit Out Year 

32. Number of Heads

33. Toilet Type

    1. Cassette, Pump-out, Compostable, Incineration 

34. Amenities

    1. Shower, Bath, Sink 

Heating Overview

35. Heating Type

    1. Multi fuel stove, Diesel heater (Electric or Gravity), Gas

36. Central Heating Connected

37. Insulation Type.

    1. Spray Foam, Fiber Glass, Celotex

[Rob-M]

Wasn't there someone on her a few years ago that built themselves a complex spreadsheet to score the boats they looked at to determine which was the best buy?

[David Mack]

2 hours ago, toppy007 said:

I was going to ask this to the community! ebay auction sales value? approach Individuals? Or acknowledge price predictions based on advertised prices and not final sales? Any thoughts, what do you think.

 

 

I wasn't thinking only of the price data. You have just listed 37 factors you are thinking of including. Where are you going to get information on all of those? Most boat ads give only partial data with no means of finding the rest unless you inspect every boat yourself. And how are you going to practically scrape the data that is published from boatyard, broker and private ads with no consistent format?

Edited November 4, 2022 by David Mack

[LadyG]

6 hours ago, toppy007 said:

Dear CANAL WORLD

 

I am currently writing an open source program to predict the value of a Canal Boat.

 

The program will be trained on collected data. I am currently sorting this data into the features that influence the price

 

My main question is about what features I should Include. 

 

For example I currently:

Boat Age,

Boat Make,

Steel Thickness, 

Size,

Material, 

Engine Type,

Heating System,

Electrical System,

 

What else do you think I should consider?

 

Thanks for any info

 

 

 

Location, N cheaper, London most expensive.

Hull builder.

Fitting out , is it owner fit out or professional, and is it cheap and cheerful or top job.

Seasonality, spring thought to be higher, winter lower.

Many boats will have two or more heating systems.

I would think maintenance would be a big factor, any rusty tatty leaky boat is going to be much cheaper. When last blacked, epoxy from new or bitumen etc etc

Paperwork, this may or may not be relevant, but could be a  positive for some people/brokers. There is no register of boats like cars, but there may be a paper trail. No paperwork is a negative.

Sale platform, private sale, or brokers

Survey available.

How are you going to find the current prices ,  offer price may not be purchase price, but in one way it must be more available 

 

 

 

 

Edited November 4, 2022 by LadyG

[Allan(nb Albert)]

I think David Mack has a point. Car value guides depend on the trade providing data including sales price, condition  and probably registration no. (as that can be used to deduce make, model, sub model and mileage).

Not sure the boating trade would be willing to provide the information you require ...

My approach would be to write software that obtains the info you need direct from the sales websites. I would be inclined to start with a very small subset -

length

year
asking price

- but also retain the complete information for later use.

The above can then be used to give a max, min and average price for a boat by year and length which would be a good starting point and useful for somebody selling.

You can then use the retained info to see what other factors have a major bearing on asking price and ask owners some simple questions to determine if the average figure should be adjusted. 

A major problem with the approach above is having the necessary skills in screen scraping, and fuzzy searching using regex. Also the large number of sites each of which will have to be analysed in different ways.  A Linux/Unix background would be more helpful that Windows.

 Hope this helps even if it suggesting a lot of work.

 

[Tony Brooks]

Personally I would turn down a boat with a Vetus engine, any other major Vetus parts based on the price of spares. I would pay a small premium for one with a Bukh engine in good condition.  Happy enough with Beta and possibly Isuzu but forget anything else, especially slow revving vintage engines. However, many pay a hefty premium for a traditional engine. I would want a basically 12V boat with only the smallest inverter, others will devalue such a boat and want a hefty inverter for loads of 240V equipment. Nothing would induce me to buy a traditional stern abut I would pay a small premium for a cruiser. That, again, is personal preference. It is the same with toilets, I want a dump through but might accept a cassette type. Nothing would induce me to buy a boat with an electric toilet or a "composting" one.

 

A boat I would pay a premium on others would want a discount. I see personal preference as being the most difficult thing to allow for.

 

[matty40s]

1 hour ago, Tony Brooks said:

Personally I would turn down a boat with a Vetus engine,..

 

 

...because its YELLOW !

[Tony Brooks]

Just now, matty40s said:

 

 

...because its YELLOW !

You know well enough it is because of the cost of Vetus parts and spares oh, and I don't like one of the gearbox options.

[roland elsdon]

Missed out the wow factor in the equation. The one thing that is unquantifiable and impossible to value.

Both my last two boats very different both with some issues ‘wow ‘made the purchase.

Neither were a logical decision, would I do it again oh yes.

 

[Murflynn]

a bit like asking how long is a piece of string I'm afraid.   IMHO there is only one factor that fixes the price of a boat - how much is the prospective purchaser willing to offer?

[MtB]

17 minutes ago, Murflynn said:

a bit like asking how long is a piece of string I'm afraid.   IMHO there is only one factor that fixes the price of a boat - how much is the prospective purchaser willing to offer?

 

There MUST be more to it than that.

 

Firms insuring narrowboats fully comp must have some sort of broadly realistic valuation methodology to avoid being defrauded by mates selling each other old rust buckets for vast sums then sinking/burning them and claiming the purchase price. 

 

 

 

 

[Tacet]

1 hour ago, Murflynn said:

a bit like asking how long is a piece of string I'm afraid.   IMHO there is only one factor that fixes the price of a boat - how much is the prospective purchaser willing to offer?

A sale determines the price but not the value.