England Under Steam!

I've been worrying that my game might be too big for my goals of a game that I can play with my gaming group in an evening. With 11 companies, it would be towards the larger side of 18xx games, although not one of the few monster games such as 18C2C or 1825 with all units in play. Even if I can streamline the system a bit (and it remains to be seen whether I can achieve that), the game might take rather longer than I'd like. Bear in mind that my companions and I aren't the fastest players.

So in a fit of either madness or inspiration, I tried hacking off some of the map, on the assumption that the corresponding companies will also be removed. Scotland has been replaced by off-board areas, as has more of East Anglia. These changes remove the NBR, Caledonian and GER. The LSWR territory has gone too. The result is noticeably smaller, with just 52 playable hexes and 7 companies.

I'm quite taken with it. It might make more sense to keep 8 companies instead of 7, so that in a 4-player game, everyone can have the chance to run 2, but I think it'll be worth experimenting with the smaller version.

More map analysis

In my last post, I mentioned that my draft map has a lot of "dot towns". I wondered how this, and other aspects of the map, compared with existing games. So one evening I wrote a table that compared certain values for 1825, 1830, 1856, 1861, 1812, 1889, 18EU and Steam Over Holland.

Starting with towns and cities, I found that the ratio of town hexes to total hexes varies from 10% (1861) to 23% (18EU). My draft map was on 38%; as I thought, it was clearly higher. For cities, the ratio in existing games varies from 15% (1861 again) to 33% (1825 Unit 2). Here my draft map was closer, with 33% equalling the highest existing game. Although it has more "big cities", which is a category that (depending on the game) includes "OO" cities named cities with higher values than the norm. Conversely, the ratio of plain hexes was far lower than the norm at 30%, compared with the 50%-75% found in existing games.

I also compared the number of town and city hexes per company in the game. Here my draft map is well within the norm, reflecting the fact that the game has more companies than many others.

I also looked at terrain. Britain Under Steam doesn't have the usual mountain and river hexes; it's more like Steam Over Holland in that it has mountain hexsides. So for both of these games, I counted the number of hexes adjacent to a mountain or river hexside and used that value. The result is 29% of the total hexes, which is squarely within the "expected" range.

Finally, I counted the number of "plain" hexes, by which I mean playable hexes that don't contain any terrain, towns or cities. At 16%, the score for my draft map is very low, matched only by the terrain-heavy (and smaller) maps for 1889 and 1825 Unit 2.

Based on this information, I plan to review each town on the map to see whether it would work better as a plain hex. Some of them represent genuinely important towns that it would not make sense to remove. Others may need to remain towns so that the tile upgrade path is appropriate for that location. But there should be some that I can remove.

To complicate matters, I may also be downgrading some of the cities, which will of course increase the number of towns again! I may well end up with more populated hexes than the average game. I don't see that as necessarily a problem, provided that they are all justified (and that the game plays well, of course!).

Analysing the map

To refine the map, I've analysed how many hexes it has of each type. I have drawn up a table that shows how many hexes have dot towns with 6 playable hexsides, or cities with 5 playable hexsides, and so forth. For hexes with fewer than five hexsides, I've noted the different configurations (e.g. K, X or Y shapes). This shows me that I have a few hexes that don't fit the more general pattern, which suggests that I might change those hexes. For example, one city hex has just 2 playable hexsides that are adjacent and I don't want to provide a tile with that shape, so I might make it a pre-printed hex. In other cases I can add track in otherwise unplayable hexes, in order to increase the playable hexsides of adjacent hexes. None of this is rocket science but the table is very useful in suggesting which hexes might need work.

I've also realised that I have large numbers of dot towns. In some cases, this could lead to boring track lays. For example, where a row of coastal hexes each contains a dot town, they would all upgrade to green "K" tiles with only one permitted orientation. I am considering adding some extra track to link such hexes along the coast, so that the number of playable hexsides is increased and so more options are available when laying tiles.

I'm pleased to be at this stage; I think I am close to having a playable map. I'm sure it will continue to change but it's taken a long time just to get this far.

The joy of skew

Having made to decision to excise the south-east, I've gone back to Corel Draw and fiddled about with the map some more. I've found that by skewing the outline of the coast and slightly changing the rotation of the hex grid, I can align a row of hexes neatly along the eastern coast of Britain, from Ipswich to Dundee. Then at right angles, the bottom of the map will run from Southampton neatly across to Southend. With a little tinkering of the half hexes on the southern edge, I have a reasonably neat looking map - better than I had before. I'll have to fit in off-board areas for Norwich and Aberdeen, but they should be possible.

On a separate note, fitting the hexes around the Severn estuary has proven a challenge. I used to have the line running from London through Bristol and into Cardiff, but historically the Severn tunnel wasn't build until quite late in the game. So instead I'm forcing track to be laid via Gloucester, and this is where the hex grid has made things tricky. I think I've handled it, by adding some mountain hexsides, including one that isn't really a mountain but blocks the geographically non-existent direct route from Worcester to Cardiff, forcing track to run via Hereford or Newport. I may have to adjust this against the tile mix, but I think it will work out.

Scotland under Steam?

In my reading about the history of railways in Great Britain, I’ve taken a particular interest in Scotland. This is partly because I’ve lived in Edinburgh for most of the last 25 years, and partly because it is one area where the design of 1825 takes major comprises with history in order to make a playable game. I’ve already made an 1825 variant for the Scottish board and recently I had some interesting e-mail discussions with Stuart Dagger on the subject. Then Geoff C designed an 1825 extension board for the north of Scotland and uploaded it to the 18xx yahoo group. So clearly there is some interest here.

From what I’ve read, it looks like Scotland should make a good 18xx game. Although two companies came to dominate the country, several reasonably-sized companies were formed along the way:

• Edinburgh and Glasgow Railway: 1838-1865
• Glasgow and South Western Railway: 1838 onwards (starting as the Glasgow, Paisley, Kilmarnock and Ayr Railway)
• North-British Railway: 1844 onwards (originally running from Edinburgh to Berwick)
• Caledonian Railway: 1844 onwards (running from Glasgow to Carlisle), or possibly 1838 onwards, starting as the Glasgow, Paisley & Greenock Railway.
• Scottish Central Railway: 1845-1865
• Scottish North Eastern Railway: 1845-1866 (The 1845 dates refers to the formation of two constituent companies, the Scottish Midland Junction Railway and the Aberdeen Railway)
• Great North of Scotland Railway: 1846 onwards
• Highland Railway: 1854 onwards, starting as the Inverness and Nairn Railway.

Company mergers, such as the Edinburgh and Glasgow merging into the North British, or the Caledonian absorbing the Scottish Central and Scottish North Eastern, could either be represented as an explicit merger as in 1861 or 1812, or could simply be modelled as the same player running the two companies to mutual advantage.

Cutting the regions

When I first planned this game, one criterion was that is should cover the whole of the UK, excepting only those extremities where hardly any lines were built. So I was happy to have off-board areas for Cornwall, West Wales and North Scotland. I still looked at including regional companies such as the Cambrian, Highland and Great North of Scotland. It soon became apparent that those regional lines were just too insignificant on the scale of this game and I dropped them.

Now I'm going a step further. I intend to drop south-east England below London and with it the LBSC, the South-East Railway and the London, Chatham and Dover Railway. The map just doesn't have the space for these to be interesting companies to run. I had thought of including them as simple investment options, but the game has plenty of companies already and can do without them.

In Scotland, I shall also drop the GSWR, because its network was similarly constrained. I might also drop the GER in East Anglia, because that region also doesn't allow much development. On the other hand, historically the GER did at one stage compete with the GNR for a route north. The GNR won, but it might be nice to leave the possibility of an alternative outcome in the game.

This leaves me with the following companies:

• LNWR (London and North Western Railway)
• GWR (Great Western Railway)
• LSWR (London and South Western Railway)
• MR (Midland Railway)
• GNR (Great Northern Railway)
• GER (Great Eastern Railway)
• LYR (Lancashire and Yorkshire Railway)
• MSLR (Manchester, Sheffield and Lincolnshire Railway, later renamed the Great Central Railway)
• NER (North-Eastern Railway)
• CR (Caledonian Railway)
• NBR (North British Railway)

Eleven is still a large number of companies, but at least it offers the possibility of a manageable game.

One way I might alter this is to vary the starting companies, e.g. by picking a set of starting companies from a larger initial pool. In this case, I might also look at some alternatives, such as the West Midlands Railway or the Scottish North Eastern Railway, which historically were quickly absorbed into one of the above companies. For now, I'm going to start with the above list and see how I get on.

Railway Towns

In 1825, several major companies have their home bases in low-value cities. The LNWR begins in Wolverton and Crewe, while the GWR begins in Swindon. In Unit 2, the GCR begins in Barnsley and the GNR begins in Doncaster. All of these cities have low, fixed values and fixed track layouts.

I was wondering whether this could be extended as a general principle. Perhaps every company could start in a small town. One of these “railway towns” could use the ordinary small town tiles but with the addition of a company marker, which would stop other companies from running routes through that hex. This would make the company “own” that piece of line. As an extension, we could let companies start in any dot town; the presence of the marker indicating that it has become a “railway town”. This would give more variation to the start of the game, rather than requiring companies to start in specific hexes.

Finding start hexes for some companies could be a challenge. In the south, the LBSC might have to start at Redhill, the SECR at Ashford, the LSWR at Woking and the GER at Colchester. In Scotland, the GSWR could start at Ayr, but I’m stretching to find start towns for the Caledonian and the NBR. But it might be doable.

I’ve been moving away from this idea recently. It would be yet another change from standard 18xx and I’m not convinced that it gives the best results for the LNWR, let alone the other companies. So for now, this is an idea that I’ll put on the back burner.

Time line

I've been a bit quiet lately. This is partly because I did some more work on my 1825 web site, played another test of Ian D Wilson's 1812 prototype, and have had a few other things on the go as well. I have still found some time for Britain Under Steam. Some of this has been spent on research; I'm reading Christian Wolmar's book Fire and Steam, and I've found a copy of Snell's Britain's Railways Under Steam to read after that. It's also worth mentioning that Wikipedia has a large amount of information about the history of early British railway companies.

I've been putting together a timeline that shows key events such as the formations of important companies, completion of certain railway lines, company mergers and so forth. I was helped immensely by Lou Jerkich's Historical Order variant for 1825, in which he collected much of this information already. To start my timeline, I took the background information from Lou's variant, put all the events into historical order, and then added other bits and pieces as I found them.

This is making a number of small effects on the game design. I've rethought the way I plan to start companies and I will come back to this soon, casting my eyes over the timeline as a whole.
Every so often I note something that leads to a tweak on the board; no major redesigns, just a tidying up or an improvement in one area or another. I've had another thought about how to represent difficult terrain, too. So all in all, this background reading has been worthwhile.

Buying trains

I reckon that newcomers to 18xx find it odd that companies buy trains at the end of their turn, rather than the beginning. In particular, this means that on their first turn, companies can't run a train for income. Now, this may be historically apt, as it is certainly the case that companies have to build track before they can run any trains, but as a game it seems odd.

There are, of course, many reasons why the games are designed this way. For one thing, if a company could buy a train from another company that has just used it, the same train could be run several times during the same operating round (by different companies each time). You would need some sort of mechanism to prevent this. For another thing, a company that just had all its trains scrapped at the beginning of a phase could just buy a new train and run that, instead of losing income and share price from having no trains for a round. This would be especially true is credit is easy to come by, in the form of loans or issuing further shares.

I think I've come up with a mechanism that may square this circle. That is to let companies buy a train when they are floated, as part of the actions taken during company formation. In fact, the rules could even require a company to purchase a train at this point. This would have a couple of effects. First, it may restrict the valid share price for starting the company (at least later in the game) - as is the case with 1825's minor companies. Second, it may dampen the effect of asset stripping, in that not all the company cash will be available to transfer to other companies. (Of course, other companies could just buy the new train, so the dampening is only mild).

I'll give this a try and see how well it works in practice.

Easy Credit

David Hecht has just posted an excellent message on the 18xx mailing list explaining how you can really push the rate of train purchases in Steam Over Holland, and noted how this is an example of how credit in various forms can be used to accelerate the flow of the game.

I think I need to check whether my design makes it too easy for players to loot companies. Currently I'm intending to allow companies to be floated once 3 shares have been bought (out of 5). This would make 5 shares worth of cash available to the company. This company could buy trains from another to transfer the cash, or possibly just buy new trains in order to start a new phase and sell the new trains to another company. The player could then sell the company, getting some cash back from the shares.

Let's assume that the shares are valued at x. The player has to pay 3x in order to float the company. The company gets 5x. When the player sells the shares, he gets 1x back (because players do not get any money from selling the director's certificate). If the company has no trains, the player would only receive 0.5x. So for an outlay of 2x or 2.5x from the player, the company gets 5x.

This is comparable to 1830 et al - actually it's less return than in 1830, which gives 10x to the company for 2x of player money (or even no loss at all, if the player can dump the company on someone else). The difference is that it may be easier to buy the second company in Britain Under Steam.

I don't see a way of avoiding this, unless I turn to strict incremental capitalisation, without the option of selling shares from the company to the bank pool. Anyway, this sort of thing makes 1825, 1830 et al rather successful games, so my hope is that is will do here too.

I will note one balancing element about the Steam Over Holland approach. In order for the company to get all its cash, it has to sell shares to the bank pool. If it sells more than one per operating round, its share price drops. So directors have to trade off the rate at which they get cash against the value of their shares. This is quite appealing and I may decide to go down this route after all, although as I previously noted, I don't like the idea of paying dividends to the company for unsold shares in its treasury.

Simplified train shuffling?

I'm wondering whether it's worth adding a simpler mechanism to transfer money between companies. The classic mechanism is for one company to buy a train from the other, at an arbitrary rate. This can be used to get a train where it's wanted or to get a large amount of money into the second company so that it can buy a new train. Managing this shuffling of trains can get complicated.

A simpler mechanism would simply be to let any company aid another when purchasing a new train. There would be no need to work out which existing train to shuffle between the two; one could simply add its money to the other's.

The downsides of this? Well, one is that neither company would suffer any disadvantage. In the existing mechanism, the transfer of the train means that the second company must lose some income (and possibly share price too) before it gets to buy the new train. So you have to trade off the benefit against the cost. With the simpler approach, there would be no loss of income. So the question is whether this level of decision making adds to the game or distracts attention from the larger issues of managing share portfolios and train runs.

In fact, that is perhaps the second downside. If this becomes too easy, will it remove an element of fun from the game? I'm not convinced; given muy target audience, I think this may be worth a try. I can always revert to the tried and tested system if the newfangled approach doesn't work.

Another reason for using 5-share companies

Way back when I started this blog, I noted that I wanted to simplify the core system of paying dividends. My idea was to pay out only multiples of £10. This would require a company to have an income of £100 before dividends would be paid, and I noted that this might be tricky at the start of a game.

If you throw 5-share companies into this mix, the problem becomes simpler. A 5-share company need only earn £50 before it pays £10/share. This should be easy.

Having two separate strands come together like this gives me a good feeling. It's one way in which the different elements of the design are beginning to gel.

5-share companies

So, if I don’t want incremental capitalisation, and would like companies to start with just a small number of shares in play, but don’t want to make asset stripping too easy, what mechanisms are still available?

One approach that might work is if companies start with just 5 shares apiece, each share representing 20% of the company. This is a mechanism that David Hecht has used in several of his games. The director’s certificate will be 40% of the company and the company will begin with half its full capital.

The distribution of shares is the same as the minor companies in 1825, but these 5-share companies are allowed to morph into 10-share companies at a later stage of the game. This “growing up” gives them an extra tranche of capital and each share becomes worth 10% of the company, therefore giving less income.. The conditions and mechanisms for this will need to be developed. In 18Ardennes, the only one of David Hecht’s games that I’ve played, the director can choose to convert the company during its turn. Some of David’s games require a company to have reached a certain destination city before they can convert.

A disadvantage of this mechanism is that it can reduce the need to withhold dividends, thus making things easier for the director. An advantage is that it could encourage more diverse shareholdings, because the director need only buy 3 shares to start the company, with another 2 available immediately and 5 more appearing if the company converts to 10 shares.

On asset stripping

I’ve mentioned that I don’t plan to use incremental capitalisation in Britain Under Steam. I’ve also suggested that companies might start after a small number of shares have been purchased, instead of the more traditional 60%. There is a major problem with this combination. If a player can start a fully-capitalised second company by purchasing just 20% of the shares, he or she could then strip the assets from this company and then dump it. Even if the shares were worthless, this would give their first company 100% of the capital at a mere 20% of the cost (which would be incurred by the player).

This is far too easy a gain. It’s not that I want to forbid asset stripping as a tactic; I just want it to have a balance between pros and cons. I conclude that the ability to start companies by buying just 20% of the shares only really works with some form of partial capitalisation.

If the director receives the current price for the shares, then the loss to the player is even less. In my post about Investors vs. Entrepreneurs, I already mentioned the rule that shares in a company without a train will be valued at half-price. I am considering adding that the director’s certificate in a trainless company will be worthless. This is an attempt to make asset stripping more costly.

This will require an anti-dumping rule. If Players A and B have 5 and 4 shares in a company respectively, and Player A sells 2 shares, then the directorship would normally pass to Player B. The anti-dumping rule will allow Player B to immediately sell the minimum number of shares required to avoid inheriting a worthless director’s certificate.

Even with these rules, floating a fully-capitalised company on the sale of 20% of its shares won't work for me.

Starting the game

The question that is exercising me at the moment is how the game should start. I am caught between two competing desires. On the one hand, I want players to be able to invest in a range of companies from the start of the game, as a viable strategy. On the other hand, I want a range of companies to be available, as opposed to imposing a fixed order for company purchase. There is a definite tension here; if each player can start a company, why would they invest in another company instead?

One twist that I’m considering is to charge a small fee for starting a company. When a player starts a company, they would have to pay this fee to the bank (representing administration charges, or “gifts” to members of parliament to approve the formation of the new company). In game terms, this would be a cost for the advantage of holding the president’s certificate, which gives double the number of shares for one slot against the certificate limit. So some players might prefer to invest in an existing company, rather than pay the fee to create one of their own.

As a detail, one player might have the special ability to create a company for nothing, either as a special power or if play reaches the last player with none of the others having started a company. This would ensure that at least one player would start a company.

Returning to the question of how the game should start, one possible approach would be to have a small number of companies available to begin with, where that number is less than the number of players. The available companies could be dealt randomly at the start of the game (thus providing variety across games). Or the players could choose which companies to start but the total number could be limited.

Another option would be to limit the number of shares that players can buy in each company. In fact, many 18xx games limit this to 60%. So if players each start a company and have money left over, then they can only invest in other player’s companies. This does potentially leave each player owning 60% of a single company, which would be a less diverse portfolio than found in 1825, for example, but it does encourage some degree of diversity.

A variation on this approach would be to set a lower maximum. At the start, it could be as low as 20%, which would both encourage cross-investment and allow scope for buying more shares as the game progresses. Perhaps the limit could increase as the game progresses through different phases. For a while I was quite taken by this idea, which could be an interesting twist on the 18XX theme – but it has one big disadvantage, which I’ll explain in my next post.

On incremental capitalisation

In some 18xx games, the money (capital) available to companies is received as players purchase each share. I.e. when the players buy the shares, the money is placed in the company instead of being paid to the bank. This is called incremental capitalisation. Usually, companies may start operations with only a few shares sold, instead of the 60% required by 1825, 1830, et al.

This seems quite natural. It does reflect, to a certain extent, the way shares work in the real world. I’m happy to play these games. On the other hand, I do think this mechanism has some disadvantages.

Incremental capitalisation certainly encourages a player-as-entrepreneur flavour rather than a player-as-investor approach, because each player is strongly encouraged to invest in their own company in order to give it more capital and therefore improve the performance of their existing shares. (This is accentuated, of course, if they are not allowed to sell the president’s certificate).

In game terms, one disadvantage is that it establishes a direct feedback between how well a player is doing and how well his or her company is doing. With full capitalisation, the company starts with all its capital available and so how well the company performs is independent of any further investment. With either approach, if the company does poorly, the return to the player will be lower, but with incremental capitalisation the reverse is also true. This can leave players stuck in a losing position.

Another disadvantage is that if a company is popular, it will do better as a result of more of its shares being bought. This can introduce a “kingmaker” element to the game, where a player in last place can directly affect the performance of the leading players. It can even lead to collusion between players, e.g. if two players agree to buy each others’ shares in preference to those of other opponents.

In most of these games, when a company pays dividends, the company itself receives income from any unsold shares. This gives a stream of income to the company. Depending on the exact flow of the game, it can remove a key element of the full capitalisation approach, namely deciding when to withhold dividends. This element is one I want to preserve in my game, because it can significantly influence decisions of when to drop an investment in a given company.

So, other things being equal, I will prefer not to adopt incremental capitalisation for Britain Under Steam. Even if I do end up going with this approach, I would still avoid paying dividends to unsold shares.

Investors or Entrepreneurs?

One of the aspects of play I really like about 18xx games is that players do not directly control companies. They invest in companies and can, to an extent, switch their investments into companies that pay better dividends.

At least, that's how it was in 1829. 1830 introduced a style of play that made investment much riskier. In particular, it banned the sale of the president's share certificate. This meant that the president of a company could asset strip it and dump the shell onto any player who owned more than a single share. The hapless recipient would be unable to sell the shell and would have to buy a train from his own pocket, often leading to bankruptcy. This was, as the box announced, a game for robber barons.

Another trend is for companies to receive their capital as shares are purchased, rather than in a lump when the company first starts. This can lead to players focussing on their own companies rather than investing elsewhere, as they need the investment in order to better their companies. Again, usually players are forbidden from selling the president's certificate.

More recently, games such as 1861 begin with each player owning several minor companies outright. There is no cross-investment until much later in the game, after the minor companies have merged to form public companies.

I like 1830, 1861 et al., but I want to see more games that focus on the player as an investor rather than as an entrepreneur. So for Britain Under Steam, I intend to encourage this style of play if I can. I will certainly allow players to sell the president's certificate, as in 1825. The idea is that they won't be trapped if they invest in a company run by another player.

One consequence is that, if there is no penalty for leaving a company trainless, players will be more tempted to asset strip. This may or may not be desirable, depending on the rest of the game. However, I'm considering halving the value when selling shares in a trainless company; an idea suggested by Mike Hutton (and used, I believe, in 1860). I will need to experiment to see if this works.

A bigger question is how to handle the initial sale of companies, but that topic deserves a post of its own.

Varying the setup

Most existing 18xx games start with a fixed setup – a given number of companies, each starting in a fixed location. Although they do have a variety of opening moves available, depending on what the other players do, they can become susceptible to analysis and even being “solved”.

There are some minor exceptions. In Steam Over Holland, for example, each game starts with only some of the available private companies. (1825 is worth a mention because all the extension kits add a variety of options).

I’m wondering if I can add more variety to the setup, to prevent formulaic play. Some ideas I’ve had:

• Randomly deal special powers to each company at the start (rather than assign them to private companies and let players choose which company they end up with)
• Add special powers for players, such as the ability (once per game) to seize the priority, or start a new company at a discount, or to sell shares without affecting the stock price.
  
• Have more possible start hexes than companies, so that companies have a choice of where to start
• Randomly assign start hexes to companies in each game
• Make companies available in tiers, so that some are available at the start of the game, some later on, and so forth. Randomly deal companies to each tier.

It’s worth noting that some of these ideas will produce unbalanced companies. This reflects my preference for players-as-investors over players-as-entrepeneurs, which I must blog about...

How many companies?

If there are any experienced 18xx designers reading my blog, I expect this post to cause some amusement at my expense, or at least to raise a wry smile. In all my thoughts on the design so far, I seem to have overlooked possibly the most basic question of all. I have ideas about how companies should be run, on players-as-investors vs. players-as-entrepeneurs, on selling the directors share and on maximum share limits. I have notions of different companies requiring different types of trains and being useful in different ways. I have some thoughts about how to vary when different companies become available and what special powers they might have. But I haven't directly addressed the issue of how many companies there should be.

Obviously, I had specific companies in mind while I was concentrating on the first draft of the map, but it was only when I printed it off and started playing with tiles that I realised the need to address how companies grow together, rather than looking at each one individually. I now suspect that this is the most fundamental design decision, affecting the length of the game more than other aspects such as the size of the map. What can I say, but "oops"?

Looking at the other 18xx games I own, 1830 is perhaps the most successful 18xx game ever and works absolutely fine with eight companies. It can still be a long-ish game. 1856 has eleven companies and is definitely longer. Steam Over Holland is quicker, with seven. 1825 can be played with several different combinations: with six companies the game is fairly short but as modules are combined, the number of companies can rise to fourteen (plus minor companies), the number of players also increases and the game can take at least a day.

For a game based on the whole of the UK, there are at least twelve obvious companies to include. They don't all have to come into play in any one game, but even if some are overlooked, this number of companies could lead to a longer game than I had in mind.

I'll need to think about this. One option would be to copy 1861 and have minor companies that merge to form larger ones, but that wasn't the style of game I had in mind. Another option would be to make the game modular, in the style of 1825, but that would bring this game directly into conflict with 1825 and I'm not convinced the world needs two modular 18xx games covering the UK. I want something different. When my thoughts crystallise, I'll note them on this blog.

London revisited - or not

In Britain Under Steam, many companies will try to build routes to London. This could become a source of contention. A company with an existing route may upgrade one of the hexes adjacent to London purely to prevent a rival from gaining a route through. In many games I enjoy such blocking play, but in this particular case I want to allow many routes through.

So I'm considering giving each company a special token, which they trade in whenever they lay or upgrade a tile adjacent to London. This would make it harder to block a rival (although not necessarily impossible). An obvious extension would be to let one private company have a token too, so that the owning public company gets a second London build.

Some companies won't run to London at all, so I could introduce other hexes where this limit applies. The hex between Edinburgh and Glasgow might be worth considering.

An alternative approach would simply be to impose a financial cost for building in these hexes. Then any company may build extra routes or lay blocking tiles but only if they deem it worth the expenditure. Currently I prefer the token option, but I'll see what happens in play.

What a difference a map makes

It's been a while coming, but I now have a printed map to experiment with. I had spent quite some time playing with different sizes and alignments of hex grids, and more time yet learning how to use new drawing software. That work wasn't wasted but it's a qualitative difference having a physical map that I can put tiles on and see how the game might work (or not!) in practice.

My first thought on assembling the map was to wonder whether I should excise East Anglia and possible the westernmost row of hexes in Wales and South-west England. This was nothing to do with routes and railways, but simply a response to the physical shape of the map on the table. I'll think about this - I may add some game tracks to fill up the gaps instead.

A first attempt at tile laying went OK. I have some new tiles for OO and OOO cities, which need tweaking, but so far I'm happy with the basic ideas. Ditto with much of the basic map layout. I now need to think laterally - what track could players lay instead of the historical routes, does this make geographical sense and are there any lays that are obviously unbalanced?

One point I hadn't noticed was that the GWR/LSWR might run out of places to lay track. The current map has relatively few cities in Southern England (though potentially plenty of towns). So once the obvious long routes are laid they are left with little space to develop. It might be worth reinstating the South-west peninsula (which is currently represented by a red off-board connection at Exeter). The track laying beyond Exeter is fairly uninteresting but it would give these companies somewhere else to expand. (Alternatively I could maybe make more use of the smaller towns).

I foresee plenty of work trying the different options.

Types of city, types of train

For a while now, I've had several ideas for types of train in the end game. What I hadn't realised before was the crucial effect of design decisions about trains in the early game. I need to decide whether a simple '2' train can stop at dot towns (as in most 18xx games), can skip dot towns (as in 1861) or even take income from all dot towns on its route. And this affects the map: which towns appear as dot towns and which need to be large towns so that companies can start in them?

In the UK, companies quickly ran long routes. E.g. the LNWR started by running London to Manchester via Birmingham, while the GWR ran London to Bristol via Reading. To replicate this and give some feel for authenticity, a company with '2' trains has to start somewhere in the middle of these routes. In 1829 and 1825, Francis Tresham handled this by drawing hexes such as Wolverton, Swindon and Doncaster with fixed track and permanently low income. The starting companies run can '2' trains from these to connect up the longer route. When the '2' trains are replaced, a larger train can run the whole route on its own.

I want to avoid fixing the track in such hexes, so that players have more options for play. But neither do I want towns such as these to compete with the larger cities as sources of income. So I'm thinking of introducing a class of city that can accept markers and be promoted to increase route options but which has permanently low income. (Actually. I might give them a choice of either adding more routes or growing in income, but that will depend on what works in practice).

The UK is also densely populated. Almost every hex should contain at least a dot town. This ubiquity makes me wonder whether it's worth having the dot towns in the game at all. I certainly don't want the annoyance of a "double-heading" rule to get past the dot towns in the early game. I could just allow trains to skip dot towns, but this makes them almost pointless (apart from controlling how the track may develop).

I have the notion of allowing T trains to run as suburban trains in the late game, alongside E trains running on the same track. (This idea is inspired by an 1825 variant by Lou Jerkich). The suburban trains would only count income from dot towns, while the express trains would have to skip them. But perhaps it would be simpler to remove the dot towns completely and just give the suburban trains an income per hex on their route?

Historical rail maps

I mentioned on the 18xx list that I would like to find some historical railway maps of the UK, and the good people there promptly recommended some. Here's a summary for future reference.

What would be even more useful would be a historical atlas approach, showing which lines were built first and how the network grew.

David Hecht: There's a book of (very detailed) maps from Ian Allan called "Pre-grouping Maps and Gazetteer", though it is currently out of print. However, there are 23 copies listed from UK booksellers on Abebooks, starting at GBP 4.00 plus shipping.. Recommended! :-) [I've just ordered a copy of this.]

Stuart Dagger: Some years ago the West Country firm of David & Charles reprinted some editions of Bradshaw's and they have maps in the back. I have a copy of the reprint of one of the 1922 ones, which must have been one of the last before the companies we all know from 1829 and 1825 were pushed into the larger groupings of the LMS etc.

It might also be worth checking out the Ordnance Survey publications, as they also do reprints of 19th century maps.

After sending the first reply I remembered another source. Try www.OldHouseBooks.co.uk. They have a London Railways map for 1897 and a Bradshaw's 1907 map for Great Britain and Ireland.

Nick Wedd: The maps at http://www.maproom.org/00/35/index.php show UK railways as they were in 1914, though not in detail, and not which companies owned them. It is a geological atlas, but the author was keen on railways: "Our lines of railway have largely aided the progress of Geology not only by opening up a ready means of communication with all parts of the country, but by furnishing many instructive sections of the strata."

The map at http://www.systemed.plus.com/New_Adlestrop_Railway_Atlas.pdf is also interesting. [This one I knew of. It's useful, but incomplete].

Roger Barnes: The Museum of London bookshop is good for old maps of London or Motor
Books in Covent Garden for railway books.

London

London dominates the English railway network. The main lines lead to London, which is the centre of finance and government. They do not run through it; when a train reaches the capital, its journey is complete. So London is both the dominant city and a block to routes across it.

There are at least 9 main line stations in London, including Liverpool Street, St Pancras, Kings Cross, Euston, Paddington, Victoria, Waterloo, Charing Cross and Marylebone. Most of these were run by separate companies. It is not practical to represent all nine stations in a single hex, so for game purposes some of these routes will branch in the adjacent hexes.

If the London tile had six station spaces, as in 1829 or 1825, this could cause an anomaly. If a company were to place a marker in London, it could use this to trace a route out vai another company's track, as if (say) Euston and Marylebone were not distinct termini. So a company connecting to London could gain access to a whole new rail network, possibly in a different area of the country.

Therefore I intend to make London a terminus only. Companies will not put markers in London, only connect to it. The simplest way to display this will be to use the same markings as for off-board areas.

Note that companies will be able to share a route into London, as happened historically. This will be controlled by the usual play of station markers in the cities that connect to the capital.

Update

I've had a couple of weeks hiatus, including a week away and another that was completely taken up with other commitments. I did manage to scribble some notes (while on a train, appropriately enough) that I will write up soon.

In the meantime, I'm continuing to play with maps and hex grids. I downloaded the UK rail map and imported it into Corel Draw so that it overlays the map I was working on. After printing the map and hex grid (just on A4 paper), I'm drwing on 18xx tiles in pencil to see what works.

There are several hexes in the Midlands and North that have connections to all their neighbours. For these, I'm toying with tiles such as the BGM tiles from 1829/1825, so that the routes through the tiles may be more restricted than a simple 6-way station. The intention is to give more strategy to the tile laying in these areas (and to represent the different historical routes more explicitly).

Further south, the routes into London tend to follow patterns that correspond to green plain track tiles. However, there are many small towns here that ought to be represented on the map. This needs more thought and experimentation.

How to comment on this blog

Some readers have encountered problems posting comments to this blog. At first this was because I hadn't changed one of the default settings - by default, the system only lets you comment if you have an online ID of some kind, but I've fixed this now.

To leave a comment, first click on the "comments" link at the bottom of the article. Under the "Post a comment" text entry box, you'll see a drop down list labelled "Comment as". Click on the down arrow and you'll see a set of options. If you have an online account from Google, LiveJournal or one of the other entries, you can use that. Otherwise select "Name/URL". Enter your name in the pop-up box. You can leave the URL box blank, or type in a URL for your home page if you have one. but I'm told the system doesn't accept e-mail addresses in this field.

Then just type your comment in the text entry box and click on Post Comment. You'll be prompted to enter a captcha as an anti-spam measure, but apart from that you'll have posted the comment in your own name.

More on map size

I've got a rather nice fit of the hex grid with a slightly larger map. With judicious use of off-board areas (for Exeter, the Welsh coast, Norwich, Stranraer and the North of Scotland), I have a map that is 15 hexes long. For the southernmost third, it is 8 hexes wide; then the width drops to 6 hexes for a short section, while the northernmost half of the map is just 4 hexes wide. This gives a total playable area of 86 hexes, which might be just about acceptable.

The result is similar in overall shape and size to 1856, or to 1825 Units 1 & 2 combined. So it would be playable, but possibly slightly bigger than I had in mind.

"It's a by-pass. You've got to build by-passes."

Stuart Dagger raises an interesting point - given that the West Coast main line runs between Liverpool and Manchester, calling at neither, how do I plan to represent that historical route in Britain Under Steam? If the map was more detailed, I could include an extra hex between the two cities, but that would require the map to be even larger than 1829/1825.

Actually, I find this strangely liberating. There are a couple of similar areas where I have been trying to tweak the map to allow the historical routes. As examples, I want the LSWR London-Exeter route to pass north of Southampton, and the Hull-Leeds route to pass between York and Doncaster. I might possibly want to represent the Rugby-Stafford route north of Birmingham too. If I can find a suitable mechanism, perhaps I can use it in all such cases.

One approach might be to use new tiles that allow one route to by-pass a city while others go through or into it. Such tiles might give an interesting twist to the game, especially for players familiar with other 18xx titles. There would remain the question of why a train would choose to by-pass a lucrative destination, rather than include the city in its route. One answer might be to by-pass another company's tokens. Another might be to gain some other bonus, such as for the shortest distance or for the longest route - both ideas I am playing with anyway. Or perhaps the tile just won't have a suitable upgrade.

Another approach would be even simpler. I could just assume that the Manchester hex represents the larger conurbation, much as the Birmingham hex includes the cities around Birmingham itself. On this scale, there would be no need to represent the by-pass explicitly. The WCML does stop at Wigan and Warrington for connections to Liverpool and Manchester, and these towns are within the scope of these hexes. This might well be the simplest approach.

A third approach would be to make one of these hexes a brown hex, with track on the board at the start. I'd prefer one of the other approaches, on the grounds they should give more options during play.

The important question, of course, is which gives the better game?

How do I get the Hexes to fit Reality?

I'm staring at several maps which overlay hex grids on an outline of Great Britain. I'm trying to work out which orientation, size and positioning of the grid best suits the game. It should be possible to build at least the main historical routes, while giving plenty of alternative choices for players to build or block routes. Certain cities will be key - London is a particular pain because no routes run through it.

There seem to be lots of choices, none of which is quite right. When I get the hexes to fit around Central Scotland, for example, it throws off either the Midlands or the South of England. The latter is particularly difficult because there are so many east-west routes in such a small space and I want to keep plenty of choice in order to make the game interesting.

Of course, I could just map the hexes smaller, but that would make the game much bigger and (presumably) longer. Currently the most likely map has a grid of 13 by 6 or 7, although the northern half of the board is only 4 or 5 hexes wide, so the total number of playable hexes is approximately 75. This is the same number as 1830, so I'm assuming it's in the right ball park for a reasonable game.

I'll fiddle with the parameters and stare at the maps some more, but at some point I'll have to accept that what I've got is the best I can achieve and go with that.

Paying by the hundred

In the previous post, I suggested that companies should pay dividends for each £100 of their income. How can we make this work in practice?

The idea raises a couple of obvious questions. First, at the beginning of a game, will a company be generating £100? If not, it won’t be able to pay a dividend, players won’t get any income, and the game won’t progress. Even if we allow a dividend of £0 to increase a company’s share price, leaving players with no cash income would cripple the game (or at least seriously change it).

Second, if a company’s income is not an exact multiple of £100, what happens to the remainder after the dividend is paid? Does the company get to keep it? This might remove the pressure on deciding when to pay dividends and when to keep income for the company itself, which is a key strategic decision that I don’t want to dilute. Alternatively, is the remainder just lost (as “administrative charges”)?

Third, will this rule have any effect on how the share price changes? In some games, the dividend paid must be more than the current share value in order for the share value to increase. Should this be based on the company’s income or on the dividend that it pays out?

Last, will this rule discourage companies from trying to extend their routes, until they can make significant changes? And would this be a bad thing? In many 18xx games, companies gradually optimise their routes, perhaps increasing the value of a station or bypassing a low-income station. If a change only adds £10 to the income, will it be worth doing? Arguably, the game might flow faster if players aren’t constantly looking for small increases in income.

My current answers for the above is that share prices will only rise if the company’s income (not dividend) is equal to or higher than the current share value. If the income is £100 more than the share value, the increase will be double. This will give some incentive for companies to keep increasing their income, while at some point small track changes may become irrelevant. This is to the good; I want the game to focus on strategic decisions rather than small-scale optimisation. Any remainder after paying dividends will be lost. (Possibly I might allow the company to keep £50 if the remainder is £50 or greater, but this is a matter for playtesting).

The key question is how to ensure that companies can earn enough to pay a dividend at the start of the game. This is a matter of map design and ensuring that companies can run enough routes. It may be necessary to increase the base value of stations – we shall see.

Simplifying the core system

A lot of gamers are put off 18xx games because they seem too complicated and too time-consuming. We who love the system have various ways of keeping the game flowing. Some people use poker chips instead of paper money, which raises the question, why are we still making games with paper money? I'm not alone in sometimes using a spreadsheet, but you can see why many people would think that any board game that requires a computer is too complicated for them. Also, it means I need to carry a laptop around as well as the game itself.

So I'm looking at ways to simplify the game mechanisms while keeping the core ideas and level of strategy. Here’s one idea that may work.

The core mechanism of 18xx is that companies get income (in multiples of £10) from running trains. They then distribute 1/10 of this income to each share, so that a player holding 3 shares gets 3/10 of the company’s income. This mechanism was introduced in the very first 18xx game and has remained basically unchanged since. It was a brilliant design and has served all of us well. It's also the main thing that slows down the game, as the cash gets handed out to each player during every company’s turn.

My idea comes in two parts. The first is to make the dividend payments be multiples of £10. This will simplify the amounts that get paid out, keeping them to a smaller range. This enables the second part, which is to record player income by moving counters on a track, instead of handing out cash from the bank. This will be much quicker than handing out paper money (or poker chips) and making change each time.

For example, the track might have 50 spaces numbered from £0 to £490. If a player’s income reaches £500, a chit could be placed underneath his or her marker to show the extra income. The track could be placed wherever it fits best on the board; some eurogames have tracks around the edge of the board.

In this design, payments can only be kept to multiples of £10 if companies pay dividends for each £100 of their income, instead of each £10. So that is the core change of this approach. For example, if the LNWR has an income of £230, it will pay £20 to each shareholder (instead of £23). If I had 6 shares, I would move my treasury marker 12 squares along the track.

It's worth mentioning some secondary effects as well. We're not just swapping cash notes for markers on a track. We're also removing all the £1/£4/£5 denominations from the game, with reduces production cost and the space required to play. We are also simplifying the arithmetic.

The big question is how this change will affect the rest of the design and the play of the game. That is the topic of the next blog post.

Acknowledgements

Many people within the 18xx community have unknowingly influenced this project so far.

I copied the notion of keeping a blog about the game from Todd Derscheid, who is recording the development of 18IA (set in Iowa) at http://18ia.blogspot.com/.

The idea of a game aimed at beginners, with a friendlier name then 18xx, came from Bart van Dijk's Steam Over Holland. Dick's game is reasonably short and friendlier than some, but I want to simplify the core systems more.

The people on the 18xx mailing list have a vast experience of playing and designing 18xx games, which is fantastically useful. I've taken ideas and inspiration from several people there; I'm sure they'll disapprove of what I've done with them ;-). I can't name them all but they certainly include Mick Hutton, Lou Jerkich, David Hecht, Steve Thomas, Robert Jasiek, John David Galt, John A Tamplin, JC Lawrence, Bruce Beard, Ian D Wilson, and many more.

Finally and firstly, to Francis Tresham, without whom, nothing.

Current Status

I'm still early in the design of this game. I have a lot of ideas, which I will flesh out and record on this blog. (In part, I'm blogging here partly to record these ideas for my own reference, as well as to welcome comments from anyone who cares to read my musings). I have created an outline map, overlaying a hex grid on the UK coastline, adding the major cities and tracing the major rail routes to ensure that the historical outcome would at least be possible. This in turn has suggested some new track tiles to add to the game, and some ideas for handling routes into London.

What I plan to do next is to draw a full size map and experiment with the options it give for route building and competition. In parallel with that activity, I will sketch out the flow of the financial side of the game. With luck, these two lines of development will fit together and produce a game that's decent enough to start a long process of playtesting and tweaking. It will be interesting to see what happens.

Quick and tournament versions

There is a tension between the twin goals of appealing to eurogamers and of keeping the essence of the 18xx system. As an example, many 18xx games begin with an auction of private companies, each of which bestow special features on the owner. A novice doesn’t know how much these features are worth, and players tend to value them differently anyway.

I plan to offer two versions of the rules. The simple version will handle the above problem by dealing out the private companies to the players, rather than having an auction. This is not a new idea; Steam Over Holland and 1889 are just two existing games that do this. But I may make a similar distinction in other parts of the game, if it becomes appropriate. So if I face a design choice between a fast-playing option and a more complex alternative that offers more depth or balance, then I will adopt the first option for the quick game and the second for the tournament version.

At least, this is my intention. We’ll see how often the need arises.

Britain Under Steam

I’m creating this blog to record my thoughts and ideas for a new game in the 18xx family. My aim is to produce a game in the mould of 1829, 1830 et al that will be popular with euro-gamers, simplifying some of the core systems while keeping the range of strategies and decisions.

For general background on 18xx games, see http://www.diogenes.sacramento.ca.us/18xx_net/

Britain Under Steam will cover most of Britain on a single map and should be playable in 2-3 hours. Why Britain? Just because it’s where I live and the geography I know best. This is not a new setting for 18xx – quite the opposite, as the very first 18xx (1829) was set in the UK. One of my favourite games is 1825, which is sort of an update of 1829, but it only offers the choice of a 3-4 hour game in one part of the UK or longer and larger games covering larger areas.

Britain Under Steam will allow a variety of opening strategies, rather than having companies start in a fixed order. This is unlike 1829 and 1825, but like many other 18xx games that have a wider choice of starting companies. However, I hope to retain the flexibility that encourages players to invest in a range of companies, in contrast to some other 18xx designs.

Most 18xx games take longer than 2-3 hours (except for very experienced and fast players). I have some ideas which I hope will speed up the game system. This blog will examine some of these suggestions.