Cross post from BGG, but think it would be valuable here as well.
Been thinking about a Ninuki Renju/Pente Variant that combines the Renju forbidden rules, with Pro Pente, and Boat pente as a way to balance the FPA issues of both Pente and Ninuki Renju.
Renju currently has an FPA of 54%, which is actually lower than chess' when you exclude draws, and, because of its asymmetric rules, it also has greater depth and variety to it than gomoku in terms of strategy.
The highest ELO rating in Renju is currently in the 2800's while Gomoku is only around 2600.
With Penju I wonder about the best of both worlds. The simplicity of the Pro rule vs Soorsov-8, the balancing effect and asymmetric play of the forbidden rules, the tactical richness of Pente's captures, and the ninuki renju "boat" rules added on for good measure.
I think it would be a really interesting and deep game to play.
So, to be clear... you're proposing overlines forbidden for the first player, but the second player can win with them unless they can be broken with a boat capture? I suppose a ninuki draw could still occur in that case, unless the overtime boat capture rule doesn't apply for both sides. Hmmm.
Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
What I'm not clear on, though, since I don't play renju much is whether they're forbidden for both players in renju? I think if it's to get more traction than ninuki has (so far) it needs to follow renju rules as closely as possible.
Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
Gotcha. I thought 6s could win for 2nd player but wasn't sure. So, in your proposal it seems like the first player could achieve a ninuki draw by making a fifth capture across an imperfect 6 and leaving a 5. But, in boat Pente, that would be a losing move, rather than creating a draw.
Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
I've been thinking more about different variants lately (ninuki, Rollie's ultimate Pente, swap2 variants, etc.) I think it's still an open question how much of a FPA there actually is in pro rules ninuki. I think swap2 options are ultimately going to be less than optimal in games which don't have the possibility of drawn positions (penju, ninuki and ultimate do have draw possibilities, but swap2 boat doesn't). Ninuki has been available for play with pro rules at PBeM for well over a decade (perhaps 2 by now) but never caught on here or elsewhere. 13x13 Pente, where overlines don't win (draw possible, but I don't know how their code counts a fifth capture which makes an overline into a 5 - anyway, likely a strong FPA) has been playable for a long time on flyordie, but hasn't ever caught on elsewhere. S Pente, proposed by Oleg Stepanov has 3 swaps, but no draws. It's playable on PBeM, but hasn't been explored much (to my knowledge). Swap2 Pente (and variants) seem to be making recent progress on other sites, but once again no draw possibility. Ultimate is a swap2 variant, close to gomoku in that overlines don't win and allows for draws without asymmetrical rules. Whether it will catch on anywhere remains to be seen, of course. ETA: I think unless someone with access to a supercomputer takes it upon themselves to evaluate different variants in terms of FPA, we have more than enough variants. Unless one or more of them converts gomoku and/or renju players as new players, I don't think any of them have much to recommend them above the others beyond single game decisiveness in live tournaments (swap2 variants) and draw potential (overlines don't win variants).
Message was edited by: watsu at Jul 24, 2022, 4:33 PM
Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
You have to mention the standart deviation. The main idea beneath this question is that in Renju chances for both sides are mathematically equal, and that's not the case for Chess.
And if you are relying on some statistics like renju.net games or anything alike, you have to know that this data consists of games played with a different opening rules. Different opening rules may give different impact on the practical 1st player percentage.
> The highest ELO rating in Renju is currently in the 2800's while Gomoku is only around 2600.
In renju we don't calculate ELO ratings anymore since 2021. The official rating list is here: https://rating.renju.net/ and it's not ELO but WHR, it's a different math approach.
And last but not the least, there was a voting about the renju opening rules during GA 2021 and only Japanese veto prevented Renju from introducing Taraguchi-10 for WC/TWC. By the way, EC, as well as YEC and YWC are being played by T10 for several years already.
So don't hate S8 rule because it's too complicated. Use 5 swaps instead =)
I'd have to go find the data again exactly but it was across hundreds of games using the Soos?rv-8 opening rule specifically. Yeah. Not in general.
This is also with draws excluded from the data set. Just black and white wins.
As for Elo I'm aware of the WHR rating system change, they use it in Tak as well actually, and yes I mispoke, sorry about that. I tend to refer to things as Elo out of habit even when its WHR or Glicko-2.
As for complexity, the pro rule is by far simpler than all of them, swap2 included. It's not effective, but the overhead is much less than most renju opening rules in general. That can't really be argued.
Pro rule is won for Black, and that is all you have to know about it. It is impossible to play with pro rule on a high level. By the way, gomoku on this site doesn't use any opening rule at all, such a pity =(
And let us discuss the opening rule complexity once more. You know, I'm a renju coach, and I do really see the complexity of different opening rules. T10 for the very beginners is like 'no opening rule'. After a while they understand that there is a possibility to make a swap at some moments - and bingo, they got it completely. 10 alternatives is a really rare scenario dedicated for professionals. And, once more, this rule gives complete equality. A huge distance from the pro rule which is just won for those who got smth like 6th dan in renju.
Yes, I agree that S8 and swap2 rules are quite complicated for beginners. Well, if you compare them with a single castling rule in chess maybe you will change your mind: these rules have moderate complexity =) Anyway, 5-swaps rule is much easier. Only playing without any opening rule goes beyond it. On the other hand, if you are going to master pro rule, it will take you nowhere. There are no serious championships, no tournaments with this opening rule: a lot of players are able to master this rule enough to earn like 80% or more points with Black.
More. Please check the standart deviation for your result. I'm dead sure your 54% lies in 3σ interval from 50% because there is NO advantage of Black in renju thanks to swaps. There may be some advantage of the 1st number by S8 because it is 1st player who puts the opening, but on the top level I doubt it's statistically significant.
For example, if you check, say, 500 games, and all of them are won by somebody, black won 270, white won 230, the standart deviation for this situation is sqrt(Npq) = sqrt(125), about 11. Here we can't say that the null-hypothesis (50% for both colours) seems to be wrong.
@ owen_ - has anyone that you know of tried pro rule ninuki with 4x4 forbidden for black? It may indeed be a sure black win, but at what what level of play? It took a few years of non pro rule 4x4 not forbidden play by high dan renju players to determine that ninuki wasn't working out. How about on a 14x14 board, where black plays to one of the central 4 intersections? A 13x13 board alters some standard Pente lines, preventing a first player win on those lines.
Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
Hi! First of all, how is 4x4 square defined? Current renju rules are stricting 1st move to 1x1 center square (exactly one point), 2nd move to 3x3 sqare (any point adjacent to the center), 3rd move to 5x5 square. What is 4x4, you mean 7x7?
--- Concerning the pro rule.
The next thing is that the vast majority of gomoku players sticks to swap2 rule since it gives equal chances because of swap possibility and a wide range of openings. So these two problems (inequality and a very narrow set of opening variations) make pro rule totally irrelevant. That's why there will be no tournaments with pro rule and such.
The same logic works for renju players who have tried 5 swaps rule (or T10, or anything alike). A huge bunch of playable positions in the opening stage, perfectly equal chances.
If some of these established gomoku/renju players will try to play Pente (actually we do play Pente, I play Pente with my students and there are some of my students on vint.ee playing Pente), they will not try to equalize chances using pro rule. They will definetely use swap2 or 5-swaps instead.
And this way seems to me to be the future of Pente actually. Either swap2 (gomoku way) or 5-swaps (renju way). If you are interested, I can try to write a comparison of these two approaches.
Nobody (or mostly nobody) will try pro rule because of the mere fact it isn't actual so (1) the game soon will focus on the best opening and optimal play for both players, quickly exhausting it; (2) narrow game tree and known outcome will bring the advantage to whom it may be brought. We can forbid some more openings as it was done on littlegolem but it isn't solving the problem.
When I said 4x4, I was referring to black's foul for double 4s. That was never used in ninuki.
I play swap2 Pente on vint.ee as well (pente_player). I'd enjoy playing your students there. Perhaps I already have played some of them. Swap rules are an improvement for Pente in that players no longer need to play sets of games, but there simply aren't balanced swap proposals in Pente. Whether 2 swaps or 5 swaps doesn't really matter, I don't think. Pente doesn't have a good possibility for draws, unlike renju and gomoku. The simplest rule to give it draws is overlines are allowed, but don't win, but those draws are rare and board size is irrelevant for them.
As to pro rules, quite a few high level renju players played with them and were among the best Pente players in the world not too many years ago. Istvan Virag, Alex Nosovsky and Viktor Barykin to name just 3. An understanding of pro rules lines is essential to becoming a great swap2 Pente player. I think pente_gon AKA Arcoroc at vint.ee (undefeated 2 time MSO tournament swap2 Pente player) will agree with me on that point. As a challenge to you and your students, I will play all of you collectively in a 2 days per move swap2 proposal (off center) of my choice.
Edit: actually, my off center 4 stone pro Pente rules style proposal is more suitable for 2 days per move with a single swap (D Pente). Either the position is balanced (lol) or you and your students can figure out and play the winning side.
Message was edited by: watsu at Dec 9, 2022, 3:07 PM
Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
> I was referring to black's foul for double 4s Ah, ok. Didn't catch it. Well, I have no idea how will Penju (Rente) be going so can really say nothing about it. But if you want to have a true mix, well, 3x3, 4x4, overline - they are all parts of Renju, so why to exclude 4x4 only.
> Perhaps I already have played some of them. Definitely you had, I've checked a couple of games. 1-1 with Alex, for example (he is 9 years old now, by the way).
> An understanding of pro rules lines is essential to becoming a great swap2 Pente player. Well, I believe you. As well as understanding how to win in gomoku with no opening rules (as it is played here) is a must to become a great Renju/Gomoku player. But it isn't why everybody should play pro rule (or play only i7/d4), is it? I doubt you can point even three strong renju or gomoku players who participated in real tournaments in past 5 years and who will prefer pro rule to swap2 or T10.
> As a challenge to you and your students, I will play all of you collectively in a 2 days per move swap2 proposal (off center) of my choice.
Sorry, but no way. I'm busy. More, why are you proposing 2 days, the idea is to give me less time to analyze because otherwise I can find better moves? When I was playing correspondence renju (under the same nickname, Owen), it was 135 days per game. It sounds a bit more reasonable but still I'm not quite interested, sorry.
And anyway that's about math. You have a good point that Pente isn't about draws. But if the game may be ended with a draw with optimal play (dunno how reasonable is it), it will after all.
If any opening may be solved (which sounds intuitively correct) than again you have to move from limited number of openings (pro rule) to a huge bunch of possible openings (swap2 or any other rule with a wide opening tree).
> Either the position is balanced (lol) or you and your students can figure out and play the winning side.
You know, I've initiated the correspondence gomoku world championship. During the first two seasons there were some players who offered me to solve their openings - and I did it several times (just to mention: it was allowed to use any programs). http://renju.in/igrat/game/?id=30177 http://renju.in/igrat/game/?id=30175 Both these openings were proven to be white win from a position with only 3 stones on the board. So I know how to do it despite the fact I'm not a gomoku player, I'm renju player. But I don't see why should I do it now, it takes a lot of time. ---- Even if the win isn't known now in some particular opening because of lack of strong programs or strong analysts, it will eventually be found. Just wait for a Katago release for Pente - and it will ruin your world =)
"1-1 with Alex, for example (he is 9 years old now, by the way)."
I enjoyed playing Alex, both in the MSO and our non tournament game. I look forward to playing him again in the future. I've been getting beaten by great 9 year old players for about 20 years, so I'm used to it
"But it isn't why everybody should play pro rule (or play only i7/d4), is it?"
No; I stopped playing TB pro rules Pente a few years ago due to the first player advantage. Is the first player advantage still there on a 13x13 (or 14x14) board with pro rule penju? If not, how great is the second player advantage (or does optimal play result in a draw) I doubt anyone has programmed it in and found out.
"why are you proposing 2 days"
Often, fast correspondence pro Pente tournaments were 2 days per move when Gary Barnes told me the best 2 days per move player was Dmitri Krasnonosov (close to 20 years ago, probably). I figured you and some of your students can find the solution to it, given enough time, yes. Correspondence Pente (swap rules included) isn't challenging for high dan renju players such as yourself. Live pro rules Penju, I would love to watch masters play, though. I'm a nobody gomoku player (and have hardly played renju), you're a renju sensei (Meijin?). Yet, I have still beaten a few renju senseis (in live swap2 Pente, which is not a very strong game for me, since I'm not good with openings and adding stones to proposals).
"Just wait for a Katago release for Pente - and it will ruin your world =)"
I look forward to that, actually. There are some slightly more complex variants of Pente which I hope will be included as well.
"And sorry for my poor English =)"
I know only a few words in Russian. Vint.ee has been a great site for allowing players who speak many different languages to play each other.
Message was edited by: watsu at Dec 10, 2022, 4:25 AM
Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
