In version 0.1, I laid out the aims of this project and put together a proof-of-concept £1-per-day diet that met most of the nutritional goals set by the UK government.
In this post, I’ll be focusing more on a meal plan, introducing a greater variety of foods, and trying to improve intangible factors like culinary enjoyment of the food.
(To avoid repeating myself I won’t go over the project goals, assumptions or rules again, so if you haven’t read version 0.1, it might be worth it to get up to speed.)
In version 0.1, we ranked foods by calories-per-£1. This was an acceptable first-order approximation of food cost, but we can do better. Calories turned out to be very cheap — white rice provides over 9000 calories per £1 — so it will be helpful to introduce some new metrics that are better attuned to other, more costly to satisfy, nutritional goals. If we want to eat 5 portions of vegetables per day, we don’t much care how many calories those vegetables provide; we can always add cheap calories by eating more rice. It makes more sense to measure the cost of each portion directly.
In version 0.1, the primary source of protein was yellow split peas alone. The primary sources of fruit and vegetables (FaVs) were bananas, carrots, onions, cabbage and tomatoes. These are the two areas in most need of more variety, so we’ll be using two new metrics – price-per-FaVs-serving and price-per-100g-protein.
Fruit and Vegetables (FaVs)
Cost per kg (£s)
Cost per 80g serving (pence)
|Brussels sprouts (frozen)||1.00||8.0|
|Aubgergine (eggplant) (approx)||1.50||12.0|
As it turns out, the 5 cheapest-per-calorie FaVs are also the 5 cheapest-per-80g-serving FaVs. We can still use this table to identify the other most promising candidates to include in our diet1. I’m particularly drawn to butternut squash because it’s a popular soup ingredient. Version 0.1 of the diet lacked many practical options for a lunch dish, and soup could help fill this hole.
Price per kg (£)
Protein per 100g (g)
Price per 100g protein (£)
|Yellow split peas (dried)||1.16||24.6||0.47|
|Red kidney beans (dried)||1.56||23.6||0.66|
|Pigeon peas (gungo peas) (dried)||1.52||21.7||0.70|
|Pork liver (raw)||1.50||21.4||0.70|
|Green lentils (dried)||2.00||22.2||0.90|
|Chick peas (dried)||1.90||19.3||0.98|
|Red split lentils (dried)||2.30||22.2||1.04|
|White kidney beans (dried)||2.72||23.6||1.15|
|Soya milk (longlife, sweetened)||0.69||3.3||2.09|
|Chicken (raw, whole) (approximate)||2.48||11.2||2.22|
|White fish fillets (frozen)||4.00||12.2||3.28|
|Chicken breasts (raw, boneless, skinless)||6.89||19.8||3.49|
|Tofu (silken, firm)||2.86||7.0||4.09|
Split peas, the cheapest protein source per-calorie, are also the cheapest protein source per 100g protein. Two other popular legumes, kidney beans and pigeon peas, come in fairly close in price.
However, not only are split peas cheaper, they’re also much easier to cook because (unlike most dried legumes) they don’t need to be soaked before cooking and they cook fairly quickly. Dried kidney beans, on the other hand, need to be soaked and cooked fairly aggressively to tenderize the beans and break down the lectin phytohaemagglutinin.
My plan is to include 2 servings of kidney beans in the weekly meal plan, in which case the beans can be cooked just once per week (and reheated for the second portion).
The only animal protein that comes close to legumes in terms of price is liver. (Pork liver is included in the table, beef liver costs around the same.) However, it’s not really a practical option, because i) it’s sold only in 1kg packs and will spoil quickly, and ii) liver is such a rich source of vitamin A that it’s not recommended to eat it more than once per week in order to avoid vitamin A toxicity.
In version 0.1 of the diet, the primary sources of cheap carbohydrate were porridge oats (4960 kcal per £1) and white rice (9125 kcal per £1).
In version 0.2, I’m going to try to include brown (wholewheat) bread (3881 kcal per £1). That’s because i) it provides more convenient lunch options with some portability, and ii) it avoids the monotony of eating the same kinds of meals for lunch and dinner. Bread doesn’t have a terribly long shelf life, but it freezes very well and can defrosted (or mildly toasted) in a toaster or under a broiler.
In version 0.1, the large majority of dietary fat came from extra-virgin olive oil.
In version 0.2, I’m going to replace some of that fat with rapeseed (canola) oil because i) it’s less than 1/3 of the price by volume, and ii) it has a more neutral flavor and can be used for cooking at a higher temperature.
Olive oil was simply taking up too much of the budget. I’m not over the moon about including more canola oil, but the money saved can be used to partially fund the overall increase in food variety. Sacrifices have to be made.
I’m also trying to include some “treat” items in this version of the diet, so I’ll be attempting to include a small amount of peanut butter (3418 kcal per £1) as a snack.
Herbs and spices
Version 0.1 left around £0.50 per week for herbs, spices and other flavorings. In version 0.2, I’m going to attempt to include the cost of herbs and spices in the budget. In most cases, I anticipate the fractional cost to be very small. The hard part will be finding space to include the up-front cost.
There are so many herbs and spices, and the serving-sizes vary so much, that I don’t think it makes sense to analyze prices ahead of time. Instead, I’ll see what flavorings would naturally fit with the meal plan, then work out the costs and see if they fit into the budget.
Version 0.2 is adjusted in a few places to try to make the plan more practical and realistic.
Firstly, all nutritional measures are now adjusted to assume a 10% loss of purchased food. This is intended to provide a buffer for food wastage. This applies not just to calories, but also to protein, fat, fiber, FaVs servings etc.
Secondly, because storing produce always carries some risk of spoilage, the budget is now weighted towards storing as little fresh food as possible. In the case where it’s only marginally cheaper to buy a much larger quantity, the smaller quantity is preferred.
For example, a 4kg package of onions is priced at £0.42 per kg and a 2kg package is priced at £0.49 per kg.
In ideal conditions, onions can be stored for months without spoiling. However, in reality, there is always risk. Because the price difference is relatively minor, we now buy onions in 2kg packages.
In version 0.2, I’m sticking to the same breakfast option as in 0.1. Porridge (oatmeal), made with water. It’s not exciting but it’s cheap and healthy and, anecdotally, people are more willing to eat the same thing every day for breakfast than for any other daily meal.
However, I’m open to the addition of flavorings to improve the taste. Salt is traditional in Scotland, but I’m more inclined towards cinnamon.
Lunch was by far the biggest weakness in version 0.1, so it’s where the biggest changes are in 0.2.
The first lunch option is hummus sandwiches with grated carrot.
I’m cheating a little – real hummus is expensive (£3.33 per kg, 970 kcal per £1). This recipe calls for a garlicy split-pea spread made without any tahini. Instead, the ingredients are split peas, extra virgin olive oil, garlic, lemon juice, salt and black pepper. It’s served on brown bread with grated carrot.
The split pea “hummus” can be made a couple of times a week and kept in the fridge. The lemon juice is taken from whole lemons which can be kept in the freezer, then defrosted (either in the microwave for a few seconds or overnight) and juiced.
The second lunch option is butternut squash and swede (rutabaga) soup.
This soup combines roasted butternut squash and rutabaga with a base of onion, carrot, garlic, extra virgin olive oil, salt, black pepper, dried mixed herbs and vegetable stock. It’s served with toast made from brown bread.
The soup can be made in a big batch and then frozen in individual portions. If we have hummus sandwiches and butternut squash soup on alternating days, we can make six portions of the soup at once and not have to make it again for 2 weeks.
In version 0.1, the only snack option was a daily banana.
I have tried every which way to include a wider variety of fruit in the diet, but it’s simply too expensive.
However, this is room for some other snacks: small servings of peanut butter and (occasionally) butternut squash seeds (the seeds left over from making the soup, roasted with extra virgin olive oil and a little salt). Each week there’s also half bar of dark chocolate for a special treat.
Each dinner still needs to contain a full portion of legumes and a full portion of rice in order to meet our nutritional targets. However, now that we can use different types of legumes, there are three fairly different dinner options from different cuisines.
The first is split pea dal with rice. This budget version of the South-Asian dish is prepared with split peas, tinned tomatoes, garlic, onion, canola oil, garam masala, ginger, black pepper, salt and served with white rice.
This is the cheapest dinner and also the easiest to prepare because the split peas don’t need to be soaked before cooking. By my reckoning we need to eat dal three times per week.
The next option is vegetarian chili with rice. This is made from kidney beans, tinned tomatoes, garlic, onion, chili powder, olive oil, canola oil, cumin, black pepper, salt and served with white rice.
The kidney beans need to be soaked and cooked for quite a long time. I think the best plan is to eat it once on the night it’s made and again two nights later (i.e. having one other evening meal in between). This means we can leave the portion of cooked chili in the fridge and all we need to do for the second meal is cook some more rice.
Lastly is rice and peas, consisting of pigeon peas, white rice, onion, ginger, garlic, chili powder, olive oil, dried herbs and salt.
Again, the pigeon peas need to be soaked. We can cook and store this dish in the same manner as the chili.
The meal plan
Putting it all together, the meal plan looks like this:
|Lunch||Split-pea hummus sandwich with carrot||Butternut squash soup with toast||Split-pea hummus sandwich with carrot||Butternut squash soup with toast||Split-pea hummus sandwich with carrot||Butternut squash soup with toast||Split-pea hummus sandwich with carrot|
|Snack||Banana||Peanut butter||Banana||Butternut squash seeds||Banana||Banana||Banana|
|Dinner||Split-pea dal with rice||Vegetarian chili||Split-pea dal with rice||Vegetarian chili||Rice and peas||Split-pea dal with rice||Rice and peas|
It might not get your mouth watering, but it does have the virtue of (mostly) avoiding eating the same thing on consecutive days (with the exception of breakfast). It certainly has a lot more variety than version 0.1 of the £1-per-day-diet.
Nutrition and budget
Price per kg (£s)
Calories per £1 (kcals)
Weekly amount (g)
Weekly Calories (kcals)
Weekly Cost (£)
|Vegetable oil (rapeseed)||1.00||8840||420||3713||0.42|
|Long Grain Rice (white)||0.40||9125||1000||3650||0.40|
|Yellow split peas (dried)||1.16||2940||450||1535||0.52|
|Olive oil (extra virgin)||3.33||2655||140||1238||0.47|
|Pigeon peas (gungo peas) (dried)||1.52||2257||160||549||0.24|
|Red kidney beans (dried)||1.56||2160||160||539||0.25|
|Dark chocolate (50% cocoa mass)||3.00||1820||50||273||0.15|
|Tomatoes (chopped, tinned)||0.78||219||800||136||0.62|
|Garam masala (dried)||7.20||451||10||33||0.07|
|Extra Hot Chili Powder||6.80||415||10||28||0.07|
|Vegetable stock cubes||6.50||563||6||22||0.04|
|(Totals adjusted for 10% food loss)|
25 ingredients in all, certainly a step up from the 9 ingredients of version 0.1. Things are looking good: even after adding a 10% buffer for our targets and radically increasing the diversity of foods in the diet, we still have just under 50 pence per week left over.
Digging a little deeper into the nutritional profile, things still look promising:
Daily amount (assuming 10% loss)
|Calories||2267 kcals||2250 kcals|
|Calories from protein||10%||15%|
|Calories from fat||35%||35%|
|Calories from carbohydrate||55%||50%|
|Fruit and vegetables||6.5 portions||5 portions|
The only target we undershoot slightly is calories-from-protein, but seeing as we’re still comfortably meeting the recommended intake of 55g per day, this seems acceptable. In fact, seeing as protein is by far the most expensive macronutrient to provide, it makes sense to get away with as little as we can (while still meeting the target).
Daily fiber, while still providing an extremely robust 47.6g per day, is down from 68g per day in v0.1. That’s good news for the tolerability of this diet — it’s less likely to cause gastrointestinal discomfort.
At this point, things are looking extremely promising — but unfortunately, there’s a sting in the tail.
Fractional costs and the monthly budget
Over the course of this project I’ve been adding to an increasingly complex spreadsheet to run a lot of the numbers automatically (download links at the end of the post). The spreadsheet can now simulate the shopping budget over multiple weeks, taking into account existing stocks of the various ingredients.
A monthly budget
Assuming we shop at Asda only once per week (in order to minimize travel costs), the first month of shopping currently looks likes this:
|First 28 days||£33.90|
This is where greater food variety really takes its toll. Even if we assume our £1-per-day budget is monthly (i.e. we get £28 on day 1, and another £28 on day 29, etc.), we’re still over budget by almost an entire week in the first month.
One problem is that the minimum purchase size for flavorings — herbs, spices, etc. — is huge compared to the amounts we’re using. For fresh produce, the most we buy at once is around a 3 week supply (for onions). But for black pepper, chili powder or garam masala, we’re forced to buy a ten week supply at a time. For salt or vegetable stock cubes, we have to buy more than a 20 week supply on the very first shopping trip.
We’re also paying the price for greater variety in other areas — two types of oil, for instance.
This means the budget is very front-loaded. The first week is relatively expensive, and that expense is gradually recouped over the next few months. That’s not acceptable under the parameters we laid out at the start of the project.
In order to reach a budget that’s at all feasible, we need to make some pretty drastic changes. If we remove all of the herbs and spices (apart from salt), the peanut butter, the dark chocolate and the lemons, we reach an average weekly cost of £5.64 and a first-month cost under £28.
|First 28 days||£27.87|
Nutritionally, this diet isn’t hugely different to the unadjusted v0.2. But from a culinary or quality-of-life perspective, the difference is extreme. I have to wonder: what value does increased food variety have if all the prepared dishes taste bland and unappealing? Meals like dal, chili, soup and rice and peas rely on herbs and spices to be appetizing. Without flavorings, it might make more sense to abandon variety as a goal and concentrate on eating the same thing every day (but at least having it taste good).
Of course, we can gradually re-introduce the herbs and spices after the first month. On this adjusted budget, every week after the first month we have around £1.30 left over to put towards these purchases. But those first 4 weeks will be pretty miserable, and it will take at least a couple more weeks to assemble a reasonably stocked spice cupboard.
A bi-weekly budget
If we assume we’re on a bi-weekly budget (i.e. £14 on day 1, another £14 on day 15, etc.), things become even bleaker. In addition to the changes above, we have to forgo olive oil (for an equal amount of vegetable oil) and bananas for the first two weeks, after which we can reintroduce them as normal.
Again, things improve fairly quickly after the first month as we can gradually purchase more herbs, spices and other flavorings. But in this version, the first month seems even more miserable than the monthly-budget version.
Goals for version 0.3
Clearly, there’s lots of work left to be done. If version 0.1 was too monotonous and impractical, version 0.2 is too lavish, resulting in an unworkable budget.
There are two options to resolve the budget problems:
- 1) Reduce the overall variety in the diet somewhat, cutting down on the number of ingredients, especially different types of herbs and spices. Or,
- 2) Increase the granularity of the simulation, allowing for a smaller range of dishes in early weeks and gradually increasing the variety of dishes as budget savings accumulate.
Option 2 is the most appealing, but also the more complicated to plan for. As my spreadsheet grows ever more sprawling, complex and muddled, I become less able to stay on top of things.
I’ll have to give some serious thought to the best way to move forward.
Other goals for the future
These might not all necessarily make it into v0.3, but more goals for the project:
- Perform a more detailed micronutrient analysis (vitamins, minerals, etc.) I’m waiting until the rest of the diet is nailed down before tackling this. As the diet happens to be vegan, there’ll be some work to do to increase the amount of micronutrients like calcium.
- Test and document the recipes in more detail. These are all foods I’ve made and eaten, but not with such precise quantities of ingredients. The palatability, cooking times, portion sizes, etc. should all be tested.
Use my data
I’ve collected a fair amount of data so far, and I’d be thrilled if anyone else would like to play around with them.
Below I’ve embedded a condensed version of my work-in-progress spreadsheet. You can play with the numbers and see how different amounts of the various foods affect the budget and nutritional breakdown.
To get started, just click or tap inside any of the blue cells and change the amount of that food eaten weekly. To readers on phones or other small-format devices, apologies if the widget below doesn’t display correctly.
Please note: any changes will be lost when you close the window or refresh the page.
Download the data
You can also download a more complete set of the data. The zip file contains both the original Excel spreadsheet (with data and formulae) or a comma separated file (data only).
The formulae in the Excel spreadsheet are probably much more convoluted and less elegant than necessary. You have been warned.
The spreadsheet includes links to each Asda product and to the corresponding entry in the National Nutrient Database for Standard Reference 25 (where available).
Predictably (for me), this project has become something of an obsession and is sucking up increasing amounts of my time.
I’ve been half considering putting together a £0.99 (or $1.49) ebook at the end of this project to present everything in a cohesive, unified and easy-to-read format. If you think you might be interested, enter your email address here for updates. (No spam, no other topics, one-click unsubscribe.) If enough people sign up, I’ll probably follow through.
Lastly, the deeper I get into this project, the more it feels like my squishy human brain is an inefficient engine to solve this particular problem.
Although food and cuisine seem deeply human, constructing a nutritionally complete diet with a limited budget is little more than a (fairly simple) combinatorics problem.
As such, it seems like a problem ripe for solving with an algorithmic approach. A computer program would have the advantage of being far more versatile than the approach I’m taking. It could be adjusted fairly trivially for different budgets or different nutritional goals.
In order to take such an approach, you would need a pool of food data that included nutritional information and price information. You could then define a set of recipes, and set your program to finding the optimal series of meals that met your goals.
I’m not a programmer but I’d be very interested in this kind of project. If it’s something you’re working on or considering, and you’d like to have a chat about it, shoot me an email.
- N.B. – many of Asda’s fresh fruit and vegetables are sold per-item and no price is given by weight. Where possible, I’ve tried to make an educated guess as to the price per kg using average weights, but in many cases this is not possible and I’ve had to leave out that food from the table [↩]