Reiner Knizia's Royal Turf is one of my favorite games. I've only played the Alea version, but most of my comments here will probably also apply to the older Turf Horse Racing and the newer The Winner's Circle.
I only play with the "hidden bets" variant.
Games don't get much simpler than Royal Turf. Roll the die, and then move one of the horses a number of spaces equal to the corresponding icon on the horse's tile. Each horse can only move once per round. Play until three horses cross the finish line, then pay out bets that were placed before the race began.
One element of the game that I would like to eliminate is the the lookup table for betting payouts. I hate lookup tables, and will always eliminate them if possible. In this case, I think it would work fine to just have three fixed payouts for the first, second, and third place horses, with each payout being evenly divided (rounding down) among all of the bets. (Probably the total payouts would have to be larger—for instance, $2000 for first, $1500 for second, and $1000 for third.) I think that this would also improve the balance of the game, as it more adequately rewards players who bet on long-shots that no one else has bet on.
One could argue that the "pace-setter" rules add complexity without really improving the game. The rule doesn't particularly bother me, but I wouldn't mind eliminating it.
The only glaringly inconsistent rule in Royal Turf is the one that states that on the third and final race, all payouts are doubled. I don't like this rule, and I prefer to play without it. I believe that it's intended to address a functional issue, which I'll discuss below.
It's extremely easy to see and understand what's going on in Royal Turf. Everyone's bets are hidden, but that counts as mystery, not opacity. There are very mild memory-elements in the game, as you sometimes want remember which horses the other players have helped or hobbled, but those aren't memory-conditions that must be remembered in order to follow the rules correctly.
Royal Turf makes excellent use of the randomness of a die roll, while the hidden betting adds an interesting element of psychology.
Royal Turf exhibits an excellent amount of variety, of a few different kinds:
Royal Turf exhibits a few minor functional problems, none of which are serious enough to cause me to lower my rating here.
One problem is that, if you fare very poorly in the initial two races, you may feel as though you have no chance of winning the game, no matter how well you do in the third race. I assume that this is the rationale behind the rule that all payouts are doubled for the third race. This smells to me like a rule that was added or requested by the publisher, although that's pure speculation on my part.
I prefer to simply ignore this rule. In fact, I'm happy to treat each race as a separate game, and just play a bunch of games in a row, like potato-chips. Another possible solution would be to keep playing races until someone reaches a pre-determined score. Since it's always possible for the leader to score no points (or even negative points) in any particular round, it will always be possible, in principle, for one of the trailing players to come from behind and win.
Another problem is that the game can exhibit a mild kingmaker effect. Sometimes by pushing your own horses forward you more or less knowingly throw the win to some other player. Uncertainty about where players have placed their "2" bets and "0" bets can mitigate the problem to some extent. Below, I suggest a betting variant that might reduce the problem even further.
A third potential problem relates to the variable line-up of the horses behind the starting line. The best strategy may always be simply to bet on whichever horses are at the front. One solution to this problem is simply to start all horses in the same space, which is apparently how it works in Turf Horse Racing. On the other hand, I'm not entirely convinced that the variable line-up is broken, because the payouts are substantially reduced for horses that many people have bet on. If you're the only player who bets on one of the trailing horses, that bet will not pay out as often, but when it does, it will pay out an amount that's many times more than the payouts of the conservative bets.
A related problem is that, all else being equal, Othello (the black horse) may be objectively the best, and Earl Gray may be the worst. At first it doesn't seem as though this can be true, since, mathematically, each horse has the exact same expected-value. But the issue is one of volatility. Earl Gray's numbers are significantly more "spikey" than Othello's. If the track were 1000 spaces long, this volatility would wash out, but given the actual track-length, Othello may have the advantage. This can be seen more clearly if you imagine starting all horses on the same space, and putting the finish line just 10 spaces away. Sure, Earl Gray can win in a single shot, but it's actually easier to roll Othello's best number twice in a row than it is to roll Earl Gray's best number once. On the other hand, it's possible that Othello's steady pace is more likely to land him in the middle of a pack, causing him to slide backward and lose movement-points.
It looks as though the standard rules of The Winner's Circle have a pre-determined line-up, with Earl Gray always in the front and Othello always in the rear. Perhaps that represents Knizia's or the publisher's attempt to fix this "problem". For my part, I suspect that the the game is perfectly functional as-is, variable line-up and all. I suspect that the hidden bets, varying payouts, and gameplay "groupthink" override any small mathematical inequities.
In Royal Turf, players spend all of their time concentrating on high-level emergent properties. They think about which horses can still move on this round, where they're likely to land, which gaps they're likely to fill, roughly how many rolls each horse is away from the finish line, which horses will likely be available on their upcoming turns, who is helping and hobbling various horses, and so on.
Administrating the gameplay in Royal Turf is extremely simple and straightforward. The one fiddly aspect of the game is that players often need to remind themselves of their own bets after they've placed them, and the betting chits are too small to be handled gracefully. Too often, players accidentally flip over their own chits while they're checking them, or even worse, flip over their opponent's chits.
We once tried the rule that you're simply not allowed to look at your chits once you've placed them, but this created an unacceptable lack of clarity. As silly as it may sound, sometimes I just can't remember where I put my "2" and where I put my "0". This leads to serious blunders, as I may play an entire race accidentally pushing my "0" horse forward.
Ultimately, I consider this whole issue to be a problem with the components rather than the ruleset, which is why I still rate the gracefulness of the game itself as excellent. If the betting chits were bigger and heavier, and if there were a better place to put them, it would be easy to re-check them without flashing them or flipping over other chits. Below, I suggest a betting variant that addresses this problem in a different way.
Royal Turf is a paradigmatic example of a game that squeezes a healthy amount of deliciousness out of a handful of simple decisions. There are only between two and seven options on any given turn (ignoring the "no-decision" turns when there's only one option), but somehow the choices always seem compelling, and there are almost always at least two different options that look appealing.
It's true that there's a somewhat destructive element to the gameplay. Sometimes your best option is to "hobble" a horse that you haven't bet on. However, other players very often end up helping you push your horses forward, so on the whole, the gameplay feels more positive than negative.
One other mildly destructive element is that you will actually lose money if you bet on the horse that comes in last. That rule doesn't bother me, but once again, anyone who dislikes it can just play without it.
Royal Turf is a mini-masterpiece, and is one of the few games that I rate as excellent in all nine categories. It is one of my guiding-lights of game design.
Although I love the game as-is, I can't resist thinking of alternative betting schemes that may reduce the mild kingmaker issue, and address the gracefulness issue of accidentally flipping over betting chits. Here's my current suggestion:
There's a deck of cards that contains (say) ten identical cards of each horse. At the beginning of each race, shuffle the deck and deal the cards out evenly to all players, discarding any extras. Set up all the horses for the race. Now, all players place their bets simultaneously, by keeping four cards and discarding everything else. Run the race as normal, and at the end of the race, everyone reveals their cards. If four "Red Fox" cards were played, each "Red Fox" card is worth a fourth of Red Fox's total payout, and so on. Notice that it's possible to place three or even four of your own bets on a single horse.
This variant should reduce kingmaker problems, because it's no longer clear which of your opponents you're helping when you push your own horses forward. It should also eliminate the fiddliness problem, because now you keep your betting cards in front of you and can easily check them at any time.
This variant is more psychological than the standard rules. You have no visible information about where players have placed their bets, so you'll have to induce it entirely through their actions. When placing your bets, you won't have as much ability to knowingly select long-shots that no one else has bet on. I haven't had a chance to test this idea, so I don't know whether or not I would miss that ability.