Tarot History Forum

A full tarot deck consists of 78 cards. According to the classical rules they have a combined value of 78 points to be divided among the (originally 3) players. 52 points are contributed from the whole card values of 0-4 points, plus 1 point for each won trick, which was absorbed into the cards as 1/3 point each. So each card has a value from 1/3 to 4 1/3 points, and their sum is 78.

There is no compelling reason on behalf of gameplay why it should be this way. The vast majority of point trick games have different card number and points to score, including most modern tarot games (e.g. French tarot has 78 cards and 91 points; Austrian 54-cards tarot has usually 70 points; etc.).

But it's also higly unlikely that this has been a coincidence.

So, how did it come that tarot has 78 card worth 78 points? Are there any hypotheses? Evidence?

(I have my own - entirely speculative - hypothesis, but will wait before presenting it - maybe this topic has already been settled.)

KnightMove wrote:A full tarot deck consists of 78 cards. According to the classical rules they have a combined value of 78 points to be divided among the (originally 3) players. 52 points are contributed from the whole card values of 0-4 points, plus 1 point for each won trick, which was absorbed into the cards as 1/3 point each. So each card has a value from 1/3 to 4 1/3 points, and their sum is 78.

There is no compelling reason on behalf of gameplay why it should be this way. The vast majority of point trick games have different card number and points to score, including most modern tarot games (e.g. French tarot has 78 cards and 91 points; Austrian 54-cards tarot has usually 70 points; etc.).

But it's also higly unlikely that this has been a coincidence.

So, how did it come that tarot has 78 card worth 78 points? Are there any hypotheses? Evidence?

(I have my own - entirely speculative - hypothesis, but will wait before presenting it - maybe this topic has already been settled.)

I agree, that it isn't an accidental "78 points".

I don't like these "1/3 to 4 1/3" points, as this is just a way to hide, that the "essential old rule" should have been just trick-counting, not card counting. 78 cards with 3 players have just 78/3 = 26 tricks, so 26 points for trick taking, and the other 52 for "special cards" (courts [40] and the trumps 0, 1 and 21 [12]).

I think, that it was "old custom" to arrange the game numbers in an elegant manner, later rules didn't care for the finesse.

For instance, if we take the old 52 card game, and assume, that ...

King = 4 points
Queen or Ober = 3 points
Jack or Unter = 2 points
Ace or Ten = 1 point
13 points for the tricks in a game with 4 players (4x13=52)

... we would get 53 points in a game with 52 cards. "53" possibly was chosen instead of "52" to have a tie-breaker.

The 52 cards game should have been older than the Tarot game.

There are some other old games, which look being arranged with "elegant numbers".

In this question it's naturally of importance, since when - to our knowledge - the game had 78 cards and since when we know of 78 points.

The sure begin of 78 cards - to our knowledge - is given with the Boiardo poem, I think, that it is from 1487, other think of something between 1461-1494.
Some assume, that the 78 cards state goes back to the begin of the Trionfi decks, I've doubts about this. I thuink, that there was an experimental stage of development with partly 16 special cards (trumps) and partly with 14.

I think we know of 78 points since the Tarotica document. We talked about this here:

viewtopic.php?f=11&t=767&hilit=casseria

That's a relatively new finding.

I personally think, that there once had been a 5x14 Trionfi deck with 4x14 usual cads and 14 trumps in a trump suit.

If I assume ...
4 points for kings and the 3 "trumps" Fool, lowest trump and highest trump
3 points for queens
2 points for knights
1 point for the Fante
.. I would get 52 points, to which I add 17 points for the tricks in game with 4 players (2 cards may play the role of the discard; 4x17 = 68, but the deck has 4x14 = 70). This makes 52+17 = 69 points, eventually another 1/2 point for the discard (69 + 1/2 = 69.5.

Then this system looks like a tiebreaker solution.
Trionfi decks with 70 cards are noted in a Ferrarese account book document in 1457.

I am revisiting this topic because of an interesting article in The Playing-Card in Oct.-Dec. 2020 by Ben Steele that attempts to take the discussion further. But the subject seems to me rather complex, with a great number of factors to consider, more than Steele takes into account. I am trying to make an evaluation, but my grasp of all the various relevant games is none too sure, and arguing from them is rather tricky. Please let me know where I have made mistakes of fact or have unconvincing arguments. Steele's article is at https://www.academia.edu/44271904/Recon ... e_of_Tarot.

I will start with his conclusion. After reconstructing the "original game" and finding that its hypothesized rules lead to a total score of 77 points, Steele says:

Did the designer(s) of the Tarot construct both the pack and the rules for the game it facilitated in such a way as to intentionally parallel the trump suit’s allegory and to achieve some manner of symmetry between the two? If we take all of the foregoing assumptions as representative of the original game of Tarot, then I think a strong, but inconclusive case can be made for it and even Michael Dummett pointed out that this sort of thematic symmetry “strikes me as just the sort of conceit that would appeal to men and women of Renaissance times.”42 If the Tarot did partake in this “Renaissance conceit” it exhibited some degree of symmetry and interrelation between the selection of the trump themes, the design choices and configuration of the pack, and the rules of the original game. In consideration of Tarot’s allegory, it seems reasonable to conclude that the constitution of the trump suit was set as precisely 21 trump cards and the matto from the very beginning. By extension then, any early pack of a different constitution, and not being a known variant of the 78-card pack, would not qualify as a Tarot pack since our hypothesis rules out any gradual development of the Tarot’s trump suit.
______________
42. Dummett, op. cit. [Game of Tarot], p. 426

That the trump suit-must have been 21 trumps and the matto from the very beginning and of 78 cards is a rather strong statement and seems to be contradicted by what Huck said in the preceding post, that a 70 card game with four players would have 69.5 points. That is so close to 70 points as to be virtually the same. In fact, if instead of getting .5 points for the discard, the dealer were awarded a whole point for his two cards, or if 1 point were awarded for making the last trick, it would be 70.

We might also consider a game played with the Visconti di Modrone (Cary-Yale) if it was a 5x16 deck. With an extra Knight per suit worth 2 points and a page worth 1 point, the points from court cards would add up to 52. Assuming no Fool present, the top and bottom trumps together would get 8. The number of tricks in a four-person game would be 20. The total is 80 points, with no discard and no points for the last trick - the same as the number of cards. If, as some maintain, the Modrone had 89 cards (25 trumps instead of 16), there would be 22 tricks, 12 points from three trumps, and 52 from the courts, adding up to 86 points. If 3 points were given for the last trick, or 2 plus 1 for the discard, that would be 89.

In general, one can get the necessary equivalence simply by supposing different numbers for the last trick and the discard.

Steele, however, considers that there would have originally been four counting trumps (including the Fool), not three, and that 1 point each would be given for the discard and the last trick. There are 19 for tricks, 40 for the courts + 16 for trumps + 1 + 1 = 77. This is one less than 78, but he says that the Fool plays a special role in the game as first, last, and the journey in between, acknowledged by the designer by leaving him out of the count.

One has to admire his adherence to the principles he formulates. It would have been easy enough to give 2 points for the last trick, so that the total would equal 78, as the Tarotica document seems to imply, as well as the game of Trappola, where the total of 78 seems contrived to add up to that specific number.

There is also common sense: whatever the special role of the Fool is, it is still a card in the deck, played to tricks like any other card, even if the rules governing it are different. In fact, he even considers it to have originally been the lowest trump.

Steele devotes considerable space to why there would have been four counting trumps (including the Fool as a trump for this purpose) and not three. First, three is characteristic of places that derive the game from Milan and its type C order. In that case:

Not only is Bologna widely recognized for its adherence to tradition, but it is also an A-type order along with Florence, which is gradually revealing itself to be the original sequencing system of the Tarot.22 Bologna’s staunch traditionalism and observance of order A makes it a much more likely candidate for the preservation of the original features of the game than Milan, an area that appears more representative of the game outside of Italy than within it.
_____________
22. For an overview of some of the more pertinent research in this area, see Pratesi, Franco. IPCS Papers #7: Playing Card Trade in 15th-Century Florence, 2012 and Caldwell, Ross. “Giovanni Del Ponte and the Dating of the Rothschild Cards in the Louvre: Some Further Considerations.”
The Playing-Card, 36(1), 2007, pp. 51-62.

And second, even where there are more than four counting trumps - in Minchiate and Sicily - they are still type A, the original order, and include the top two in that order among them.

Dummett came to a different conclusion about Florence. For that city, Dummett, favored three trumps over four (still including the Fool). It is in Game of Tarot Ch. 21, which I have put online at viewtopic.php?p=19806#p19806), p. 420:

On the other hand, both in Sicilian Tarocchi and in Minchiate, though not in Tarocchino, there are signs that the three usual counting cards have a special role. In Sicilian Tarocchi this lies in the fact that these cards have a point-value twice that of a King or of one of the other four arie; in Minchiate, it consists in the existence of a versicola consisting of the Trombe, Uno and Matto, corresponding to the Angel, Bagatto and Fool.

Sicily is an indicator for Florence because Sicily got its game from Rome, which got its game from Florence. Minchiate is assumed to be Florentine. On the basis of Florence, therefore, the evidence suggests three counting trumps over four. The problem is that then, in a four-person game, there would not be 78 points - 19 points from tricks + 40 from courts + 12 from trumps = 71 - unless there were also several points from the last trick and the discard, of which there is no evidence in France and hence from Milan, our only source of information for that city. It is possible that the four-person game was never popular in Milan, or the lack of points for the last trick and discard were fixed at the time of the Modrone and never changed and there was no interest in having the total points equal 78 until the three-person game was dominant. But Florence remains.

In favor of four original counting trumps, Steele refers to Bologna as preserving the original game, based on its method of play being "staunchly traditional". But it seems to me that in one important way it was not so traditional: at some point, probably in the 16th century, it reduced the deck from 78 to 62 cards. Since that reduces the number of points from tricks from 19 to 15, it is a change that could well affect the total number of points. How the points in the 78-card game were counted is unknown. If there were 78 points, it could mean that originally there were only three counting trumps, allowing the points from tricks to then be 4 more in order to keep the 78-point total. Bologna did have, in the 62-card game, 6 points for the last trick and 1 point for the dealer's discard (see https://www.pagat.com/tarot/ottocento.html). Three counting trumps would yield 12 points from trumps, 40 from courts, 19 from tricks, 6 from the last trick, and 1 from the discard = 78.

Steele's other arguments for four seem rather flimsy. He says that four would correspond to the four counting suit cards, Kings through Pages, treated as a kind of sub-suit. But they are not a sub-suit. If they were meant to parallel the trumps, it would be just as logical to take the 1 or 2 lowest cards in the suit, the Ace and Deuce in Swords and Batons and the Ten and Nine in Cups and Coins, or the Ace and Deuce in all four. If the trumps were meant to parallel the courts, we would expect them to have progressively lower point values as they went down.

There is also the question of whether the Fool would have originally been the lowest trump. Even if it was true at some point early on, was it the lowest trump at the time when it was determined that the total number of points should add up to 78, i.e., sometime after there were 78 cards? We cannot assume that it was, if that is the conclusion we want to arrive at. And it may well be that the Fool was never a trump. If the allegory governing playing cards generally was one representing the social order, nobility and commoners, the character on the early Fool cards looks like someone with a lower status than any of them. He is not someone who would be expected to master the virtues and climb the heights of contemplation and valor.

Steele cites Andrea Vitali, who in one essay imagines the tarot giving the Fool "the power to reason" so as to ascend "the mystical staircase" of the sequence. But Andrea is speaking here of the time when the name of the game changed from "trionfi" to "tarocco", which is long past that of the first Fool card, and even the first tarot, whose number of trumps he considers might have been smaller.

I will move on to the other parts of Steele's presentation. That the court cards would add up to 40 is plain enough - unless, of course there were more or less than four of them per suit. There remains the number of points for the last trick and for the discard. Steele's principles for these are three.

First, he assumes a "simplified scoring system." By that he means that everything besides the counting cards gets 1 point: tricks, discards, last trick. His inspiration, he says, is Dummett, p. 427 of Game of Tarot. But Dummett is careful to restrict that principle to the tricks; Steele is not, despite all evidence to the contrary. Bologna gave 6 points to the final trick. Sicily, whose game came from Rome, which got it from Florence, gave 5; minchiate, from Florence, gave 10. France, and presumably Milan, gave none.

The latter contradicts Steele's second principle - "a score for the final trick" - and so, he says, was not part of the original game: That is because type A is the original trump order, where we always see points for the final trick. France is type C. The problem is that this assumes that the impetus for giving points for the last trick was not precisely, when first done, to make the total add up to 78. If it was the impetus, then no points needed to be given when not required. That would explain why in the classic non-Italian 3 person game with 78 cards, no points were given for making the last trick.

As to why just 1 point for the last trick, as opposed to 5, 6, or 10, Steele refers us to his principle 1, simplicity - even though all the evidence is to the contrary. (What is simplest, it seems to me, isFrance and Milan's solution: to have no points for the last trick or discard. Bologna gave 6 points for the last trick in the 62-card game. For the 78-card game earlier, the points for the last trick might have been adjusted to keep the total at 78, instead of the number of counting trumps. So instead of 6 points for the last trick, it might have been 2. So now, working backwards from 62 to 78 cards in Bologna (ignoring the criterion of "simplicity"), there are two ways the points might have added up to 78. One is with 40 points for courts, 12 for trumps, 19 for tricks, 6 for the last trick, and 1 for the discard = 78. The other is 40 points for courts, 16 for trumps, 19 for tricks, 2 for the last trick, and 1 for the discard = 78. Of these, the former seems to me the most reasonable. Neither is Steele's.

Steele's third principle is the necessity of points for a discard. Again he cites Dummett, but goes further than Dummett would have dared. Dummett says that the discard would have been a feature of "an early stage of the game." This is not the same as saying "the original game." In fact, Dummett goes to some length to question whether points for the discard existed before the 16th century. The games reported in the 1659 Maison Academique had no discard. One version does not specify how many cards are distributed to each player; the other specifies 12, with up to six players (Dummett and McLeod, Games played with the tarot pack, pp. 23-24). The likelihood is that there will be cards left over not in play. It is possible, he says in Game of Tarot, that these games go back to the earliest times, and that the practice of the dealer's picking up the cards left over from the deal only came in later, as a result of influence from another game, Scartino, which means "discard," in the years just before and after 1500. If so, a constant number of points in the game would not have been possible before then (p. 427):

If Scartino did influence Tarocco, it is possible that the practice whereby the dealer took extra cards and made a corresponding discard was not an original feature of Tarot games, but was incorporated into them about the beginning of the sixteenth century, in time for it to be imported, as a feature of the game, when Tarot arrived in Switzerland.

There is even a chance that the name change to "tarocchi" had something to do with discards of "empty" cards, since a likely root, according to the etymologists, was the Arabic "tarh" with the "h" pronounced like the "ch" in Bach, meaning the part of a load of merchandise that is removed from consideration as far as payment, as either packaging or defective, in other words discarded. Thierry Depaulis advanced that idea in his 2013 Le Tarot Revele.

A difficulty with postulating the discard at that late date, ca. 1500,, Dummett admits, is that it assumes considerable interaction between centers. Bologna is not far from Ferrara. Florence and Rome are more problematic, but interactions were common enough.

Would the game even have been played without all the important cards in play, due to a couple of cards left over unclaimed? Well, the 1659 French rules mention a 2-person game where an entire third of the deck is "dead."

Another issue about the discard, ignored by Steele, is whether more players than just the dealer would have participated. In the 1637 rules reported by the Abbe Morales, the rules stipulate that the dealer discards 3 cards and the others 1 (https://www.pagat.com/tarot/marolles.html):

The dealer deals 4 rounds of 5 cards to each player; then one more card to each other player and 4 to himself.

The two non-dealers each discard one card and the dealer discards four. In each case these cards count in the discarder’s favour at the end of the play. No trumps or Tarots may be discarded, on penalty of paying two marques to each opponent.

This will not change the number of total points. But if there is a connection between Morales and the game in Ferrara, as some have indicated, it reinforces one way in which Dummett interprets Lollio's game in ca. 1550: after the deal, two players discard 5 cards and the dealer 8, leaving them just enough cards for 20 tricks. 6 points for the last trick would put the total at 52 + 20 + 6, assuming no points for the discards. If so, in Bologna, there could have been 16 tricks rather than 19, even with 78 cards total and 4 counting trumps, if all the players had discards, 3 cards for three of them and 5 for the dealer. If there were no points for the discards, the total would be 16 + 56 + 6 = 78. Dummett concludes (p. 429):

Such a game, played with the full pack and without cricche or sequences, might be the remote ancestor of the Bolognese game.

So much for simplicity!

So there are quite a few possibilities. I think it is clear that having three counting trumps is at least as logical as having four, and the numbers for the last trick and discards quite variable. If so, the total points and the number of cards can be the same for a variety of packs. Or not, if that coincidence had not yet occurred to anyone as a nice "conceit," or if there was no fixed total.

However, like Steele, I want to take the argument a little further. Was the game originally one primarily about capturing counting cards? I wonder if it was not instead a game in which what mostly counted were the number of tricks.

The game described by Viti in the 1490s, using Boiardo's cards, fits that of tarot in most respects: 21 trumps plus a Fool (which may or may not be a trump), the usual 56 suit cards, following suit if one can, playing a trump if one cannot follow suit; but there are no counting cards at all. At the end of play, everyone counts up the number of cards they have won (https://web.archive.org/web/20100811191 ... Conclusion, discussed by Dummett p. 422 of Game of Tarot). The only difference is that cards in two suits count positively and the other two negatively. That is simply a variation on the usual idea of the order going up from Ace in two suits and down from Ten in the other two.

There is also the evidence of a 1477 statute about minchiate: payment is to be made directly on the basis of the difference between the number of cards taken. In other words, no cards with special points. Pratesi cites this statute in two of his essays, "Carte da gioco a Firenze: il primo secolo (1377-1477)," The Playing-Card 19 No. 1 (1990) 7-17. https://naibi.net/A/30-PRISECO-Z.pdf and "De l'utilité des jurons pour l'histoire de minchiate," L'As de Trèfle , N. 52 (1993) 9-10, https://naibi.net/p/51-JURON-ZOCR.pdf.

Also, there is the game with ordinary cards that the tarot inspired, sometime in the late 15th century, first in Spain, namely that called "triumphs," the game that allegedly forced the name change to "tarocchi". Neither the Spanish nor the French version had counting cards (see Wikipedia entry for "triomphe".) The English and the German versions did (giving points to the courts), but they were surely later.

Karnoffel, another game with trumps, documented by 1450 and with a possible relationship to tarot even if on the other side of the Alps, also did not have counting cards.

Marziano, in the main predecessor game to tarot, likewise says nothing about counting cards. His is clearly a trick-taking game with a trump-like special suit. On the other hand, it is possible that the Kings were counting cards, because he makes a point of mentioning them. If so, that connects Marziano with minchiate's practice of a few hundred years later, in which Kings were the only counting court cards. Kings are all-important in chess, where all that matters is being able to capture the King. On the other hand, Marziano, in mentioning Kings, may only have wanted to indicate that they were the only court cards.

There is also the possibility that combinations earned points in Marziano's game: 3 or 4 Kings, 3 or 4 of each of the orders of gods, or sequences of gods or birds or both together, using the association of groups of gods with suits of birds. Viti lists sequences as one of his "games" (parts of games, in Dummett's view). Viti's game allegorically also fits Marziano's: positive points for the two more virtuous suits, negative for the two less virtuous ones, then the taking of cards of their choice by those with more points from those with less, according to the difference in points, the winner being the one who has the longest sequence (see my previous references to Viti). In Marziano, that might include cards both above the King - the corresponding 4 gods - as well as the birds below. In short, an allegorical parlor game, one especially suitable for children. As we know, the decks of 1442 were ordered for the ducal children aged 8 and 10. Then, I am surmising, others wanted a game with more complexities and opportunities for strategy: so the game would have expanded the point-getting opportunities accordingly.

My conclusion is that the acquisition of points from counting cards (trumps, courts) might well have been a late development, when the allegories became less important and game strategy more. Then various numbers of total points equaling the total number of cards were possible, as long as all the cards were in play. If not, the idea of a fixed total is meaningless. Even in Bologna, with 78 cards and 78 points, the evidence points in various directions.

This is a difficult subject, and I found some of both Steele and Dummett hard to figure out. Please let me know what I have missed.

Hello.

This all seems interesting. Now, I'll offer some superficial reflections in case any might inspire you. Don't pay too much attention to me. Today I've read this thread and the one on tarot.

I don't understand what you mean when you talk about three or four trionfi. I imagine they are the cards called "tarocchi" in Bologna, since the Angel is sometimes included. The other three are the World, the Fool, and the Bagatto.

An other question. I don't know what you mean when you mention Steele. I think it could be the ancient manuscript or a text written by Robert Steele. However, someone said that Steele quotes Andrea Vitali, or so I read in the automatic translation of this thread. Robert Steele lived decades before Andrea Vitali. Therefore, I don't know what Steele means in this thread.

Otherwise, I'm trying to understand the debates about numbers. I thought the game of Tarot was traditionally played between four players. You say that Tarot was originally a game for three players. This reminds me of the ancient Chinese game whose name refers to a horse hanging by one leg, so that three legs are free. I believe that in that game there were four players, three of whom played against the fourth player, who was the bank. I may be inaccurate in what I just said, but I have a rough memory. What is our basis for stating that the Tarot was initially a game of three players?

On the other hand, it has occurred to me that perhaps the total number of cards in the deck was intended to reflect the maximum lifespan of a human being. At 63 years of age, a great decline arrives, according to some philosophers and astrologers who took into account the seven-year cycles that Saturn takes to orbit the sun, if I remember correctly. Tarocchino is played with 63 cards, and the Goose game has 63 spaces. Other traditions could estimate that human beings tend to die at 78 years of age (Marseilles Tarot) or at 97 years of age (Minchiate II). This occurrence isn't very important. In any case, it's an occurrence compatible with other explanations about points and such.

--------*******--------

On the other hand, I found it interesting to read something about Savoy and the Jews. You say there was a Jewish calendar compatible with the structure of the Tarot. And you say the Jews left Savoy but remained in one locality.

I think the Trionfi deck emerged with 16 trumps or less and was later expanded. I think it was expanded to 78 cards after 1465 and before 1502. I think perhaps the Tarot didn't yet have 78 cards when Boiardo wrote his poem about a deck of 78 cards that had different trumps than the Tarot. We have discussed the dating of Boiardo's poem and the dating of the Sermon discovered by Robert Steele. I think the Tarot reached 78 cards approximately between 1481 and 1502.

I don't know if the 78-card structure originated in Italy or perhaps it was a variant that emerged in France. An interesting hypothesis could be that of Piedmont and Savoy, which was located between Italy and France. Perhaps crypto-Jews had an influence on the consolidation of this 78-card structure, or perhaps Jews were not important to the Tarot. I believe I have read that in France there was not as much of an Inquisition as in Spain and Italy. That could perhaps have facilitated the consolidation of a Hebrew or Jewish initiative more easily.

Personally, I think the logic behind 22 Tarot cards is that there are 20 cards plus an end on each side, as we see in Tarocchi Met-Budapest, where the Fool and the World are the only ones without a printed number.

Magic Moon wrote: 07 Oct 2025, 16:45
An other question. I don't know what you mean when you mention Steele.

MikeH begins his post with:

"I am revisiting this topic because of an interesting article in The Playing-Card in Oct.-Dec. 2020 by Ben Steele" and gives a link to his article on academia edu site.

His following references to Steele are to this Ben Steele and his article. Basically, as far as I understand it, Steele argues that the Tarot pack and its rules were deliberately designed to create numerical symmetry, with 21 trumps plus the Fool from the very beginning, yielding 78 points in the “original game.” Steele suggests four counting trumps (including the Fool), simplified scoring, and fixed values for the last trick and discard, arriving at a total of 77 points. MikeH disagrees, saying old decks had different numbers of cards, and one can change how points are given for tricks or discards to reach various totals in several ways. Some traditions counted only three trumps, some gave no points for discards or the last trick, and so on. He posits that early Tarot was mostly about winning tricks, and the “78 points for 78 cards” idea was probably a later conceit.

Magic Moon wrote: 07 Oct 2025, 16:45 Otherwise, I'm trying to understand the debates about numbers. I thought the game of Tarot was traditionally played between four players. You say that Tarot was originally a game for three players.

In La Maison académique contenant les Jeux (1659) a "Swiss" variant of The Pleasant Game of Tarot
in which there are three players is described, the dealer has 28 cards and discards three and the two others are dealt 25. A similar game for three players, with the variation that each player is dealt 26 cards and each discards one, is described in the Tarotica (1589) text. The Tarotica text also has a somewhat ambiguous line that one can sort of make sense of if one interprets it in light of the 78 point variation, which would make it the earliest refence we have to 78 points if correct.

In the rules as written by Abbé Michel de Marolles,1637, we have:

"At most three people should play this game, and it is not very enjoyable with two, since even then one must suppose a third, called the Dead, from whose hand as many cards are drawn at random as the others have, to be carried off by the strongest player. Nevertheless, those who are exceedingly fond of the game may sometimes amuse themselves in this fashion."

Mais auant que d'entrer plus auant dans le destail de ce jeu, il me semble a propos de dire qu'il n'y faut estre que trois personnes au plus, & qu'il n'est pas fort agreable a deux, estant mesme encore necessaire d'y en supposer vn troisiesme que l'on appelle le Mort, duquel l'on tire selon le hazard autant de cartes que les autres font de mains pour estre emportées par celuy qui est le plus fort. Neantmoins ceux qui l'ayment extresmement s'y peuuent quelquefois diuertir de la sorte.

On the Tarotica text see: viewtopic.php?f=11&t=767&hilit=casseria

The main argument for the originality, or at least ancientness, of the four person game is a 1456 comment by the Ferrara jurist Ugo Trotti, of which the most relevant part is the following (viewtopic.php?t=379, based on Ortalli, which I reproduced at viewtopic.php?p=27865#p27865):

Ex hiis infertur quid de ludo cartarum qui hodie multum frequentatur, qui tamen multiplex est et quandoque plus habet industrie quam fortune. Veluti si quattuor bipertiti ludunt ad triumphos. Interdum eorum plus habet fortune quam industrie, ut si in 3a vel in 4a vel falcinellis ludatur.

which Marco translated as

From those [other games?] the game of cards, that today has many followers, stands out, because it is varied and it depends more on skill than on luck, in the case in which four people divided in two couples play a game of triumphs. If the third, the fourth, or falcinello are played, it must be judged in the same way as a dice game, because it depends more on fortune than on skill.

It is clear that the four-person game of trionfi is referred to. No other number of persons is mentioned, but this does not mean that there were not games for other numbers of players, somewhere in northern Italy at that time. I say "somewhere," because a well-known fresco in 1440-1450 Milan shows what appear to be five persons playing a card game with large cards suggesting trionfi. I read Trotti's sentence as being mainly about the issue of skill vs. luck, and he is saying that the four-person game of trionfi is a good example of a card game in which this is true. I would amend Marco's translation slightly to insert the word "as" after "luck,...", i.e. "luck, as in the case...", because the word "veluti" means "as in", according to Wiktionary. I can imagine that a partnership game, involving much cooperation between partners, requires a good deal of skill, more than a three-person game where each player plays individually. The other games mentioned, "in 3 or in 4" and falcinello, are mentioned in the Steele Sermon (different Steele); they are examples where luck plays more of a role than skill.

Huck in his example of a game with a 70 card deck in Ferrara of 1457 was assuming a 4-person game at that time and place. That seems to me reasonable. On the other hand, Lollio in 1550 Ferrara describes a 3 person game, on Dummett's analysis of it in 1980 and I think everybody's subsequent opinion. Steele says that the game would have shifted from a 4-person game originally to a 3-person game in Ferrara. Well, maybe so. But maybe that wasn't true everywhere, in the sense of one form predominating early on.

It is not that I think that either the four- or three-person game was "the original". If I have any position at all, it is that probably there were various forms from the start, or close enough to the start as not to make any practical difference, and that the 4 person game probably required the most skill from all the players; but which form was more popular may have varied from place to place. But Galeazzo Maria Sforza, for example, might not have liked playing with a partner. Also, it is not clear to me that there would have been any fixed number of points in the game "originally," whether 3 or 4 person. In part, this is because we have no idea if all the "counting cards" would have always been in play, as opposed to being in a small pile of cards left over from the deal but not claimable by anyone. More fundamentally, we have no idea to what extent it was a "points" game at first at all, as opposed to being primarily a matter of the tricks won, as in other 15th-century games with trumps that might have been related to trionfi.

SteveM's summary of my presentation is quite accurate. Thanks.

My argument against Ben Steele was not against his assumption of a four-person game as "original" but rather the assumption that there would have been four counting trumps, i.e. World, Angel, Fool, and Bagatto. It is not that I think there were only three: I think there is not evidence to say with certainty one way or the other, but that it is slightly more reasonable to think that it was originally three, because the only evidence of four is from a 62-card game, which would have had 4 fewer points from tricks. In that situation, it is reasonable that the number of counting trumps would have increased by one accordingly. This is still with a 4-person game.

I would welcome more questions, where what I say is not clear. In general, I am presuming the previous posts in this thread as background. And of course I welcome disagreements as well.

Hello.

Thank you very much for both explanations, which are very useful for understanding this interesting subject.

It seems possible to me that a three-player tradition that emerged in Ferrara spread to France, as was later recorded in the Academy. It seems difficult to determine whether it was originally a game for three or four, or another number of players. Perhaps it was a game for four when the deck supposedly had about 14 trumps, and later a three-player variant emerged for the extended deck of 77 cards.

Steve wrote:
<< Mikeh wrote: "I am revisiting this topic because of an interesting article in The Playing-Card in Oct.-Dec. 2020 by Ben Steele" and gives a link to his article on academia edu site. His following references to Steele are to this Ben Steele and his article. Basically, as far as I understand it, Steele argues that the Tarot pack and its rules were deliberately designed to create numerical symmetry, with 21 trumps plus the Fool from the very beginning, yielding 78 points in the “original game.” Steele suggests four counting trumps (including the Fool), simplified scoring, and fixed values for the last trick and discard, arriving at a total of 77 points. >>

I am interested in the 77-point option because I think the Tarot possibly had 77 cards in Rosenwald. I should study why Ben Steele said 77 points.

*********

I don't know the Tarot game forms in detail. I know that those three or four trump cards are special, at least in the game tradition of Bologna, France, and Germany. The rules are different in those three places, but those three cards are usually special. You say they are counting cards. I don't know what that means exactly. But they was special cards, ok.

Nice!

Counting cards are cards with special points based on what they are, by themselves rather than in combinations. Typically, they are the court cards plus some of the trumps (Bagatella, Fool, World, in Bologna also Angel, in Rome and Sicily a few more - World, Sun, Moon, Star-, in minchiate even more, but of the court cards only the Kings).